MRC Cognition and Brain Sciences Unit

You are in:  Home » History of the Unit

3 COGNITION AND EMOTION GROUP PROGRESS REPORT

Overview

The aims of the Cognition and Emotion programme are 1) to increase basic understanding of the representation, elicitation, maintenance, and control of emotion, at neurobiological, cognitive, and computational levels of explanation, and 2) to apply that understanding to the analysis and treatment of disorders of emotion. Research spans a range from the detailed analysis of perceptual mechanisms of emotion recognition, using neuroimaging, neuropsychological, and cognitive methodologies, through to the development of novel therapeutic interventions and their evaluation in clinical trials. This breadth of inquiry is co-ordinated within multi-level theoretical frameworks developed by our group (e.g. Mathews & Mackintosh, 1998; Power & Dalgleish, 1997; Teasdale & Barnard, 1993). It is increasingly recognised that such frameworks are necessary for a comprehensive understanding of the multiple interacting processes that contribute to emotion production and control. Our research has generated a wealth of specific findings. Here, we preview some of the more important general conclusions that emerge from our studies. These conclusions offer guidance for deepening basic understanding, formulating further research, and improving patient care.

Neuroimaging and neuropsychological studies conducted in Projects CE1 and CE6 have produced critical new evidence converging on the conclusion that multiple brain systems interact in processing emotional information. In summary, some brain systems (e.g. amygdala, insula) have been shown to be involved in processing affective stimuli associated with specific basic emotions while others (e.g. medial frontal cortex) appear to be involved in a range of emotions, drawing on multiple sources to compute more complex emotional and social meanings. Yet others (e.g. anterior cingulate cortex and dorsolateral frontal cortex) are concerned with higher level control of both emotional and non-emotional processing. Differences in neurotransmitter receptor densities in some of these systems have been shown to correlate with trait differences in emotionality, providing a new means of testing theories of how emotions and personality traits are related. The propensity to experience emotions thus depends on the outcome of interactions among several different brain systems. The co-existence of specificity and generality of emotion-related brain functions is precisely what our multi-level analyses of emotion predict, and reinforces our strategy of combining detailed investigation of individual component systems, with the development of frameworks that specify how those components interact.

Work on recognition of emotional expression (Projects CE2 and CE6) illustrates how the detailed investigation of particular component systems can yield an elegantly simple model that captures key features of a range of functions. Building on earlier work on recognition of facial identity, research in the reporting period has shown that multidimensional principal component analysis provides compelling models. These can account not only for recognition of identity, but also of emotional expression, gender, the dissociation of identity and expression recognition, effects of emotional intensity, and the configural nature of expression processing. The success of this work indicates the potential fruitfulness of focusing analysis at the level of component systems.

At the same time, the strategy of focusing investigation on component systems and subsystems needs to be complemented by analyses that focus on the patterns of interaction between subsystems that support more complex functions. This is a formidable challenge. Project CE3 describes the early stages of development of a radically new approach to computational modelling which has explored the distributed control of exchanges between component subsystems within a wider, distributed system. This modelling work has been complemented by empirical work in both clinical and non-clinical populations. This has shown that important aspects of central executive function can be understood in terms of changes in the patterns of interaction between subsystems, rather than deficits or anomalies in the subsystems themselves. This work validates the potential usefulness of this unique approach to computational modelling. It indicates that we can now move forward to model further aspects of the exchanges between subsystems that our models suggest are of central importance in cognitive-affective function.

The interaction of automatic affective and controlled cognitive processing is particularly important in understanding the maintenance of affect. Project CE4 addressed instances of such processes associated with vulnerability to anxiety and depression, such as selective attention to threatening information, and perception of the more negative meanings of ambiguous events. These processes have the effect of making potential threats more intrusive, and in this way can maintain negative emotions. New evidence shows that opposing positive biases occur in healthy people, but are not present in anxious patients. A similar failure of positive bias can be induced experimentally, producing symptoms characteristic of anxiety disorders. In healthy individuals, task-irrelevant aversive stimuli are typically easily ignored, but work in this project has revealed that those vulnerable to negative affect are less able to disengage from such stimuli, and to inhibit or forget unwanted aversive meanings. These findings show that stable individuals are able to inhibit mildly aversive information, but that inhibition failure is associated with vulnerability to anxiety or depression, and that such processing causes congruent changes in affective symptoms. In demonstrating the importance of such failures of disengagement and inhibition, these studies point to a need to shift the emphasis in existing explanatory models. Rather than attempting to understand emotional disorders solely in terms of excessive activation of dysfunctional representations and processes, such accounts also need to recognise more clearly the importance of failures of appropriate inhibition.

Self-perpetuating emotion-related cognitive biases, such as those studied in Project CE4, have figured prominently in the multi-level analyses of emotional disorders developed by our group. Given the weight attached to these biases in the maintenance of emotional disorders, such as major depression and generalised anxiety, it is important also to understand their role in the initial onset of disorder, To this end, Project CE5 included extensive investigations of emotion-related cognitive biases in children with emotional disorders, many of whom were experiencing their first episodes, and in children who were at risk for disorder but who had not yet experienced an onset of the disorder. Findings demonstrated a pattern of disorder-specific cognitive biases that were very similar to those found in adult disorders, and that, in some cases could be shown to pre-date onset of disorder. These findings are important in suggesting that the identified self-perpetuating cognitive biases may be risk factors contributing to the initial onset of emotional disorders, rather than being primarily aftereffects of the experience of the initial episode.

Neurosurgical interventions on the intact human brain offer a rare opportunity to investigate the neural substrates of emotional disorder with the precision that is more usually associated with studies of experimentally induced lesions in animals. Project CE6 took advantage of such an opportunity in studying the effects of stereotactic subcaudate tractotomy for the relief of treatment resistant depression. The findings indicated a remarkable degree of specificity in the effects of this operation on neuropsychological measures. These effects were specific to reductions in patients' sensitivity to negative feedback information. This reduced sensitivity was shown only by patients who benefited clinically from the operation, and was not shown by either operated patients who failed to benefit from the operation, or by patients who did not receive the operation but whose depression improved as a result of pharmacotherapy. Such specificity is consistent with a causal role for reduced sensitivity to negative feedback in patients' clinical improvement, and for a role for heightened sensitivity in the maintenance of chronic depression. This study demonstrates the potential utility of this investigative strategy, as a way of elucidating simultaneously both the nature of depressive psychopathology, and the role of specific regions of the human brain.

Our multi-level analyses of emotional disorder implicate self-perpetuating cognitive-affective processing configurations in the moment-to-moment dynamic maintenance of states such as generalised anxiety and major depression. A crucial test of such models is the extent to which particular cognitive processing styles, that we believe play a causal role in perpetuating aversive emotional states, can be experimentally controlled. Project CE7 has shown that a threat-related attentional bias, of the kind assumed to maintain generalised anxiety, can be induced to previously neutral stimuli by associating them with threatening meanings, without participants' awareness. Further, practice in accessing negative meanings of ambiguous material was shown to lead to a related bias in processing new ambiguous information, together, most interestingly, with the development of a tendency to respond with anxiety to such information. These findings provide the first evidence that processing biases can indeed be a cause of anxiety. In major depression, brief attentional manipulations were found to reduce overgeneral memory, a characteristic feature of individuals prone to depression. These findings are consistent with the idea that overgeneral memory, like other features of depression, is dynamically maintained by self-perpetuating processes that can be interrupted, rather than an enduring trait-like feature, as was generally previously assumed. Further, this work has characterised two distinct modes of self-attention, one associated with perpetuation of depressive cognition, the other with its amelioration. Thus, laboratory-based manipulations can bring cognitive processes characteristic of emotional disorders, under experimental control. By manipulating the form of these processes, we have shown that they have causal effects on symptoms of emotional disorder. Further, this work has identified potential therapeutic processes for inclusion in clinical treatments.

An important motive for our more basic research on the interaction between cognition and affect in normal and disordered emotion has been to provide a firm basis for the development of new and more effective approaches to the treatment and prevention of emotional disorder. Project CE8 illustrates how our theoretical and experimental studies led to a radically different way to understand the effectiveness of an existing treatment for depression, cognitive therapy, which then provided the basis for the development of a new cost-efficient programme to prevent relapse in recurrent depression. Specifically, we gathered support for our hypothesis that cognitive therapy works through implicitly changing the mode of processing of negative thoughts and feelings, rather than through changing belief in the content of negative thoughts, as is generally assumed. This hypothesis guided the development of a programme explicitly designed to achieve the same ends as cognitive therapy, through systematic training in the control of modes of attention. In two clinical trials, we showed that this intervention, firmly rooted in theoretical and experimental work conducted within the Cognition and Emotion group, was highly effective in reducing relapse in recurrent depression, and did so in a way consistent with the processes that we had specified. This work validates our strategy of developing new approaches to patient care grounded in our basic research on cognitive-affective interaction.

Project CE1: Affect-related representations.

Scientific Direction: Calder (5%), Lawrence (70%), Mathews (20%), Teasdale (5%).
MRC-supported scientists: Murphy (100%)
Grant-supported scientists: Bishop (15 months, 100%), Potts (from January '02, 100%).
Research support: Green (5%), Keane (5%), Yiend, (20%).

The multi-level information processing analyses of emotion developed within the Cognition-Emotion group identify both sensory/perceptual and cognitive/interpretative levels of representation in the generation and regulation of affect, and in the interaction of cognition and emotion. The aim of project CE1 is to use functional neuroimaging and complementary techniques to identify the neural substrates of these interacting levels in cognitive-affective processing. Findings suggest that some brain systems are involved in processing specific types of sensory/perceptual emotional stimuli (e.g. amygdala, insula), while others (e.g. medial frontal cortex) probably draw on multiple sources of information to compute more complex emotional meanings. Yet other systems (e.g. dorsolateral frontal cortex) are concerned with higher level control of both emotional and non-emotional processing.

CE1.1 The cognitive generation of affect
The Interacting Cognitive Subsystems (ICS) framework (Teasdale & Barnard, 1993) suggests that the cognitive generation of affect depends on the integration of propositional meanings with perceptual input, to form schematic affective representations. Teasdale, in collaborative work at the Institute of Psychiatry, has conducted fMRI studies aimed at identifying brain networks mediating such cognitive generation of affect. These studies compared the activation elicited by presenting pictures and captions combined so as to represent a meaningful emotional event, or in a re-combined form so that they no longer meaningfully cohered. This methodology controls precisely for all aspects of stimulus materials other than the extent to which captions and pictures cohere. It follows that comparison of coherent and non-coherent caption-picture pairs identifies structures mediating the specifically meaning-related contribution to the generation of affect. Using this approach, Teasdale, Howard, Cox, Ha, Brammer, Williams & Checkley (1999) identified medial prefrontal networks related to the processing of schematic-level meanings and the cognitive generation of positive and negative affect. Networks activated were similar to those reported in other affect-elicitation imaging studies (see CE1.3); the importance of this study is the precision with which the contribution of complex emotional meanings to affect-elicitiation can be specified.
A subsequent study using the same methodology (Kumari et al., in press) replicated these findings for normal subjects, and compared them with treatment resistant depressed patients. Compared to the controls, patients showed significantly less activation in the identified medial prefrontal networks to positive caption-picture pairs, suggesting deficits in the processing of cognitively generated positive affect in treatment resistant depression.

CE1.2 Eye gaze processing
The direction signalled by another person's eye gaze not only carries information about where the other person is looking (i.e., attentional cues), but also that something, or someone, has captured the gazer's attention. Consistent with the latter, Baron-Cohen (1997) has proposed that the interpretation of gaze plays an important role in a normally functioning theory of mind (ToM) system. Calder validated this role by showing that a meta-analysis of functional imaging research demonstrated that both ToM and eye gaze tasks engaged a similar region of posterior superior temporal sulcus (STS) (Calder et al., 2002). In addition, he noted that a second, more prominent brain region associated with ToM, the medial prefrontal (MPF) cortex, had not been identified by eye gaze research. In collaboration with Lawrence, Calder identified methodological issues that might account for the absence of MPF activation in these experiments. He then conducted a PET study that controlled for these factors and addressed the neural correlates of processing direct and averted gaze (Calder et al., 2002). The results showed that the MPF regions associated with ToM were indeed involved in processing gaze, but particularly averted gaze. Moreover, because participants were not explicitly asked to attend to the faces' gaze, the study demonstrates that simply viewing a face with averted gaze is sufficient to engage the mechanisms involved in ToM. This provides the first demonstration that the mechanisms involved in processing another person's mind state are engaged automatically. The MPF regions involved were very similar to those implicated in the cognitive generation of affect in CE1.1, consistent with a role for these brain regions in the processing of affect-related implicit meanings, and contributions from such meanings to gaze-related affect generation.

CE1.3 A novel imaging meta-analytic approach to understanding the structure of neural emotion space

The conclusions that can be drawn from any single functional imaging experiment are, necessarily, limited. To overcome this difficulty, Murphy, Nimmo-Smith & Lawrence (Lawrence & Murphy, 2001; Murphy, Nimmo-Smith & Lawrence, submitted) have applied novel statistical techniques to the meta-analysis of very large numbers of functional imaging experiments. Using a variant of the Kolmogorov-Smirnoff statistic, they examined the distribution of patterns of activation foci in functional imaging experiments, and how these distributions relate to models of the neural representation of emotions. Using data pooled from over 100 experiments, they tested predictions made by categorical (e.g. basic emotions) versus dimensional (e.g. approach/avoidance; valence/arousal) frameworks. In support of basic emotions accounts, the distribution of activation foci associated with discrete affects (fear, anger, disgust) were significantly different from each other, and from happiness and sadness. In contrast, the activation peaks for happiness and sadness were not different from each other. In addition, fear, disgust and anger activations were maximal bilaterally in regions which, when damaged, are associated with selective deficits in processing these emotions (fear - amygdala, disgust - insula, anger - ventral frontal cortex - see CE6); whereas happiness and sadness activations tended to cluster in the anterior cingulate. Activation of anterior cingulate cortex was associated with the majority of emotions, with the exception of disgust, suggesting that this region plays a generic role in the processing of a range of affects - consistent with multi-level representation of emotion. There was also some support for a differentiation of emotions in terms of associated action tendency (approach / avoidance), but no evidence for coding based on valence or arousal dimensions.
CE1.4 Emotional associative learning
The study of reward learning in animals has revealed an interrelated set of limbic and cortical structures involved in reinforcement and motivation. Emotional associative learning in humans has not yet been much investigated from a similar neuroscientific perspective. Conditioned place preference is one of the most common procedures for assessing stimulus-reward associative learning in animals, and Johnsrude, Owen, Zhao, & Whitney (1999) have developed an analogous procedure for use with people. Using this procedure, Johnsrude, Owen, White, Zhao & Bobhot (2000) have shown that unilateral removal of the temporal lobe, including the amygdaloid region, in humans (for the treatment of epilepsy), abolishes either the formation or expression of such conditioned preferences. In contrast, patients with unilateral frontal-lobe removals were unimpaired. This result provides clear evidence that, in humans as in other animals, reward-related learning critically depends on a circuit involving inferotemporal cortex and the amygdaloid region.

CE1.5 Interactions of emotional and cognitive control systems in the brain

Interactions between emotional processing and cognitive control systems are assumed to play a central role in the maintenance and regulation of affect. Mathews & Mackintosh (1998) proposed that threatening representations (of stimuli such as fear-related pictures) receive additional activation from a threat evaluation system (involving the amygdala complex), so that they are more likely to capture attention. Such activation may be inhibited, if it would conflict with other current goals, by competing activation of task-related representations from a control system, involving medial and dorsolateral areas of the frontal cortex. Mild emotional distracters may thus be ignored, and only more severe threats will disrupt attention to ongoing tasks (Mackintosh & Mathews, in press). Unintended attention to threatening distracters, which characterizes individuals vulnerable to anxiety states, can thus arise due to a lower response threshold in the threat evaluation system, and/or relatively weak inhibition by the control system.

Using fMRI, Mathews and Lawrence investigated the neural correlates of encoding manipulations on threat related material, in high and low anxious subjects, to test predictions made by the above model. Subjects viewed sets of pictures without specific instructions, and other sets while performing tasks: either judging if the current picture is more fear-inducing than the previous one (emotional task); or if it required more preparation by the photographer (non-emotional task). Findings so far are that threatening pictures provoke more activation in the primary visual cortex, parietal regions, and amygdala than neutral pictures. A parallel psycho-physiological study employed a noise-induced startle response to index amygdala activation. Startle magnitude was augmented when viewing threatening versus neutral pictures, and this difference was enhanced by judgements of fear, but was abolished by the non-emotional judgement task. These findings suggest that affect-related amygdala activation can be modulated by conscious cognitive control.

In collaboration with Bishop, Brett and Duncan (attention group), Lawrence conducted an event-related fMRI study to investigate the role of different prefrontal cortical structures in the processing of emotional material. Using a task which requires matching pairs of houses in the presence of emotional distractors (fearful faces), combined with a trial structure manipulation (frequent versus infrequent distractors) known from previous work to manipulate levels of cognitive control, this study demonstrated a role for the anterior cingulate (affective division) in detecting the presence of response conflict engendered by the presence of emotional distractors. In contrast, the dorsolateral prefrontal cortices were involved in implementing attentional control. Differences in dorsolateral prefrontal cortex were also seen as a function of anxiety levels. This work showed, for the first time, that different structures (anterior cingulate and lateral prefrontal cortex) play dissociable roles in control of emotional material (Bishop, Duncan, Brett & Lawrence, in submission). Further, it showed that areas already known to be involved in controlling attention in non-emotional tasks (the anterior cingulate and dorsolateral frontal cortex) are also employed when controlling response to emotional information.

CE1.6 Individual differences in affective style and temperament and their neurobiological correlates

Better understanding of the neurobiological substrates of stable individual differences in temperament and affectivity is centrally relevant to improved models of psychopathology. One of the most powerful theoretical frameworks for explaining such individual differences and their relation to emotion systems is the reward sensitivity theory of Gray (Gray, 1982). Gray's model postulates two major dimensions of temperament, anxiety and impulsivity, that represent individual differences in the sensitivity of two neural systems involving processing of aversive (Behavioural Inhibition System, BIS) and appetitive (Behavioural Activation System, BAS) cues, respectively. Certain forms of psychopathology can be characterized in relation to these major axes of personality variation (Pickering & Gray, 1999). For example, a link between the positive symptoms of schizophrenia and BAS function has been proposed (Pickering and Gray, 1999) and clinical anxiety is thought to be related to high levels of BIS activity.

Lawrence, in collaboration with Goerendt, Rabiner, McGowan, Brooks and Grasby at MRC Clinical Sciences Centre, London, have shown that individual differences in serotonin 5HT1a receptor densities in the hippocampus and amygdala are associated with individual differences in BIS function (as measured by the difference between scores on extraversion and neuroticism), but not with neuroticism or extraversion per se (Rabiner et al., 2002a and unpublished data). These data are consistent with Gray's model.

An intriguing finding from these initial studies was that 5HT1a receptor levels were associated also with scores on the lie scale from the Eysenck Personality Inventory (Rabiner et al., 2002a). Although initially thought to be a measure of false but socially desirable responses, it has been known for some time that this factor is in itself a substantive aspect of personality, and individual differences in this trait (sometimes called defensiveness) have been related to lifetime risk of psychopathology. Rabiner et al., (2002b) have replicated this finding in a new group of individuals. The finding of a relationship between scores on the lie factor and 5HT1a receptor levels has implications for conceptualisations of a serotonin-linked trait as well as factors predisposing to risk of affective disorder, and suggest revisions to reward sensitivity theory will be required.

Gray's model also predicts that individual differences in BAS system function will relate to indices of dopaminergic activity. Lawrence, together with Goerendt and Brooks, MRC CSC, have used [18F]Fdopa and [11C]Raclopride PET to measure individual differences in dopamine uptake and D2 receptor status, and have found that measures of impulsiveness are correlated with ventral striatal activity, but only in male participants. In addition, they have used H2[15]O PET to map the structures involved in processing cues thought to activate the BAS (financial rewards), and the modulation of these structures by dopamine neurotransmission (studies in Parkinson's disease). This work has shown that ventral fronto-striatal circuits mediate BAS function, and these structures are less activated in unmedicated PD patients, especially those showing symptoms of apathy, but can be restored by treatments that increase dopamine levels. In addition, behavioural work has shown that these circuits mediate the incentive motivational (speed enhancing) effects of monetary rewards.

A neuropsychological study found that individuals with anorexia nervosa, a profound disorder of appetitive function, show highly selective deficits in reward-based learning, consistent with a rather specific deficit in a BAS system. In contrast, they show no impairment on difficult pattern recognition memory tasks associated with medial temporal lobe function (Lawrence, Dowson, Foxall, Summerfield, Robbins & Sahakian, in press). Gray's model predicts an association between the neural systems implicated in BAS system function and the neurobiology of positive schizophrenic symptoms (Pickering & Gray, 1999). In collaboration with McGowan and Grasby (MRC CSC, London), Lawrence has used PET imaging (blood flow and [18F]Fdopa to image dopamine function) to test this hypothesis. In line with Gray's theory, schizophrenic patients showed increased [18F]Fdopa values in the ventral striatum, related to their degree of positive symptomatology. In addition, [18F]Fdopa values were correlated with performance on tasks requiring the use of stored regularities to guide behaviour (such as verbal fluency) (McGowan, Lawrence & Grasby, 2001b, submitted).

Project CE2: Recognition of emotion

Scientific Direction: Calder (55%).
Research support: Keane (55%).

The majority of research in facial expression processing has focused on the communicative value of facial expressions, and few studies have investigated the perceptual processes preceding this stage. As a result, there is no detailed model of facial expression recognition. Project CE2 addressed the perceptual mechanisms underlying the recognition of facial affect, with a view to developing a model of facial expression processing. The resultant multidimensional framework, based on principal component analysis (PCA), can account for all of the phenomena discussed below.

Figure 1: The first eight eigenvectors (eigenfaces) extracted from a PCA of the Ekman and Friesen (1976) pictures of facial affect. The eigenvectors have the same dimensionality as the original starting images so they can be displayed as visual images that have a ghost-like facial appearance. Note that the eigenvectors clearly capture aspects of the different expressions.

CE2.1 Front-end coding of facial expressions

It is generally agreed that facial identity recognition (who the person is) and facial expression recognition (what they are feeling) share the same front-end (perceptual) system. Previous research has shown that a principal component analysis (PCA) of the visual information in faces provides an effective front-end account of facial identity processing (Burton, Bruce, & Hancock, 1999). Hence, two critical questions are whether PCA can also support the recognition of facial expressions, and whether it provides a suitable means of coding a face's identity, expression, and sex within a single dimensional framework. Calder has addressed these issues by submitting the pixel intensities of pictures of facial expressions from the Ekman and Friesen (1976) series to a PCA (Calder, Burton, Miller, Young, & Akamatsu, 2001a) (Figure 1). The results showed that PCA provides a highly effective means of coding all three facial attributes. For facial expressions, the correct recognition rates and false positives derived from the principal components were well matched to human performance. In addition, the model exhibited properties of two competing accounts of facial expression processing (dimensional and category-based models), providing a means of reconciling what were generally perceived to be distinct theoretical accounts. Consistent with research showing that facial identity and facial expression recognition can be selectively disrupted, Calder found that cues to identity and expression were coded by largely separate sets of principal components. A similar dissociation was found for expression and sex, while, consistent with recent cognitive research, identity and sex were coded by largely overlapping sets of PCs (Ganel & Goshen-Gottstein, 2002). This research shows that linearised compact coding of human faces provides a highly plausible account of the psychological data.

CE2.2 Configural coding

It is well established that configural information (the relationships between facial features) plays an important role in coding a face's identity. However, its contribution to facial expression recognition is less well understood. In fact, some researchers have suggested that facial expressions are processed in a part-based (non-configural) manner (Ellison & Massaro, 1997). Calder addressed this issue using a composite paradigm (Calder, Young, Keane, & Dean, 2000d). Participants were slower to identify the expression in either half of composite facial expressions: that is, faces in which the top half of one expression (e.g., anger) was aligned with the bottom half of another (e.g., happiness) to create a novel expression configuration relative to a noncomposite control condition in which the two face halves were misaligned. These findings support the role of configural coding in facial expression recognition, and parallel the composite effect for facial identity (Young, Hellawell, & Hay, 1987). However, Calder also showed that the identity and expression effects operate independently of one another (i.e., the configural cues to these two facial attributes are qualitatively different). This research complements the findings of the PCA (see above) which showed that identity and expression are represented by separate principal components. In line with this observation Cottrell (California, San Diego) and Calder have shown that PCA provides an effective model of Calder's composite data (Cottrell, Branson, & Calder, 2002).

CE2.3 The representation of facial expressions

Facial expression recognition has been discussed in terms of two types of models. In category based systems, each expression is identified by activating a discrete representation. By contrast, in low dimensional accounts, facial expressions are identified by judging their values on two, or three continuous dimensions. These models have been applied to the recognition of emotion in general. However, as illustrated by the principal component analysis of facial expressions discussed above, they are also competing candidates for the perceptual representation of facial affect. To address this debate, Calder conducted a number of experiments using photographic-quality morphs (blends) and caricatures of facial expressions prepared using a computer-based procedure.

CE2.3.1 Categorical perception

Two studies used a categorical perception paradigm (Harnad, 1987) to address participants' perception of morphed facial expression continua ranging between pairs of facial expressions (e.g. anger and disgust; Calder, Young, Perrett, Etcoff, & Rowland, 1996a; Young et al., 1997). Results showed that all morphed images were consistently identified as the expressions at either end of the relevant continuum, with a sharp category boundary at the middle. In addition, pairs of morphs that straddled the category boundary were better discriminated than pairs of equal physical magnitude from either side of the boundary. Calder and Young have argued that these findings are consistent with the view that facial expressions are coded as discrete categories. In his work on PCA, Calder has proposed that facial expression categories can be viewed as clusters of vectors in a multidimensional space, with each cluster representing an attractor state for one facial expression (Calder et al., 2001a). Hence, categorical perception is consistent with a multidimensional account of facial expression coding. This research has also led to the production of a neuropsychological test of facial expression recognition (The Emotion Hexagon) that uses morphed facial expressions. To date, the test has been used in a number of studies, including those in CE6.

Figure 2: An example of a caricature (+50%) and anticaricature (-50%) image of an afraid facial expression. The images were produced by exaggerating (caricature), and reducing (anticaricature), the differences between the original afraid expression (0%) and a picture of the same model posing a neutral expression (left).

CE2.3.2 Emotional intensity

An important factor that distinguishes facial identity and facial expression is that the latter vary in intensity. Calder investigated the intensity issue using photographic-quality caricatures of facial expressions. Caricatures were produced using a computer graphics technique that exaggerated the physical differences between a particular facial expression (e.g., fear) and a picture of the same model posing a neutral expression. Similarly, anticaricature expressions were produced by reducing these differences (Figure 2). Subjects were significantly faster to recognise caricatured expressions relative to the original undistorted images, whereas anticaricatures produced slower recognition rates (Calder, Young, Rowland, & Perrett, 1997). An additional study showed a positive linear relationship between level of caricature and rated emotional intensity – but only for the emotional scale corresponding to the emotion displayed (i.e., fear caricatures were rated as expressing more fear, but not as more disgust, anger, etc.; Calder et al., 2000c). Caricatured expressions have provided a valuable stimulus material for Calder's collaborative functional imaging projects (see CE6). These studies mirror the findings discussed above by showing that caricatures of fear and disgust engage different brain regions, with changes in neural activity being positively related to level of caricature (Morris et al., 1998; Morris et al., 1996; Phillips et al., 1997; 1998). Calder has shown that caricature effects for facial expressions are best accounted for by an exemplar-based model of facial expression coding, analogous to the PCA architecture discussed above (Calder et al., 2001a; Calder et al., 2000c). In such a system, each expression category is coded as clusters of exemplars (vectors) in multidimensional space, and caricaturing operates by shifting the expression along the vector.
Summary: This research programme has addressed the perceptual processing of facial expressions from front-end analysis through to the point of representation. The results demonstrate that a multidimensional system based on a form of linear encoding analogous to PCA, provides a suitable metaphor for processing facial affect. Moreover, this single PCA-based system provides a model of coding a number of facial attributes (i.e., facial identity, expression, and sex). In addition, it can account for the dissociable effects observed for facial expression and facial identity processing (e.g., different configural cues for expression and identity), and single route coding of identity and sex (Ganel & Goshen-Gottstein, 2002). The same multidimensional metaphor also provides a suitable account of the effects of categorical perception, caricature, and configural processing described.

Project CE3: Multi-level representation and processing dynamics

Scientific Direction: Barnard (90%).
Research support: Ramponi (90%).
MRC-supported students: Battye (3 years, 100%).
Grant-supported students: Tavares (1 year, 100%).

Multi-level theories focus on different types of representation and the parts they play in the generation and maintenance of affective states. The need for theories of this class, their properties, and their potential for linking information processing to neural level analyses have been reviewed (Teasdale, 1999a). Such theories provide an explicit basis for the distinction between a traditionally cognitive, or propositional, level of representation, within which emotion-related material may be discussed rationally, and non-conceptual levels of representation and meaning that are directly linked to affect.

The Interacting Cognitive Subsystems architecture (ICS) is an example of a multi-level theory that is formulated with sufficient precision to support both empirical studies and computational research. It views central executive functions as an emergent property of reciprocal exchanges between one subsystem that represents specific meanings as propositions and another subsystem that represents implicational meaning. Implicational meaning is cognitive-affective and takes the form of abstractly encoded schematic models of self, others and the world. These bring together the products of propositional cognition and the products of sensory processing, including body states. Exchanges between two levels of representation act as a real-time control loop for ideation and memory access, without theoretical reference to general-purpose resources such as a unitary central executive or a limited capacity working memory system (Barnard, 1999).

As a system-level approach, exchanges among levels of representation within ICS can operate in a number of alternative configurations. The configurations differ in the parts played by processes in the propositional and implicational subsystems (Barnard, 1985). A process can operate in automatic mode, relying solely on its immediate input to generate an output. Alternatively, a process can be re-configured to use an extended representation of recent input (buffered mode), or it can access longer-term regularities stored in memory (record access). Propositional-implicational exchanges can be co-ordinated by buffering propositional meanings, and their specific content, or by buffering schematic meanings with their more generic properties. In the propositional mode, internal attention and experience is focussed on specific meanings, while in the implicational mode, it is focussed on more generic, and potentially affectively charged meanings. As with many other theories, dysfunctional ideation is assumed to call upon abnormal schematic models of self, others and the world. What ICS adds is the hypothesis that key properties of normal and dysfunctional processing arise from an interaction between schematic models and the alternative modes in which they are processed while attending to meaning or retrieving it. This dynamic approach led to what has now become an influential account of depression (Teasdale & Barnard, 1993). It links depression to a state of interlocked processing between two levels of meaning in which attention is predominantly focussed on the propositional meanings that are generated by depressogenic schematic models. The general hypothesis is that executive control of normal and dysfunctional ideation depends upon models and modes. Specific control parameters, such as the salience and extent of discrepancies, or rates of change in information content, are hypothesised to determine how use of modes is dynamically co-ordinated over time.

In the current reporting period specific empirical projects have investigated how schematic models and modes influence the executive control of ideation in healthy volunteers and in patient groups. Theoretical work has tackled the difficult problem of how to resolve the computational issues that must be addressed when modelling complex systems with distributed rather than centralised control, and several forms of novel computational approaches have been developed. Additional theoretical work extending the ICS account from depression to other psychopathologies is described in CE5.

CE3.1 Attention to meaning in healthy volunteers

In everyday life there is natural variation in the extent to which we attend to detailed aspects of meaning. When asked a question like "How many animals of each kind did Moses take into the Ark?" many people will answer "Two" - not noticing that the relevant Biblical figure was actually Noah (Erickson & Mattson, 1981). Both Moses and Noah fit the same schema and this kind of phenomenon can be taken to index the extent to which respondents adopt an executive mode in which they analyse propositional meaning, or rely on more generically encoded schemata or models. The Moses illusion reflects the fact that much useful ideation can occur simply by glancing at schematic models in implicational mode. If no global discrepancy in meaning is detected from the glance, then a response can be generated without switching to the more time-consuming propositional mode required to take a detailed look at specific meanings. As a background for studies of patient populations, we investigated this type of phenomenon in normal healthy volunteers using two paradigms, one involving question-answering and the other involving watching rapidly presented lists of words belonging to a generic category and detecting targets that belonged to a semantically specific category.

CE3.1.1 Modes, schematic models and question-answering

Laboratory studies of executive function in healthy volunteers are likely to have over-estimated the extent to which ideation works in the analytic propositional mode, and underestimated how use of schematic models constrains understanding. Most tests take repeated measurements on similar problems and thus encourage participants to develop special-purpose executive strategies to minimise error. We asked single unexpected questions requesting a single large number in between two bounds (e.g. Can you please give me a random number between one million and ten million?). Small differences in question phrasing invoked very different distributions of responses. These allowed us to determine the generic features of numbers, such as their magnitude, parity, and complexity, that had been attended to, and hence we could infer what schematic models had been activated by the different questions (Scott, Barnard & May, 2001). As with the Moses illusion, we found that volunteers often failed to focus their attention on the relevant bounds, and did not check their responses prior to generating them. They frequently erred by reporting numbers outside the requested range. A follow up study also varied question phrasing but this time varied processing load by increasing the length of the number sequence to be generated. Here, we found that executive error rates in healthy populations occurred with 50% of the volunteers under conditions of high load (Barnard, Scott, & May, 2001). In both studies we demonstrated that the kind of executive errors often committed by frontal or manic patients can be understood as phenomena that also occur within the range of normal processing, albeit perhaps at the extreme of that range (see also CE5.2.2). Both the occurrence and nature of, these errors can be understood as resulting from attention primarily being paid to generically encoded schematic models rather than propositional details.

CE3.1.2 Modes of processing meaning and the Attentional Blink.

In another series of studies, we developed a novel variant of the attentional blink (AB) paradigm (Raymond, Shapiro & Arnell, 1992). We asked volunteers to detect an instance of a semantic category (jobs/occupations) in rapid serial visual presentation where most of the words belonged to a category (nature words) unrelated to the target. Our procedure ensured that volunteers could only determine what was and was not a target by processing meaning. If the sequence included non-target distractors, (e.g. miser) that were related in meaning to a subsequent target, then report of that target was substantially impaired. While evaluating the precise meaning of the distractor, they missed the target much more frequently than in a control condition with no distractor. Only minor impairment occurred in response to a category change unrelated in meaning to the target category. To have an effect the distractor had to be semantically salient. We used a sophisticated form of statistical semantics called Latent Semantic Analysis (Landauer & Dumais, 1997) to index semantic relatedness and to eliminate a simple priming explanation. The results supported an explanation in which an initial glance at implicational meaning marks an item as salient and triggers an extended look at its propositional meaning (Barnard, Scott, Taylor, May & Knightley, submitted). This demonstrated that AB onset and duration could be taken to index the time course of processing generic and specific meanings. It also led to a more controversial conclusion. Results obtained in the standard AB paradigm with coloured letters are now potentially open to explanation based on purely semantic mechanisms.

CE3.2 Attention to meaning in depressed, manic and anxious states

Taken together, the question-answering and AB paradigms provided evidence that modes adopted when paying attention to two levels of meaning are open to measurement and varied systematically in healthy volunteers. Since implicational meaning encompasses affect, the same tasks can be used to investigate the effects of different emotional states on the executive control of attention to meaning as a function of hypothetical parameters such as rates of information change or personal salience.

CE3.2.1 Executive modes in manic and depressed states.

In depressed states, Teasdale & Barnard (1993) argued that an analytic, propositional mode is preferentially adopted when processing negative self-models. Repeated generation of the same thought also implies a low rate of change in generic self-models or representations, with rather more moment to moment change in specific propositions. We have also argued that a more experiential, schematic mode is adopted in mania where high rates of change in generic representation occur in the form of flights of ideas (Barnard, in press). Our theoretical analysis argued that attention moves to the level of meaning with higher rates of change. This predicts that the Moses illusion should be less likely to occur in depressed mood states than in elevated mood states. It follows that bipolar patients, who display both extremes and can act as their own controls, should detect false presuppositions more readily when depressed than when manic. This prediction was confirmed in a longitudinal study using a task in which the pragmatic implications of a preceding statement were questioned. Patients in a depressed state were more likely to detect a false presupposition than when they were in a manic state. Patients' ability to attend to successive non-semantic targets (i.e. digits), as indexed by the SART (Robertso, Manly, Andrade, Baddeley, & Yiend, 1997), showed comparable decrements in both states (Barnard, Palmer & Scott, & Knightley, submitted). Hence, a simple explanation based on general attentional mechanisms being more widely disrupted in mania than in depression could be ruled out, and our effect linked to the preferential adoption of different modes of processing meaning.

CE3.2.2 Executive modes and schematic models in anxious states.

The ICS analysis of alternative modes applied in processing two levels of meaning offers a novel theoretical perspective on the processing of threat-related material by people in anxious states. Monitoring for threat related material could plausibly rely on the mechanism taking brief glances at implicational meaning, while the engagement of attention with threatening material could involve taking a more extended look at propositional meanings. Task-related and individual difference variables should both influence a generic salience parameter controlling mode shifts. We investigated this possibility in a series of three studies using the attentional blink paradigm in which we introduced threat-related distractors.

We first showed that physical threat words, unrelated in meaning to the target category, were capable of capturing attention in the AB paradigm, but only for trait anxious individuals who were state anxious at the time. Furthermore, the serial position curves indicated that the blink for threat material occurred later, and was briefer in duration, than that obtained previously for distractors more highly related in meaning to the targets. This suggested that attention under these circumstances was captured but not followed by a detailed look of the kind that occurred in our earlier experiment with healthy volunteers (CE3.1.2). A second experiment showed that social threat words had only minimal effects, like those of a simple category change in our earlier experiments. The respective contributions of semantic similarity and threat-relatedness were then examined in a more sophisticated design that used Latent Semantic Analysis to index both threat-relatedness and semantic similarity. In a context where jobs/occupations are the targets, distractors like villain and miser both have high levels of semantic association to the target category, but differ in their threat associations. Words like seducer and vegetarian show low levels of semantic associations to targets, but again differ in threat associations. The results replicated the basic effects of both semantic similarity (see 3.1.2) and of threat. All these words capture attention significantly to a greater or lesser extent, but the time to taken to engage a detailed propositional evaluation of a distractor's meaning, the depth and duration of that evaluation, vary with indices of state and trait anxiety. Anxious individuals do not just show a deeper blink to threat. Rather, the way they pay attention to meaning follows different patterns over time. These studies showed that an adequate theory of executive aspects of attention to affect-related meaning will require detailed modelling of task-induced salience, personal salience attributable to alternative self-models, and parameters determined by current affective state (see CE3.3.1 below).

CE3.3 Depression, self-models & memory

In the ICS architecture, representations are preserved in both the short and long term in memory records linked to specific subsystems. Recall and recollection are intimately bound up with attention to different levels of meaning or form (Barnard, 1999). Properties of encoding and retrieval in depressed states are theoretically related to both modes of executive control and schematic models (Teasdale & Barnard, 1993). In the current reporting period, we have investigated relationships between executive mode and schematic models in depression.

CE3.3.1 Autobiographical recall in depressed adolescents

Working in collaboration with Park (Dept. of Psychiatry Cambridge), we investigated autobiographical memory in adolescents with major depression. Independent groups in episode or in remission were given a series of autobiographical prompts. Case grammar analysis, based on linguistic distinctions concerning the generic semantic roles assigned to human agents, was used to analyse the verbal protocols. We found that self-reference in natural discourse differed between individuals currently depressed and those in remission. Currently depressed adolescents made proportionately fewer references to self-as-agent in an event, more state-based references to self and self-experience, and they also exhibited heightened used of negation, when compared to those with a diagnosis but not currently depressed. This supports the hypothesis (Teasdale & Barnard, 1993) that the model in place changes on generic features, such as agency, when depression remits. Reference to self-as-agent was negatively correlated not only with depressed state but also with an index of ruminative response style. Hence this result extends previous research by demonstrating an effect using an implicit measure of self-modelling expressed in natural discourse, as opposed to those based on questionnaires. When compared to a matched non-clinical sample, both the depressed and remitted groups gave reliably more propositions that referenced content beyond the context cued, an index of non-specific, or overgeneral recollection within this paradigm. This result effectively confirms previous findings obtained with the standard Galton cuing paradigm but using a new measure. The amount of overgeneral, or non-context specific material produced did not correlate with measures of current state or rumination (Barnard, Park & Ramponi, in preparation). Whereas the standard Galton procedure simply differentiates specific and overgeneral memories, our more refined methodology suggests that some attributes of the schematic models in place vary with current state. However, the attributes that change do not appear to be those that lead to overgeneral memories, which implies that other attributes do not vary with current state. Plausible determinants of this contrast were pursued in the next study which examined recollection in both autobiographical and recognition memory for word lists.

CE3.3.2 Recollection, self-models, modes and dysphoric mood

Using non-clinical samples of dysphoric and non-dysphoric participants, we investigated whether or not autobiographical recollection and non-autobiographical recollection for word lists were governed by the same determinants. Interest was focused on an assessment of the contributions of two factors. A predisposition to use a propositional mode of executive control was indexed by a ruminative response style. This was predicted to be associated with the production of fewer specific autobiographical memories and poorer recollection of list items. Self-complexity has been hypothesised as a vulnerability factor for depression (Linville, 1987). Those with more complex schematic models are argued to be less vulnerable. In a previous reporting period we had shown that those with more complex descriptions of daily routines showed greater recollection of specific autobiographical events (Eldridge, Barnard & Bekerian, 1994). We therefore hypothesised a general relationship between the complexity of self-models and recollection. Those with more complex models should be able to elaborate processing of both autobiographical and non-autobiographical material, and hence show enhanced recollection. The effects of this variable should be independent of the mode of executive processing. The complexity of self-and other-models was assessed using the Levels of Emotional Awareness Scale (Lane, Quinlan, Schwartz, Walker, & Zeitlin., 1990). Autobiographical recollection was assessed using the standard Galton cueing technique, while recollection of emotional neutral word lists was assessed in the remember-know paradigm (e.g. Gardiner & Richardson-Klavehn, 2000). The same samples of volunteers were assessed on both tasks. The dysphoric group showed a substantial deficit in recollection of list items and also recalled fewer specific autobiographical memories. The two measures were related: those recalling fewer specific autobiographical memories also recollected fewer list items. In both tasks multiple regression showed that unelaborated schematic models and a ruminative response style independently predicted a substantial proportion of the variance in recollection, with dysphoria contributing little. From this we conclude that the properties of schematic models and modes of executive control are the principal determinants of impaired recollection across tasks, rather than depressed mood per se (Ramponi,& Barnard, submitted). Such an analysis provides a more precise foundation for developing computationally explicit models than those theories that simply attribute memory impairments to a deficit in controlled rather than automatic processing.

CE3.4 Addressing the challenge of modelling highly concurrent architectures

The development of computationally explicit models is particularly challenging in the domain of cognition and affect. Multilevel theories emphasise highly concurrent processing both within local networks and on the wider scale of multiple interacting subsystems. In the current reporting period three avenues of modelling research have been pursued.

CE3.4.1 Analysis and simulation of multi-component, concurrent architectures

For modelling interactions among multiple subsystems our approach has been to explore and develop concepts for systems analysis (Barnard, May, Duke, & Duce, 2000, 2001). We proposed four particular concepts to organise more refined analyses of the behaviour of complex information processing systems: the configuration of information processing resources, their capabilities, the requirements that must be met for those capabilities to work, and the mechanisms that control and co-ordinate resource use. These concepts support the development of methods, and notations for analysing the resource requirements of particular mental tasks (Barnard & May, 2000; May & Barnard, in press). They also provide the foundation for developing computationally explicit models of highly concurrent architectures where resources can be rapidly re-configured in real time.

In one form of modelling, production system techniques were reported that automate ICS-based theoretical reasoning (Barnard & May, 1999). A second approach used an augmented form of modal action logic to demonstrate how mathematics can give rise to formal proofs of the application of ICS in behavioural tasks requiring complex actions and the integration of multi-modal sources of information (Duke, Barnard, Duce, & May, 1998). In a third approach, we have used process algebra to develop a running computational model of the attentional blink phenomena (Bowman & Barnard, 2001; Barnard & Bowman, submitted). This implemented attention switching between two levels of meaning. It relied on the idea of monitoring salience by glancing at incoming implicational representations and then switching to a more detailed look at propositional meaning only for those items found to be salient. Our use of process algebra simulates mental operations at a level of abstraction rather higher than that afforded either by traditional connectionist or symbolic methods. It operates at a level of process exchange and control directly analogous to the box models of cognitive psychology. We believe that this novel modelling technique, and our specific model of the attentional blink, have both broken substantial new ground. By actively researching alternative modelling methods our commitment is to find the best way of modelling the behaviour of interacting subsystems, rather than focussing all our efforts around one modelling technology, such as symbolic, connectionist, or hybrid methods.

CE3.4.2 Relating the modelling of cognitive systems to the modelling of neural systems.

In parallel with the other strands of modelling research one of our PhD students (Battye) has developed a connectionist model of the emotional Stroop effect (Battye, Barnard and Page, submitted). This work identified problems with a prior connectionist model of the emotional Stroop (Matthews and Harley, 1996) and rectified them. One key insight obtained from this new model was that apparently conflicting psychological theories of the emotional Stroop effect can all be traced back to the activation of common pathways in an underlying connectionist model. Battye has also developed a connectionist model of the attentional blink effect that offers the prospect of accounting for related cognitive phenomena on the basis of common underlying mechanisms, and relating those mechanisms to candidate brain systems such as the anterior cingulate cortex (Battye & Barnard, 2002). Another graduate student (Tavares) is currently using animated vignettes to determine how schematic models that differ on affect related dimensions, such as affiliation or antagonism, differentially activate frontal and posterior brain regions.

The Barnard and Bowman (submitted) process algebra-based model of the attentional blink implemented two modes of processing meaning with a delay line mechanism for holding active representations. Delay lines, which are essentially a linear arrangement of circuits which pass information states "down the line," have already been proposed as plausible functional correlates for the network structures observed in, for example, cerebellar cortex (Abeles, 1991) and could equally well exist in other cortical areas. In the current reporting period we have completed an initial phase of research (Barnard, report in private circulation) that has worked out how a subsystem of cognitive processes, like those proposed in ICS, could emerge from interactions among neural delay lines. Oscillations in the build up of delay line content offer one means for predicting when evoked electrical potentials should occur in response to changes in salience parameters in tasks such as the attentional blink. The qualitative nature and number of delay lines needed to represent a principal components analysis of the material being processed can also form a basis for predicting patterns of brain activation. This particular outcome of the theoretical work in the current programme forms a key component of our future proposals summarised later under Emotion programme E3.

Project CE4: Interactions of "automatic" affective and "controlled" cognitive processing in the maintenance of affect.

Scientific Direction: Dalgleish (15%), Mathews (35%).
Grant-supported scientists: Ridgeway (2 years, 100%).
Research support: Yiend, (35%).
MRC-supported students: Rafter, (3 years, 100%), Wood (3 years, 100%), Yiend (3 years, 100%).

CE4.1 Attentional bias and anxiety

Individuals vulnerable to anxiety are more attentive to threatening cues than are less anxious individuals (see Mathews & MacLeod, 1994). The methods used in previous studies have not enabled a distinction to be made between attentional engagement (a process in which attention is initially captured by stimuli) and attentional disengagement (in which attention is removed from one object or location and allocated elsewhere). Engagement and disengagement may involve different brain systems, and establishing which is most involved in differential attention to threatening cues could have implications for modifying vulnerability to anxiety (see Project CE7).

Yiend & Mathews (2001a) investigated this issue using a method developed by Posner, in which a single emotionally neutral cue stimulus indicates the location of a target that follows. These so-called valid trials lead to faster detection of the target, while occasional invalid trials (in which the target appears in an alternative location) lead to slowing due to having to relocate attention. When the cues were pictures having threatening or neutral content, anxiety-prone individuals were particularly slow relative to controls in the invalid trials, but only when the pictures were threatening. The distinction between anxiety-prone groups was present when cues were displayed for 500ms but was reduced after 2 seconds. It was concluded that the attentional bias in anxiety involves a difficulty in disengaging attention from threatening cues at early stages of processing.

In a further series Mathews, Fox, Yiend & Calder (submitted) showed that this extends to locations having threat connotations. Faces with neutral or fearful expressions were displayed centrally, with eyes gazing to the right, left, or straight ahead. Target letters appearing randomly to the left or right, were detected more slowly when gaze had been directed to the other location than in the direction of gaze. In highly anxious individuals, target detection was particularly speeded in valid trials, when targets appeared in the direction indicated by a fearful (rather than neutral) gaze, and was particularly slowed in corresponding invalid trials. There was no such distinction due to fearful expression in low anxious individuals. It can be concluded that attending to the direction of another person's gaze is relatively automatic, and that the attention of anxiety-prone (but not low anxious) individuals is more strongly guided by eye gaze when a fearful expression implies the presence of danger. This evidence supports further the view that the attentional processing biases associated with vulnerability to anxiety will increase awareness of potential dangers, and thus may serve to maintain anxious mood (see also CE7).

CE4.2 Attention to threat in patients with life-threatening disease

Work on the prediction of anxiety in breast cancer patients (supported by a grant from the CRC to Mathews) has now been completed. A number of measures obtained immediately prior to diagnosis, including a measure of attention to emotional words, were predictive of distress four months later. These results show that patients who are prone to anxiety and health worry prior to diagnosis will continue to experience relatively greater emotional problems on follow up (Mathews, Ridgeway, Warren & Britton, 2002). These data have implications for predicting which patients may profit from anticipatory counselling following diagnosis of cancer.

CE4.3 Interpretation of emotional ambiguity

With Hirsch at the Institute of Psychiatry (London) Mathews has investigated the extent to which social phobia is associated with "on-line" interpretation of ambiguously threatening social events. Based on earlier work (Hirsch & Mathews, 1997), it was supposed that people without severe social anxiety typically attend to external cues (e.g. others' reactions), and tend to make positive inferences about their own social performance. In contrast, socially phobic individuals appear not to make such on-line inferences, either positive or negative, perhaps because their processing resources are depleted by anxiety and/or diverted to internal cues. Nonetheless, social phobics judge their own social performance harshly, presumably based on their prior expectations and internal feelings rather than external information.

To test this hypothesis, social phobic patients and non-anxious controls were asked to read descriptions of ambiguously threatening social events (interviews) while imagining themselves in those situations (as the interviewee). At unpredictable points in the text they made speeded lexical decisions for words that matched either a possible positive or a negative inference. Relative to an unambiguous control condition, at ambiguous points in the text non-anxious individuals were slower to endorse negative (but not positive) words matching possible inferences. Social phobics were always slower at ambiguous points than when the text was unambiguous. These data are consistent with the view that normal positive inferences are blocked in social phobia, and that this may contribute to the persistence of anxiety even in benign situations (Hirsch & Mathews, 2000).

One reason that social phobics fail to make positive inferences is that they report being distracted by negative images of themselves in social situations. When trained to replace these negative images with more benign versions, social phobics' anxiety decreased, and this improvement was apparent to blind observers of their social performance (Hirsch, Clark, Mathews & Williams, in press). Furthermore, when non-phobic volunteers were trained to hold negative self-images of themselves in mind, they came to resemble social phobics in failing to make normal positive inferences about social situations. It thus appears that an internal focus of attention on negative images both enhances anxiety and blocks protective positive inferences in social phobia.

CE4.4 Inhibitory processes and emotion

Mathews has explored negative priming of emotional and non-emotional words as a possible index of the ability to suppress responses to previously attended emotional words. Such failures of inhibition have previously found to be characteristic of Obsessional Disorders. Negative priming refers to the finding that when two items are presented with only one to be selected for report, it takes longer to name an item that was ignored in the preceding trial. Theoretical accounts of this phenomenon suggest that prior ignoring of the later to-be-selected item is crucial. In a novel procedure originally developed by MacDonald, Joordens, & Seergobin (1999), two words are presented together for a comparative judgement (e.g. which animal is larger?). In the critical trials, the non-selected item in a preceding trial becomes the to-be-selected item in the current trial. MacDonald et al. claimed that prior attention to the non-selected item did not prevent negative priming, but actually enhanced it. Mathews, Mackintosh & Holden (in press) found that, to the contrary, correct matching of ignored repetition and control item content revealed no negative priming at all if the to-be-selected items were previously attended and then rejected. Earlier conclusions from studies using this method, suggesting that Obsessive Compulsive patients show large negative priming effects (implying intact inhibitory processes) are thus invalid.

Mathews and Dalgleish (Wood, Mathews & Dalgleish, 2001) used a different technique to investigate the prediction that emotionally vulnerable individuals have particular difficulty in the inhibition of emotional processing. In a series of experiments, a reading task was performed that required the suppression of emotional or neutral meanings. Participants read sentences, some of which ended in threat/neutral homographs (e.g. Some puddings are made using batter) and were then required to decide if another word was related to the meaning of the sentence (e.g. injury). In these critical sentences, the threatening meaning was irrelevant and so must be rejected. Decisions involving homographs were significantly slowed relative to control sentences. When threatening (rather than neutral) meanings had to be suppressed, anxious participants made more errors and were also sometimes slower than control groups. It was concluded that inhibition of threat meanings may be more difficult in anxiety-prone individuals, consistent with the difficulty they experience in disengaging attention from threatening locations (see also Yiend & Mathews, 2000a).

This difficulty in inhibiting congruent emotional material in individuals prone to emotional problems was further investigated by Dalgleish and collaborators in patients with clinical depression using a directed forgetting (DF) methodology (Power, Dalgleish & Claudio, 2000). The DF paradigm requires individuals to try to forget information that they had previously been instructed to remember. The DF literature shows that successful forgetting in this context involves direct inhibition of the to-be-forgotten material. In this series of studies depressed and control participants were instructed to forget emotional and neutral material. DF was measured relative to a (remember) control condition in which no instruction to forget was given. The data showed that the depressed patients, but not the controls, showed no DF effect for emotional words. This provides further evidence of a difficulty in inhibiting emotional information in vulnerable and clinical groups.

Dalgleish and Mathews (Dalgleish, Mathews & Wood, 1999) outlined a theoretical framework for conceptualising inhibitory processes in cognition-emotion relations that rested on the distinction between automatic and controlled cognitive processing. A two-by-two taxonomy of inhibitory processes was outlined involving automatic and controlled processing of emotional cognitive content and automatic and controlled processing of emotional feelings and behaviours. Dalgleish (in collaboration with Andrews at the University of Essex) investigated this controlled/automatic distinction in a large sample of emergency service workers who were asked to respond to test items describing different putative ways in which they inhibited aspects of affect in relation to a specified traumatic event. Two clear factors emerged reflecting controlled (e.g. Do you make an effort to not talk about what happened?) and automatic (When you go over the event do your emotions feel numb?) inhibition processes. These factors significantly predicted levels of post-traumatic stress in a longitudinal study of a subset of the original sample. The factor structure was confirmed in an additional study of patients with eating disorders who had experienced traumatic life events (Andrews, Joseph, Troop Van Rooyen & Dalgleish, submitted).

Project CE5: Multi-level analyses of emotional disorders

Scientific Direction: Barnard (10%), Dalgleish (60%), Teasdale (20%).
Research support: Golden (80%), Green (10%), Ramponi (10%).
MRC-supported students: Dunn (3 years, 100%).
Grant-supported students: Sheppard (3 yrs,100%), Du Toit (2 yrs, 100%), Bishop (3 yrs, 20%).

Unique features of the work of the Cognition and Emotion group include: (1) integration of basic research on the interaction of cognition and emotion; (2) the development of comprehensive, multi-level, information processing theoretical frameworks on the basis of that research; (3) the development and investigation of analyses of specific emotional disorders within those frameworks; and (4) application of these analyses to understanding and developing psychological treatments for emotional disorders. Work in Project CE5 in the report period has taken forward step 3, multi-level analyses of emotional disorders, in two new directions: (a) inclusion of a developmental perspective; and (b) applications to new disorders.

CE5.1 Developmental aspects of emotional disorder

Multi-level analyses of emotional disorder in adults (e.g. Mathews & Mackintosh, 1998; Power & Dalgleish, 1997; Teasdale & Barnard, 1993) focus on the way that self-perpetuating patterns of information processing, frequently including mood-related cognitive biases, contribute to the onset and maintenance of disorder. For example, increased accessibility of self-devaluative material in depressed mood in individuals vulnerable to major depression is assumed to contribute to the maintenance and escalation of those moods; attentional biases to threat in anxious individuals are assumed to contribute to the development of generalised anxiety disorder. It is assumed that these self-perpetuating patterns are learned, but there has been very little research investigating the cognitive processing of emotional information in child and adolescent emotional disorders to determine when, developmentally, these patterns are established (Vasey, Dalgleish & Silverman, in press). Such research is important because in other areas of developmental psychology, there is evidence that cognitive processing follows a clear developmental course and is markedly different in younger groups relative to adults across a number of paradigms and cognitive domains.

CE5.1.1. Cognitive biases in childhood and adolescent emotional disorders

Dalgleish, with colleagues at the Institute of Psychiatry, carried out a large set of studies to address the following questions: 1) Do processing biases for emotional information exist in younger populations with emotional disorders? 2) If so, to what extent are they similar to such biases in adult samples? 3) What is the developmental course of such processing biases? The research involved 3 clinical samples of children and adolescents with diagnoses of Major Depressive Disorder (MDD), Generalized Anxiety Disorder (GAD), and Posttraumatic Stress Disorder (PTSD). The cognitive tasks were similar to those used with adult disorders and included measures of attention (the dot probe paradigm and the modified Stroop task), recognition and recall memory (for emotional words and trait adjectives), prospective cognition (probability estimates for future negative events), and the interpretation of ambiguity (homograph interpretation).

The most striking finding of these studies was that childhood and adolescent emotional disorders were each characterised by particular biases in processing emotional information that were very similar to the biases shown in corresponding adult disorders, and different from the pattern of biases shown in other disorders (Dalgleish et al., in press). The GAD and PTSD groups exhibited clear attentional and interpretative biases for threat-related material: (Dalgleish, Moradi, Taghavi, Neshat-Doost & Yule, 2001; Moradi, Taghavi, Neshat-Doost, Yule & Dalgleish, 1999a; Taghavi, Dalgleish, Moradi, Neshat-Doost & Yule, in press; Taghavi, Moradi, Neshat-Doost, Yule & Dalgleish, 2000; and Taghavi, Neshat-Doost, Moradi, Yule & Dalgleish, 1999). In contrast, attentional effects were absent in the MDD sample (Neshat-Doost, Moradi, Taghavi, Yule & Dalgleish, 2000; Neshat-Doost, Taghavi, Moradi, Yule & Dalgleish, 1997). Conversely, MDD children and adolescents, and a sub-clinically depressed sample, exhibited memory biases for negative self-encoded material (Neshat-Doost, Taghavi, Moradi, Yule & Dalgleish, 1998; Bishop, Dalgleish & Yule, submitted), but such biases were not shown by the GAD and PTSD children. The broad similarities between the cognitive biases shown by adult and childhood presentations of individual emotional disorders suggest that the disorder-related self-perpetuating processing patterns to which these biases contribute can become established relatively early in development.

The main difference between the findings observed by Dalgleish and colleagues and previously reported findings in adults was that, on the prospective cognition measure, GAD, MDD and PTSD children, unlike their adult counterparts, showed no tendency to estimate negative events as more likely to happen to themselves than to others (Dalgleish, Moradi et al., 2000; Dalgleish, Taghavi et al., 1997). Follow-up studies suggested that the lack of a negative self-referent bias in clinically distressed children and adolescents reflected the operation of strategic inhibition mechanisms that served to protect the young patients from distressing material (Bishop, Dalgleish, Nimmo-Smith & Yule, submitted; Dalgleish et al., 1998; Dalgleish, Wood & Yule, submitted).

The developmental course of the different forms of processing bias in younger clinical groups was investigated using regression techniques. These showed that, within the age range studied (8 to 18), there was no evidence for a developmental progression in the measures of attentional bias, suggesting that these biases are established early and are relatively constant subsequently. In contrast, the strength of the relationship between depressed mood and memory bias in the MDD sample increased significantly with age. This finding is consistent with the hypothesis that such processing is underpinned by negative schematic self-representations. Other developmental evidence would lead us to expect that such representations would become more elaborate over development in ways that would increase self-referent memory biases.
CE5.1.2 Overgeneral autobiographical memory in major depression

In overgeneral autobiographical memory (OGM) individuals, asked to recall memories of specific events, produce, instead, categoric descriptions summarising repeated instances of such events. OGM occurs in major depression, predicts the course of this disorder, has been linked with cognitive avoidance and rumination (CE7), and is assumed to reflect core aspects of depressive psychopathology. All previous studies of OGM have been in adults. Teasdale, in collaboration with Park and Goodyer (Park, Goodyer & Teasdale, 2002) has demonstrated, for the first time, OGM in first episode major depression in adolescents. This finding is important in demonstrating that OGM is not an aftereffect of previous episodes of depression, and that it characterises developmentally early forms of major depressive disorder just as much as adult forms.

The tendency to generate overgeneral autobiographical memories has been associated with reports of childhood abuse in depressed patients (Kuyken & Brewin, 1995). In order to disentangle the relationships of depression and early trauma to OGM, Dalgleish, Tchanturia et al. (submitted) examined OGM in a group of patients with eating disorders for whom the primary diagnosis was not depression, but who, nonetheless, reported moderately high levels of both childhood abuse and depression. These patients generated more OGMs than controls, and the degree of OGM was positively associated with levels of self-reported abuse, even when levels of current depression were controlled for. These data, along with those of Kuyken and Brewin (1995) indicate that perceived early trauma, as well as significant levels of depression, are important contributors to OGM.

CE5.1.3 Cognitive vulnerability to first episode major depression in children and adolescents

Teasdale's (1988) differential activation hypothesis suggests that vulnerability to major depression is related to the kind of negative thinking patterns activated in states of mild depressed mood, rather than to the more trait-like aspects of thinking style emphasised by previous hypotheses. Activation of such patterns by current depressed mood is assumed to reflect a history of association between negative thinking and depressed mood in the past. This hypothesis can account for the observed progressive increase in risk for further episodes of depression with every consecutive episode experienced; each such episode provides further opportunities for associations to be formed between depressed mood and negative thinking. Studies in adults consistently support the differential activation hypothesis: recovered depressed patients (known to be at high risk for depression) show more negative depressogenic thinking than controls when assessed in induced or natural depressed mood, but not in non-depressed mood. It is not possible to tell from such studies whether the tendency for depressogenic thinking to be activated in dysphoria also preceded patients' first episode of major depression, constituting a prior vulnerability, or whether it was primarily a consequence or "scar" of earlier episodes. To address this issue, in work with Goodyer and colleagues (Kelvin, Goodyer, Teasdale & Brechin, 1999), Teasdale examined the thinking patterns of high emotionality children, known to be at risk for depression, who had not yet experienced an episode of major depression. As in adults, in induced depressed mood, high risk children showed more globally negative self-referent thinking than low risk children, but the groups did not differ in neutral mood. These findings demonstrate, for the first time, that the tendency to switch in depressogenic thinking in dysphoria constitutes a prior vulnerability for major depression.

CE5.1.4 Everyday memory in children and adolescents with PTSD

Research in adults with PTSD has shown that they present with impaired memory functioning (e.g. Yehuda et al., 1995) on a range of neuropsychological tasks. However, no such research has been carried out on younger populations. This is important because children and adolescents with PTSD are normatively pursuing their education and any impairments in memory, however short-lived, could have profound long-term consequences. Dalgleish and collaborators therefore examined memory functioning in children and adolescents with PTSD using the Rivermead Behavioural Memory Test, a set of tasks that taps everyday memory processes (Moradi, Neshat-Doost, Taghavi, Yule & Dalgleish, 1999b). Overall, the PTSD youth showed poorer memory performance than the controls. In particular, they were worse on tests of prospective memory where they had to remember an intention to carry out an act. These data suggest that attention needs to be paid to everyday cognitive functioning in young trauma survivors.

CE5.2 Applications to new disorders

CE5.2.1 Seasonal depression

In Major Depressive Disorder with Seasonal Pattern (MDD-SP), or seasonal depression, patients are in episode in the winter and in remission/recovery in the summer. Because variation in this disorder is so closely linked to biologically relevant seasonal rhythms, it is often assumed that these variations may be more directly biologically determined than in more common forms of unipolar depression. For this reason, Dalgleish used the study of seasonal depression as an opportunity to examine the generalisability of previous findings in unipolar depression to a condition that was clearly related, but that might have a different aetiology.

Dalgleish, Spinks, Kuyken and Yiend, (2001) found that, unlike patients with typical major depression, seasonally depressed patients did not differ overall from healthy controls on overgeneral autobiographical memory (OGM) when the 2 groups were compared in the winter, when the seasonally depressed patients were symptomatic. This lack of a significant cross-sectional OGM effect in seasonal depression was surprising given that the performance of seasonally depressed patients on other cognitive measures such as attributional style (Levitan et al., 1998) is similar to that in non-seasonal depression. Further, winter OGM scores, although not significantly elevated in the seasonally depressed, did predict the extent to which these patients recovered on observer-rated symptoms of depression the following summer. It was hypothesised that seasonal depression is not associated with fundamental shifts in elaborative cognitive processing, as indexed by explicit memory tasks, while nevertheless being associated with negative biases on tasks with a large self-report/response bias component, such as questionnaire measures. This idea was examined in a follow-up study (Dalgleish, Spinks, Golden & Du Toit, submitted) in which performance on a different memory measure - recall memory for trait adjectives - and on a self-report measure - the attributional style questionnaire (ASQ) - were assessed in healthy controls and a group of seasonally depressed patients in the winter when they were symptomatic. Again, patients were followed up in the summer and symptoms of depression were assessed by interview. The MDD-SP group differed from controls on the ASQ (in the predicted depressogenic way) but not on the recall measure. However, as with OGM, recall of self-referent negative adjectives predicted later symptom outcome.

These studies suggest (a) that aspects of the mnemonic processing of emotional information in seasonal depression play an important role in the maintenance and recovery from the disorder but (b) that seasonal depression seems to differ from non-seasonal depression in that it is not associated with overall biases in memory for emotional information, as measured by both autobiographical recall and recall of trait adjectives. This may be because the aetiology of seasonal depression is more closely associated with biological factors than with cognitive ones (Dalgleish, Rosen & Marks, 1996).

CE5.2.2 Four sources of variation in dysfunctional ideation and affective states

ICS has previously been applied to the analysis of unipolar major depression (Teasdale & Barnard, 1993; and see CE7 and CE8). Barnard (in press) has extended the ICS analysis of symptom expression to a wider range of cognitive-affective psychopathology. He identified four theoretical sources of variation in the processing of propositional and implicational levels of meaning: Self-models, modes, how they are switched as a function of control parameters (e.g. salience, rate of change), and the consequences of asynchronous exchanges between subsystems. The schizophrenic spectrum of symptoms was related to the consequences of asynchronous exchanges between two levels of meaning. The symptoms of mania were related to high rates of change in implicational representations and to a predominance of experiential buffering at this level of representation, rather than the analytic buffering of propositional representations linked to depressive states. This illustrated how different, yet overlapping constellations of symptoms can arise when specific properties of executive processing move outside their normal range of operation. Some of the evidence leading to or predicted by this analysis was summarised in section CE3.

CE5.2.3 Theoretical modelling of PTSD

Dalgleish (submitted-a) has reviewed the major cognitive theories of PTSD and argued that they represent a paradigm case for the evolution of theorizing in psychopathology from single 'level' associative network or mental schema models of disorder to the types of multi-representational architectures that have been developed in the Cognition & Emotion group - SPAARS (Power & Dalgleish, 1997) and ICS (Teasdale & Barnard, 1993). This paper argues that three core representational systems (schemas, propositional representations, and associative representations) are necessary to provide comprehensive accounts of a range of psychopathology, using PTSD as the paradigm case. However, the cognitive theories of PTSD are called to account for their underspecification of the mechanics of theorising, for example how information is transferred from one type of representation to another, and this is targeted as a major item on the theoretical agenda.

CE5.2.4 Counterfactual reasoning in trauma survivors

A key feature of the clinical presentation of trauma survivors is the experience of intrusive and distressing thoughts about the original traumatic event. Often these thoughts are of a counterfactual (CF) nature whereby individuals try to imagine how the event might have turned out differently if a significant aspect had been different - mental undoing (Dunmore, Clark & Ehlers, 1999). Events can be mutated or undone in an upward direction to provide a better outcome (e.g. "if only I had not gone out that day then the accident would not have happened") or in a downward direction to provide a worse outcome. Mutations can be about self-related or other-related behaviours. Upwardly mutating the event has negative affective consequences as the individual is presented with an imagined way in which the trauma could have been avoided or ameliorated. Despite this emotional cost associated with upward mutations, clinical anecdotal evidence suggests that these are the most common form of CF intrusion in Posttraumatic Stress Disorder (PTSD).

Dalgleish (submitted-b) investigated this issue empirically in trauma survivors with and without PTSD to determine whether upward CFs were indeed normative in the disorder and, if this was the case, to examine what the function of such CFs might be. The first study revealed that trauma survivors did invariably make upward, self-referent CFs about their trauma, even though they had minimal control over the unfolding of the traumatic event at the time. However, the second study showed that this pattern was the same regardless of PTSD status. It was hypothesised that this pattern of CF reasoning might reflect attempts by all trauma survivors to integrate an element of self-control into their representation of the trauma by hypothetically mutating aspects of their behaviour such that the trauma could have been prevented. This cognitive gain might outweigh the negative affective cost of such CFs. A third study investigated whether this CF style transferred to events other than the original trauma and whether it varied with the objective controllability of the event. Trauma survivors and never-traumatised controls generated CFs to high- (car crash) and low-controllability (hurricane) accident scenarios. The results revealed that trauma survivors generated self-referent CFs to both high- and low-controllability scenarios. This was again independent of PTSD status. In contrast, controls showed this pattern for high controllability events, but not for low-controllability events, where the CFs were more likely to be other-referent.

It remained possible that this prevalent, self-referent CF style in trauma survivors was a function of priming of the original trauma memory by the research recruitment process which emphasised the psychological processing of trauma. In order to examine this possibility, a final study involved priming healthy volunteers to reflect on either a neutral life event or a negative, uncontrollable life event, prior to carrying out the scenario task described above. The results revealed that priming with a negative event did not induce a more self-referent CF style. Indeed, it had the opposite effect, indicating that such priming is unlikely to be a significant factor in the pattern of results in the trauma survivors. This set of studies indicates that the experience of a traumatic life event, regardless of whether or not that event leads the victim to develop PTSD, seems to result in a CF reasoning style that emphasises the upward mutation of self-referent behaviour, even for non-autobiographical events where changes in the person's behaviour could not reasonably have made any difference at the time. This is consistent with upward, self-referent CF reasoning having the cognitive function of reinforcing a sense of self-control following the experience of trauma

Project CE6: Neuropsychological deficits in recognition, perception and experience of emotion

Scientific Direction: Calder (40%), Dalgleish (25%), Lawrence (30%).
Research support: Golden (20%), Keane (40%).
MRC-supported students: Croucher (1 yr, 100%).

Project CE2 addressed the perceptual processing (i.e., front-end encoding, configural processing, etc.) of facial expressions. A full understanding of facial expression recognition requires an appreciation of the mechanisms underlying the identification of the emotions displayed (i.e., that a facial expression signals fear, rather than anger or disgust, etc.). Calder has addressed this by studying different patient populations. In earlier research, Calder and collaborators found initial evidence that the recognition of facial signals of fear and disgust may be supported by separate neural systems (Calder, Young, Rowland, Perrett, Hodges, et al, 1996b; Sprengelmeyer, Young, Calder, Karnat, Lange, et al, 1996). Bilateral amygdala lesions were found to impact primarily on the recognition of fear, and to a lesser extent anger, (Calder et al., 1996b) whereas patients with Huntington's disease showed a disproportionate impairment in recognising facial expressions of disgust (Sprengelmeyer et al., 1996). Project CE6 has built on these findings, characterising more completely the functional and neurological dissociation between fear and disgust processing. In addition, work with Lawrence has addressed the neural underpinnings of anger processing. These findings have important theoretical implications and are reviewed in a recent Nature Reviews Neuroscience article by Calder, Lawrence, & Young (2001b). The research addresses a longstanding debate concerning whether the representation of emotion involves individual systems for separate emotions, or an integrated system for all emotions (Calder et al., 2001b). More specifically, the research provides insights into the neural mechanisms underlying the recognition of individual emotions, and highlights parallels between human and comparative research (LeDoux, 2000; Garcia, Forthman Quick, & White, 1983). On the basis of these findings we have proposed that an effective approach to the neuropsychology of human emotion is to use phylogenetic data to guide a search for dissociable emotional systems.

CE6.1 A cross-modal system for recognising fear

Calder's collaborative neuropsychological projects have confirmed the amygdala's role in processing facial expressions of emotion, and in particular fear (Adolphs et al., 1999; Broks et al., 1998). In collaboration with Scott (University College London), Calder has addressed the contribution of the amygdala to the recognition of emotion from vocal cues (Scott et al., 1997) in a case study of DR, a lady with selective bilateral amygdala damage who shows impaired recognition of fear and anger (Calder et al., 1996b) from the face. DR showed an identical pattern in the vocal domain, supporting the view that the amygdala contributes to the recognition of these emotions across different sensory modalities (Calder et al., 2001b). This proposal is also supported by collaborative functional imaging (fMRI) projects with Phillips, and Morris and Dolan, which showed enhanced amygdala signals for facial (Phillips et al., 1998; Phillips et al., 1997; Morris et al., 1998) and vocal (Phillips et al., 1998) signals of fear. These observations lend support to Calder et al.'s (2001b) proposal that fear recognition is achieved by a mechanism that codes emotional information from multiple sensory modalities.

CE6.2 A dissociation between fear recognition and memory for negatively-valenced material

Recent functional imaging and patient-based research has implicated the amygdala in enhanced memory for negatively-valenced material. In collaborative work (Papps, Calder, Young, & O'Carroll, in press), Calder has investigated memory for emotional material in DR (see CE6.1; Calder et al., 1996; Scott, et al., 1997). DR showed intact memory for negatively valenced pictures and words, providing the first evidence that memory and recognition of negative material rely on dissociable systems. These findings were attributed to the fact that DR's amygdala damage is incomplete.

CE6.3 Disgust recognition in OCD and Tourette's syndrome

Calder's collaborative work with Sprengelmeyer has demonstrated that Huntington's disease (an autosomal dominant neurogenetic disorder that, in its early stages, particularly affects the basal ganglia) causes disproportionate impairments in recognising facial expressions of disgust (Sprengelmeyer, 1997). To further investigate the role of the basal ganglia in disgust processing, Sprengelmeyer and Calder (Sprengelmeyer et al., 1997b) examined two psychiatric disorders associated with abnormal metabolic activity in this brain region, obsessive compulsive disorder (OCD) and Gilles de la Tourette syndrome (Braun et al., 1995; Rapoport, 1989; Rapoport & Fiske, 1998). The results showed that the OCD group and a sub-group of the Tourette's participants with co-morbid OCD symptoms showed a selective impairment in recognising disgust facial expressions. One interpretation is that these findings emphasise the role of the basal ganglia in recognising disgust. In addition, it was proposed that the presence of OCD symptoms in the patients' childhood years may have led to a weakened mapping between self-experienced emotion and the facial expressions of others.

CE6.4 Functional imaging studies of disgust processing

Huntington's disease, OCD and Tourette's syndrome are not characterised by focal neuropathology. Hence, although these patient-based studies point towards the probable involvement of the basal ganglia in disgust, the evidence is indirect. In this respect functional imaging research has been particularly informative. Using stimuli developed from program CE2, Calder's collaborative work with Phillips (Phillips et al., 1998; Phillips et al., 1997) has identified two areas involved in processing facial expressions of disgust - the insula and the basal ganglia. Insula involvement is particularly interesting given its role in gustatory function (Augustine, 1996; Small et al., 1999). Of equal relevance is research showing that lesions to the insula or pallidum of rats interfere with conditioned taste aversion (Dunn & Everitt, 1988; Hernadi, Zaradi, Faludi, & Lenard, 1997). Together these findings concur with the theory that disgust has developed from a more primitive system involved in distaste (Rozin & Fallon, 1987; Rozin, Lowery, & Ebert, 1994).

CE6.5 A cross-modal system for recognising disgust

While functional imaging of healthy participants identifies brain regions that are correlated with certain behavioural tasks, it provides no information about whether these regions are necessary and sufficient for successful performance of these tasks. Hence, it is important that Calder has provided further evidence for the role of the insula/basal ganglia regions in processing disgust in the form of a case study of a man (NK) with a focal lesion affecting these areas (Calder, Keane, Manes, Antoun, & Young, 2000b). NK's damage is lateralised to the left and includes the insula, putamen, internal capsule, globus pallidus, and the head of the caudate. NK showed highly selective impairments in recognising disgust from facial and vocal cues, and in his self-reported experience of this emotion. NK's results are consistent with damage to a system that is involved in the recognition of disgust from different modalities, and in the experience of disgust (Calder et al., 2001b).

CE6.6 Differential effects of ageing on the recognition of fear and disgust

In line with the proposal that separate neural systems underlie the recognition of fear and disgust, Calder and Manly have shown differential effects of ageing on the recognition of these emotions (Calder et al., in press). On two tests of facial expression recognition with five age groups ranging from 20 years to 70 years, increasing age produced a progressive reduction in the recognition of fear and, to a lesser extent, anger. In contrast, older participants showed absolutely no reduction in recognition of facial expressions of disgust; rather, there was evidence of an improvement. Recognition of other facial expressions showed no significant evidence of deterioration (or enhancement) across age groups. These results are consistent with the differential effects of ageing on two brain regions underlying the recognition of, respectively, fear and disgust. In relation to fear, research has shown that medial temporal pathology (including the amygdala) is a consequence of normal ageing (Anderton, 1997), while fMRI research has demonstrated reduced amygdala activation to negative facial expressions with increasing age (Iidaka et al., 2001). In contrast, the gross structure and neurochemistry of the pallidum, a region of the basal ganglia implicated in fMRI studies of disgust (Phillips, et al. 1997; Phillips, et al. 1998) and taste aversion (Hernadi et al., 1997), is largely spared by ageing (Raz, 2000).

CE6.7 The contribution of frontal systems to facial expression recognition

The work discussed above identified separate neural mechanisms involved in processing fear (amygdala) and disgust (insula and basal ganglia). Other studies have emphasised the important role of the frontal lobes in processing emotional cues in general, and some have suggested that the systems involved in coding individual emotions may feed into more general emotion systems in frontal cortex (Sprengelmeyer, Rausch, Eysel, & Przuntek, 1998). If this is correct, then we would expect to see general emotion recognition impairments following frontal cortex damage. Keane, Calder, Hodges and Young (2002) investigated this issue in a series of patients with frontal variant frontotemporal dementia (fvFTD), a condition that largely affects the frontal regions of the brain but particularly the ventromedial frontal lobes. Results showed that fvFTD was associated with impaired recognition of a number of emotions from both facial and auditory cues. In contrast, there was no evidence of impaired recognition of identity from faces. These results emphasise a role for the frontal lobes in processing emotional cues from different sensory modalities. In addition, they suggest that previous reports of impaired facial expression recognition in the absence of impaired facial identity recognition, may have been incorrect to interpret this pattern as the antithesis of prosopagnosia (impaired facial identity recognition). Rather, as suggested by the results of the fvFTD study, this pattern may instead reflect impaired recognition of emotion.

CE6.8 Perceptual and motor codes involved in facial expression recognition

It is tempting to think of the perceptual mechanisms underlying facial expression recognition as analogous to those for facial identity. However, we should be cautious in adopting this view because we not only recognise expressions in other people's faces, we generate them ourselves. Hence, the mental representation of facial expressions has the added requirement of a motor-program code (to describe how to produce the expression) in addition to a visual code. The extent to which these two codes interact is unclear. To investigate this issue, Calder studied a group of subjects with a rare congenital disorder that causes facial diplegia (Möbius Syndrome) (Calder, Keane, Cole, Campbell, & Young, 2000a), meaning that they have never produced normal facial expressions. Anecdotal reports had suggested that this group are severely impaired at recognising facial expressions, but until now, there has been no systematic research. Calder found no evidence of marked deficits in facial expression recognition in the Möbius individuals. These findings suggest that there may be minimal interaction between motor-code and visual representations for facial expression.

CE6.9 Selective deficits in anger recognition

Lawrence and Calder have used data from ethology and ethopharmacology to make and test predictions about selective deficits in emotion recognition. In particular, they argued that certain emotions (fear, disgust, anger) can be linked to processing in defense systems involved in detecting and coordinating flexible responses to different ecological threats. For example, appetitive aggression occurs in the context of resource/dominance disputes in a wide variety of species. Hence, the possibility arises that a specific neural system may have evolved to detect and coordinate responses to this specific form of challenge or threat. The dopamine system has been implicated in the processing of signals of aggression in social-agonistic encounters in several species. Acute administration of the dopamine D2-class receptor antagonist sulpiride was used to induce dopaminergic antagonism in healthy male volunteers. This produced a selective disruption in the recognition of facial expressions of anger (signals of appetitive aggression in humans), in the absence of impairment to other emotions (e.g. fear, disgust) or to facial identitiy processing (Lawrence, Calder, McGowan, & Grasby., 2002). These results provide strong support for evolutionary approaches to emotion, but are difficult to reconcile with approaches based on a limited number of dimensions, such as valence and arousal. Such approaches have difficulty explaining, for example, selective deficits in the processing of anger, but not fear.

CE6.10 Mechanism of psycho-surgery for depression

Patients who suffer from severe, chronic depression that is resistant to the standard interventions of pharmacotherapy and psychological therapy have the option of psychosurgery as a treatment-of-last-resort. Dalgleish and Teasdale, in collaboration with colleagues at the Institute of Psychiatry, investigated a group of patients who had undergone Stereotactic Subcaudate Tractotomy (SST) as anti-depressant neurosurgery. Typically SST leads to remission/recovery from depression in 60% of patients. However, it is unclear what the psychological mechanisms associated with this anti-depressant effect are, nor whether there are secondary cognitive/neuropsychological costs associated with the operation.

In order to address these two questions, patients who had recovered from depression following SST (SST-recovered) were compared to patients for whom the operation had not been successful (SST-depressed) on a broad range of cognitive and neuropsychological measures (Dalgleish, Yiend et al., submitted). The results revealed that the SST-recovered patients were insensitive to negative feedback information on a widely used decision making task (Bechara, Damasio, Damasio, & Anderson, 1994) relative to the SST-depressed group. However, there were no other differences between the groups on a broad neuropsychological test battery indexing language, general intelligence, attention, memory and executive functioning. The data were re-examined in a correlational analysis and it was also found that the degree of recovery from depression following SST, as indexed by depression self-report measures, was associated with increased insensitivity to negative information.

In order to investigate which group was performing normally, a sample of matched, never-depressed controls completed the task. The performance of the controls did not differ from that of the SST-depressed group but was significantly different to that of the SST-recovered sample, with the recovered sample showing greater insensitivity to negative feedback relative to the healthy, never-depressed individuals.

It remained possible that the performance of the SST-recovered patients was a function of recovery from depression per se rather than a function of recovery following psychosurgery. In order to examine this possibility, patients who had recovered from depression with medication alone (medication-recovered) also completed the task. Again the results revealed significantly augmented insensitivity to punishment in the SST-recovered group and the medication-recovered participants were no different to the healthy controls and SST-depressed samples.

This pattern of results suggests that SST for depression may work by inducing a relatively enhanced insensitivity to negative information. Such insensitivity would potentially disrupt the vicious cycles of processing of negative information that are seen as central to the maintenance of the disorder in the theoretical models developed in the Cognition and Emotion group (Power & Dalgleish, 1997; Teasdale & Barnard, 1993). The results also indicate that the anti-depressant effect generated by SST does not seem to be at the expense of a more general deficit in neuropsychological functioning as measured by a standard test battery. This is important clinical information for those who are contemplating this irreversible treatment-of-last-resort.

Project CE7: Attention and the manipulation of affect perpetuating processes

Scientific Direction: Mathews (45%), Teasdale (25%).
Research support: Green (75%), Yiend, (45%).
MRC-supported students: Hoppitt (I year, 100%), Potts (3 yrs, 100%).

The manipulation of cognitive processes assumed to produce or maintain emotional states and disorders provides an opportunity both to test multi-level analyses of those disorders, and to identify the key processes to be modified in clinical interventions to treat or prevent those disorders. Studies in both anxiety and depression have investigated the effect of attentional manipulations on affect perpetuating processes.

CE7.1 Threat-related biases

CE7.1.1 Manipulation of attentional bias

This project is concerned with the issue of whether emotional processing biases can be manipulated. Establishing such experimental control would allow us to investigate the mechanisms underlying individual differences in these processing operations, their causal relationship with emotional states, and consequently, whether and how this control can inform therapeutic applications. In a first series of experiments in the area, Mathews addressed specific hypotheses about the critical factors underlying individual differences in attention to mild threat cues such as words or pictures. One hypothesis is that these cues are more aversive for anxiety-prone individuals because they have experienced more unpleasant life events associated with them. A second is that, even if there are no such differences in association frequency, anxiety-prone people learn negative associations more readily. The third is that, even if threat cues are no more aversive for anxiety-prone individuals than for others, they are less able to prevent attention capture by these cues.

In a series of experiments, in collaboration with Mackintosh (and Fulcher), Mathews found evidence consistent with this last hypothesis (Fulcher, Mathews, Mackintosh, & Law, 2001). Neutral pictures were paired with neutral, positive or negative captions, while participants formed an image linking the two. After a delay, the negative-paired pictures were rated as less liked, with no differences between groups high or low in negative emotionality. However, in a task in which participants had to find targets superimposed on the pictures, the group with high negative emotionality were slowed by negative-paired pictures, while the low group was slowed more by positive paired pictures. These data suggest that, even when matched for emotional learning, new negative cues capture attentional resources more in those prone to negative emotions.

In a further study of this kind, Mackintosh & Mathews (in press) used a spatial cueing task to contrast attention to mildly emotionalised cues with that to more severely threatening pictures. In a non-anxious group there was a general tendency for attention to the more mild cues to be actively avoided, although the more severe pictures typically held attention. This provides support for the view that non-anxious individuals can adaptively prevent their attention being captured by mildly emotional cues but this control is overridden in most people as threat severity increases.

CE7.1.2 Evidence that biased processing has causal effects on mood

Although Mathews and colleagues have postulated that attentional and interpretative biases play a central role in the self-perpetuating patterns of information processing in anxiety disorders, until recently there has been no convincing evidence that this relationship was causal. Recently, however, techniques have been developed for inducing emotional processing biases in the laboratory, and have been used to reveal the conditions under which they can cause changes in vulnerability to anxiety.
In an initial series of experiments (Grey & Mathews, 2000) volunteers were presented with threat/neutral homographs as clues to solve word fragments (e.g. batter – in-ury or batter - pa-cake), or in the context of relationship judgements (e.g. Injury - batter. Related?). Participants were randomly allocated to decisions involving either threatening or neutral meanings of homographs, and then tested with new homographs not previously exposed. Both when tested with similar decisions as used in earlier, and with a completely new lexical decision task, these participants showed evidence that their prior experience continued to influence their interpretation of new homograph meanings. Most critically, in current experiments (in collaboration with Colin MacLeod), Mathews has shown that subsequent exposure to an ambiguously threatening event (videos of accidents) leads to different emotional consequences. Those exposed to negative homograph meanings reported increases in state anxiety, while those exposed to benign meanings did not. Post - experimental interviews indicated that participants were unaware of any connection between their prior exposure and later anxiety reactions.

In a parallel series of experiments, Mathews & Mackintosh (2000) have used more complex descriptions of real-life events, having threatening or benign outcomes determined by a final word. In some experiments, volunteers simply read these descriptions while imagining themselves in the situations, while in others they read the same passages up until the final word, which was presented in a to-be-completed fragment form. In all experiments, participants were randomly allocated to threatening or benign outcomes, and then read a new set of descriptions with ambiguous outcomes. Their interpretation of these ambiguous events was established indirectly using a recognition test involving negative and positive versions of the original items. This procedure has revealed a very robust interpretative bias that lasts at least into the following day. Another critical finding was that anxiety changed in a congruent direction when (and only when) training involved participants in the active generation of emotional meanings (e.g. by resolving the final word fragment).

In reviewing these (and other experiments on training attention), Mathews & MacLeod (2002) concluded that experimental induction of processing biases can readily be achieved in the laboratory, without the awareness of participants. The induction procedures themselves do not normally induce emotional changes directly (with some predictable exceptions), but do influence how later events are interpreted. When these interpretations have emotional implications for the individual concerned, they elicit congruent changes in mood. These findings thus provide the first evidence that processing biases can indeed be a cause of anxiety, albeit indirectly by influencing how new events are interpreted.

CE7.2 Modes of self-attention

Teasdale (1999c) drew attention to the apparently anomalous relationship of self-attention to the perpetuation and amelioration of emotional disorders. On the one hand, a tendency to focus attention on the self has been consistently associated with emotional disorders, particularly depression. On the other hand, interventions that involve individuals intentionally focusing attention on aspects of emotional experience lead to reductions in disordered emotion. Using the Interacting Cognitive Subsystems (ICS) theoretical framework, Teasdale suggested that this anomaly could be resolved by recognising the existence of different modes of self-focus. These modes are characterised by distinct configurations of processing resources (see project CE3), and, for that reason, are also characterised by distinct relationships to the perpetuation of emotion. The ICS analysis (Teasdale, 1999c) suggested that perpetuation of depression involves a ruminative mode of self-attention, focused on processing discrepancies between conceptual-level (propositional) representations of current and goal self-states. This analytical mode involves thinking about the self as a conceptual object. Conversely, therapeutic change involves a mode of self-focus controlled at a more schematic (implicational) level. This experiential mode involves less goal-related discrepancy processing. In this mode the subjective self is experienced directly as feelings, thoughts, and body sensations in the moment. Teasdale and colleagues have developed methods to induce experimentally these two modes of self-focus and have demonstrated differences in their effects on aspects of autobiographical memory.

CE7.2.1 Effects on overgeneral autobiographical memory in clinical depression

Measures of overgeneral autobiographical memory (OGM) (see CE5) predict the course of depressive disorders, and, for that reason, OGM has been assumed to reflect processes central to the maintenance of depression. Because OGM is shown by both acutely depressed and recovered patients, and is unrelated to the severity of depressed mood, it has been assumed to be a persistent, trait-like characteristic of depression-prone individuals. However, in collaborative experimental work Teasdale has shown, for the first time, that OGM in both depressed adults and children can be modified by manipulations of attentional focus as brief as 8 minutes (Park, Goodyer & Teasdale, submitted; Watkins, Teasdale & Williams, 2000; Watkins & Teasdale, 2001; Watkins & Teasdale, submitted). These findings suggest that the apparent stability of OGM in the depression-prone reflects its dynamic maintenance by processes that can be disrupted by competing tasks, rather than the trait-like characteristic, previously assumed. Further, these studies have identified the interruption of analytical, ruminative thinking as a feature of interventions that reduce OGM. From their findings, Watkins and Teasdale (2001) proposed that OGM reflects chronic ruminative attempts to analyse and understand current and past difficulties.
These studies of brief attentional manipulations in clinically depressed individuals have also provided an opportunity to examine the effects of analytical versus experiential self-focus on OGM, a variable known to be related to the maintenance of depression. Watkins and Teasdale (2001) induced analytical and experiential self - focus in clinically depressed individuals. The inductions had similar effects on depressed mood, but experiential self-focus significantly reduced the overgenerality of memories, compared to analytical self - focus. In a subsequent study (Watkins & Teasdale, submitted) in which the analytical and experiential inductions used identical items and differed only in the instructional set within which participants processed aspects of self-experience ("think about" versus "experience directly"), the same pattern of results was obtained. These results are consistent with Teasdale's (1999c) proposals that ruminative self-focus acts to perpetuate depression, whereas experiential self-focus can ameliorate it. The latter suggests that more extended training in experiential self-focus could be a valuable clinical intervention to modify depression. This suggestion is developed further in CE8.

CE7.2.2 Effects on the at-oneness of autobiographical memories

In parallel with the above studies of depressed patients, experiments in non-depressed individuals have examined the effects of ruminative analytical versus experiential self-focus on a novel measure of autobiographical memory. This measure is assumed to reflect affect-related discrepancy-based processing. Within Teasdale's (1999c) analysis, such processing is seen as a characteristic feature of ruminative analytical self-focus. It was assumed that the autobiographical memories most accessible at a given time provide a window into the processing configuration active at that time. Ratings of how "at one with things" individuals felt in their most accessible memories were used as a measure of goal-related discrepancy-based processing: low at-oneness was assumed to reflect high discrepancy processing, characteristic of a ruminative, analytic, form of self-focus. Consistent with this assumption, low ratings of at-oneness of autobiographical memories uniquely and substantially predicted high scores on dispositional ruminative self-focus and neuroticism in a sample of 130 normal volunteers (Teasdale & Green, submitted). Three experiments examined the effects of experimentally induced analytical self-focus versus experiential self-focus on the at-oneness of autobiographical memories. Compared with analytical self-focus, experiential self-focus consistently increased the at-oneness of autobiographical memories, while having no differential effect on the happiness or unhappiness of memories, or on mood. Putting the negative correlation of at-oneness with ruminative self-focus, observed by Teasdale and Green, together with the positive effects of experiential self-awareness on at-oneness observed experimentally, these findings support Teasdale's (1999c) suggestion that analytical self-focus and experiential self-focus are distinct forms of self-attention, with different functional properties. Further, the experimental findings in non-depressed individuals provide further evidence that, relative to analytical self-focus, experiential self-focus reduces a process (self-related discrepancy processing) assumed to maintain depression.
In summary, Project CE7.2 supports a distinction between analytical self-focus and experiential self-focus, and shows that experiential self-focus can reduce processes implicated in the perpetuation and escalation of depression. These studies provide an experimental foundation for the training in mindfulness (a form of experiential awareness) that is a central component of the innovative relapse prevention programme described in Project CE8.

Project CE8: Development and evaluation of clinical therapeutic procedures and processes

Scientific Direction: Teasdale (50%).
Research support: Green (10%).
Self-supported students: Ma (3 years, 100%).

Major depressive disorder is a lifelong, recurring condition. It follows that prevention of relapse and recurrence is a centrally important therapeutic challenge. Teasdale, in collaborative work, funded by external grants, has investigated the outcomes and therapeutic mechanisms of two psychological interventions for relapse prevention in recurrent depression. One, cognitive therapy (CT), is an established treatment. The other, mindfulness-based cognitive therapy, is a radically novel approach developed by Teasdale and colleagues on the foundation of prior experimental and theoretical work.

CE8.1 Cognitive therapy and the prevention of relapse and recurrence in residual depression

Depressed patients treated with antidepressant medication who still show residual symptoms of depression are at high risk of relapse. In a two-centre trial (Paykel, Scott, Teasdale, et al., 1999) 158 such patients were randomised either to continue with medication and clinical management, or, additionally, to receive CT. CT significantly reduced relapse by approximately 40%. This result is very important (and this trial was identified as one of MRC's highlighted achievements for the year 1999) as it is the first trial to show, unambiguously, that CT can reduce relapse, compared to continuing pharmacotherapy, in just those patients for whom it is most relevant. Effects of CT on relapse were not accompanied by reductions in background levels of depressive symptoms, suggesting that CT acted through reducing the escalation of symptoms at times of potential relapse, rather than through a generalised reduction in symptom levels (Scott, Teasdale, et al., 2000).

This trial also provided an opportunity to investigate the processes through which CT prevents relapse. CT explicitly aims to reduce belief in the content of negative thoughts and dysfunctional assumptions. It has been widely assumed that these changes in belief mediate the relapse prevention effects of CT. However, there has been a consistent failure to support this view empirically. By contrast, an ICS analysis (e.g.Teasdale, 1997a) suggests that CT works, not through changing belief in thought content, but through changing patients' way of processing, or relationship to, their negative thoughts and feelings. For example, the existence of two levels of meaning in ICS (CE3) suggests that representations at one level of meaning can concurrently be the topic of representations at the other level of meaning. This has particular advantages in analysing the role of metacognitive processing in the maintenance and modification of mood disorders (Teasdale, 1999b). Teasdale's (1997a, 1997b) analysis suggested that increased metacognitive awareness (the ability to see negative thoughts and feelings as events in the mind, rather than identifying with them, or regarding them as necessarily true) is an important aspect of the shift in relationship to negative thinking effected by CT. To test this hypothesis, a memory-based measure of metacognitive awareness was developed and included in the above trial. Consistent with the hypothesis that CT prevented relapse through increasing metacognitive awareness, low metacognitive awareness predicted subsequent relapse, and CT significantly increased metacognitive awareness (Teasdale et al., 2002).
More detailed mediational analyses in this trial (Teasdale et al., 2001) provided further support for the view that CT prevents relapse by changing the way that depressive thinking is processed, rather than by changing belief in its content. First, CT significantly reduced measures of the form of thinking (the extent to which it was absolutistic or "black and white"); second, these measures predicted relapse; and third, the effects of CT on relapse could be accounted for by changes in these measures. By contrast, as in previous studies, measures of belief in the content of negative thinking failed to provide any evidence that changes in belief mediated therapeutic effects. This is the first time that effects of CT in preventing relapse in depression have been shown to be mediated by changes in any cognitive variable. It is particularly important, therefore, that mediation was demonstrated for a measure related to the way that thinking was processed, rather than its content.
This mediational analysis suggested that CT prevents relapse by teaching patients, in depressed mood, to switch out of an habitual ruminative mode of negative thinking into an intentional processing mode in which initial, dysfunctional, "automatic" cognitive products are reappraised using controlled processing resources. In quite separate work (conducted under CE5 but reported here), Sheppard and Teasdale (2000, submitted), using decision latency measures in novel experimental paradigms, independently identified a deficit in such metacognitive monitoring as a feature of acute major depression. Further, they showed that recovery normally involves a transition through a phase in which dysfunctional schematic products are effortfully re-appraised to produce more functional products, before more automatic access to functional schemas is established with full remission. These findings suggest that CT may be effective because it operates through reinforcing the use of a pre-existing meta-cognitive strategy.

In summary, this trial of CT for residual depression provided very important evidence for the clinical efficacy of CT in reducing relapse. It also provided the first peer-reviewed evidence demonstrating the mechanisms through which CT achieves these preventive effects. This evidence supported Teasdale and colleagues' proposals, derived from ICS, that CT works by changing the way in which negative thoughts are processed, rather than through changing belief in thought content. These proposals were the basis from which the novel preventive programme described in CE8 was derived.

CE8.2 Mindfulness-based cognitive therapy for prevention of relapse in recurrent depression

The differential activation analysis of vulnerability to relapse (CE5.1.3) suggests that prevention of relapse and recurrence in major depression requires that patients, particularly those who have experienced multiple episodes, learn how to prevent the escalation of negative, ruminative, thought patterns reactivated by dysphoria. In an analysis of the effectiveness of CT, Teasdale, Segal & Williams (1995) proposed that CT, although explicitly targeted on changing belief in depressive thoughts and assumptions, implicitly teaches patients to prevent escalation of rumination in dysphoria by switching to an intentional processing mode in which negative thoughts and feelings are treated as passing events in the mind, rather than as self or as accurate readouts on reality. The evidence in CE8.1 supports this analysis.

Mindfulness-based cognitive therapy (MBCT) (Segal, Williams, & Teasdale, 2002; Teasdale, 2000) is a radically novel, highly cost-efficient intervention, explicitly designed by Teasdale, Segal and Williams to teach recovered recurrently depressed patients the response prevention skills that the above analysis suggests are necessary, and that are taught, implicitly, by CT. Unlike CT, MBCT does not focus on changing belief in the content of negative thoughts. Instead, it teaches patients skills that allow them to intentionally disengage from negative, ruminative thought patterns reactivated by dysphoric mood at times of potential relapse. MBCT does this by training patients to bring experiential self-awareness (mindfulness) (CE7.2) to their thoughts, feelings and bodily sensations. In parallel, MBCT cultivates metacognitive awareness, the capacity to relate to thoughts and feelings as mental events.

In a three-centre clinical trial, 145 recovered recurrently depressed patients were randomised either to continue with treatment-as-usual or, additionally, to participate in the MBCT programme (Teasdale et al., 2000). For patients with 3 or more previous episodes of depression (77% of sample), MBCT significantly reduced relapse from 66% to 37%. For patients with only two episodes (23% of sample) MBCT was of no benefit. MBCT also increased metacognitive awareness (Teasdale et al., 2002), and reduced overgeneral autobiographical memory (Williams, Teasdale, Segal & Soulsby, 2000). This trial was the first demonstration that a group psychological intervention, administered in recovery, could significantly reduce relapse and recurrence in major depression. It was also the first multi-centre trial of a mindfulness-based clinical intervention.

A subsequent clinical trial (Ma & Teasdale, submitted) with the same design, replicated the finding that MBCT is a highly cost-efficient approach to preventing relapse in patients with 3 or more episodes: relapses were reduced from 78% to 36% for an average therapist contact of only 2 to 3 hours per patient. Again, patients with only two previous episodes failed to benefit from MBCT. This trial provided evidence to support the hypotheses (Teasdale et al., 2000) 1) that MBCT is specifically effective in reducing relapses triggered by dysphoria reactivating autonomous ruminative thought patterns, rather than those provoked by major life events, and 2) that this is the reason why MBCT is ineffective with patients with only 2 episodes; for these patients, relapse was primarily provoked by major life events. MBCT was also found to reduce measures of rumination, thought suppression, and salivary cortisol, compared to the control condition. A therapist manual describing the development, delivery, and evaluation of MBCT has been published (Segal, Williams, & Teasdale, 2002), and a scale to assess therapist adherence to MBCT has been developed and validated (Segal, Teasdale, Williams & Gemar, 2002).

MBCT has attracted wide attention internationally. For example, in 2001, Teasdale was invited by the US National Institute of Mental Health, as the only non-North American participant, to make a presentation on MBCT to an invited working group reviewing approaches to relapse prevention in major depression. In contrast to conventional CT, MBCT makes no attempt to teach patients to address belief in the content of negative thoughts. Instead, it trains patients to access a mode of intentional cognitive processing in which negative thoughts, feelings, and body sensations are regarded as passing objects of attention. The success of this programme in preventing relapse in recurrent depression supports this shift in therapeutic strategy, and opens up new directions for the development of therapeutic procedures that are complementary to existing cognitive therapy approaches across a wide range of disorders. This success also supports the general strategy, inherent in the Cognition and Emotion group, of basing therapeutic interventions on empirically grounded theoretical analyses of emotional disorders and their treatment.

AWARDS AND HONOURS

Dr P. Barnard was appointed to a Visiting Professorship in the Departments of Psychology and Computer Science, University College London, from 2002 to 2007. Dr T. Dalgleish received the May Davidson Award from the British Psychological Society in 1999/2000 for early career contribution to clinical psychology. He also became an Honorary Senior Lecturer in psychology at the Institute of Psychiatry in 1999, and holds a non-stipendiary research fellowship, at Clare Hall College, Cambridge from 1999 to 2002. Dr A. Mathews was appointed to a Visiting Professorship at the Institute of Psychiatry from 1997 onwards. Dr J. Teasdale was elected Fellow of both the British Academy and the Academy of Medical Sciences in 2000.

PUBLICATIONS
(Note: Some papers are included that were published prior to 1998. These represent relevant work that appeared during the interval between the last formal Unit Review and the current reporting period.)
Refereed Journals
Adolphs, R., Tranel, D., Hamann, S., Young, A.W., CALDER, A.J., Phelps, E.A., Anderson, A., Lee, G.P., & Damasio, A.R. (1999). Recognition of facial emotion in nine individuals with bilateral amygdala damage. Neuropsychologia, 37, 1111-1117.
Andrews, L., Joseph, S., Troop, N., Van Rooyen, T. & DALGLEISH, T. (submitted). The structure of avoidance following trauma: Preliminary development of the Posttraumatic Avoidance Acale (PAS).
BARNARD, P. J. & Bowman, H. (submitted). Rendering information processing models of cognition and affect computationally explicit: Distributed executive control and the deployment of attention.
BARNARD, P. & May, J. (1999). Representing cognitive activity in complex tasks. Human-Computer Interaction, 14, 93-158.
BARNARD, P., May, J., Duke, D. & Duce, D. (2000) Systems, interactions and macrotheory. ACM Transactions on Human-Computer Interaction, 7, 222-262.
BARNARD, P. J., Palmer, A., Scott, S., & Knightley, W. (submittted). The immediate processing of schema discrepant meaning in Bipolar Disorder.
BARNARD, P, Scott, S.K., & May, J. (2001). When the central executive lets us down: schema attention and load in a generative working memory task. Memory, 9, 209-221.
BARNARD, P., Scott, S.K., Taylor, J., May, J, & Knightley, W. (submittted). Paying attention to meaning.
BATTYE, G., BARNARD, P.J. & Page M. (submitted). Connectionist models of the Emotional Stroop Effect.
BISHOP, S., DALGLEISH, T. NIMMO-SMITH, I., & Yule, W. (submitted). Anxiety, judgement and availability: a similar story for children and adults?
BISHOP, S., DALGLEISH, T., & Yule, W. (submitted). Memory for emotional material in depressed children.
BISHOP, S., DUNCAN, J., BRETT, M., & LAWRENCE, A.D. (submitted). Prefrontal cortical function and anxiety.
Bramham, J. & DALGLEISH, T. (2000). Assessment of dissimulation of reading ability using the modified Stroop task: A case report. Clinical Psychology and Psychotherapy, 7, 411-414.
Broks, P., Young, A.W., Maratos, W., Coffey, P.J., CALDER. A.J., Isaac, C., Mayes, A.R., Hodges, J.R., Montaldi, D., Cezayirli, E., Roberts, N., & Hadley, D. (1998). Face processing impairments after encephalitis: amygdala damage and recognition of fear. Neuropsychologia, 36, 59-70.
Cagnin, A., Myers, R., Gunn, R.N., LAWRENCE, A.D., Stevens, T., Kreutzberg, G.W., Jones, T., & Banati, R.B. (2001). In vivo visualization of activated glia by [11C] (R)-PK11195-PET following herpes encephalitis reveals projected neuronal damage beyond the primary focal lesion. Brain, 124, 2014-2027.
CALDER, A.J., Burton, A.M., Miller, P., Young, A.W., & Akamatsu, S. (2001a). A principal component analsysis of facial expressions. Vision Research, 41, 1179-1208.
CALDER, A.J., KEANE, J., Cole, J., Campbell, R., & Young, A.W. (2000a). Facial expression recognition by people with Möbius Syndrome. Cognitive Neuropsychology: Special Issue on Face Processing, 17(1/2/3), 73-88.
CALDER, A.J., KEANE, J., Manes, F., Antoun, N., & Young, A.W. (2000b). Impaired recognition and experience of disgust following brain injury. Nature Neuroscience, 3, 1077-1078.
CALDER, A.J., KEANE, J., MANLY, T., Sprengelmeyer, R., Scott, S.K., NIMMO-SMITH, I., & Young, A.W. (in press). The effect of ageing on facial expression recognition. Neuropsychologia.
CALDER, A.J., LAWRENCE, D., KEANE, J., Scott, S.K., Owen, A.I., Christoffels, I., & Young, A.W. (2002). Reading the mind from eye gaze. Neuropsychologia, 40, 1129-1138.
CALDER, A.J., LAWRENCE, A., & Young, A.W. (2001b). The neuropsychology of fear and loathing. Nature Reviews Neuroscience, 2, 352-363.
CALDER, A.J., Rowland, D., Young, A.W., NIMMO-SMITH, I., KEANE, J., & Perrett, D.I. (2000). Caricaturing facial expressions. Cognition, 76, 105-146.
CALDER, A.J., Young, A.W., KEANE, J., & Dean, M. (2000d). Configural information in facial expression perception. Journal of Experimental Psychology: Human Perception and Performance, 26, 527-551.
CALDER, A. J., Young, A. W., Perrett, D. I., Etcoff, N. L., & Rowland, D. (1996). Categorical perception of morphed facial expressions. Visual Cognition, 3, 81-117.
CALDER, A.J., Young, A.W., Rowland, D., & Perrett, D.I. (1997). Computer-enhanced emotion in facial expressions. Proceedings of the Royal Society of London. Series B., B264, 919-925.
CALDER, A.J., Young, A.W., Rowland, D., Perrett, D.I., HODGES, J.R., & Etcoff, N.L. (1996). Facial Emotion Recognition after Bilateral Amygdala Damage: Differentially Severe Impairment of Fear. Cognitive Neuropsychology, 13, 699-745.
Constans, J., MATHEWS, A., Brantley, P., & James, T. (1999). Attentional reactions to an MI: the impact of mood state, worry and coping style. Journal of Psychosomatic Research, 46, 415-423.
COX, S.M.L., Stefanova, E., JOHNSRUDE, I.S, Robbins, T.W, & OWEN, A,M. (2001). Preference formation and working memory in Parkinson's disease and normal ageing. Neuropsychologia, 40, 317-326.
DALGLEISH, T. (1997). An anti-anti-essentialist theory of emotion: A reply to Kupperman. Philosophical Psychology, 10, 85-90.
DALGLEISH, T. (1997). Once more with feeling: The role of emotion in self-deception. Behavioural and Brain Sciences, 20, 110-111.
DALGLEISH, T. (2000). Roads not taken: the case for multiple functional-level routes to emotion. Behavioural and Brain Sciences, 23, 196-197.
DALGLEISH, T. (submitted a). Cognitive theories of Posttraumatic Stress Disorder (PTSD): A paradigm case for the evolution of multi-representational theorizing in psychopathology.
DALGLEISH, T. (submitted b). What might not have been: Counterfactual reasoning in survivors of trauma.
DALGLEISH, T., Joseph, S., Williams, R., & Yule, W. (2000). The Herald of Free Enterprise disaster: Lessons from the first 10 years. Behaviour Modification, 697-723.
DALGLEISH, T., Moradi, A., Neshat-Doost, H., Taghavi, R., Yule, W., & Canterbury, R. (2000). Judgements about emotional events in children and adolescents with Posttraumatic Stress Disorder. Journal of Child Psychology and Psychiatry, 41, 981-988.
DALGLEISH, T., Moradi, A., Taghavi, M., Neshat-Doost, H.T., Yule, W. (2001). An experimental investigation of hypervigilance for threat in children and adolescents with Post-Traumatic Stress Disorder. Psychological Medicine, 31, 541-547.
DALGLEISH, T., Neshat-Doost, H., Taghavi, R., Moradi, A., & Yule, W. (1998). Information processing in recovered depressed children and adolescents. Journal of Child Psychology and Psychiatry, 39, 1031-1036.
DALGLEISH, T. & Power, M.J. (submitted). The I of the storm: The role of the self in conscious emotion experience.
DALGLEISH, T., Rosen, K., & Marks, M. (1996). Rhythm and blues: The assessment and treatment of seasonal affective disorder. British Journal of Clinical Psychology, 35, 163-182
DALGLEISH, T., Spinks, H., Golden, A., & Du Toit, P. (submitted). Cognitive processing of emotional information in seasonal depression: A longitudinal study.
DALGLEISH, T., Spinks, H., Kuyken, W., & YIEND, J. (2001). Autobiographical memory style and future symptom remission: An investigation of seasonal affective disorder. Journal of Abnormal Psychology, 110, 335-340.
DALGLEISH, T., Taghavi, R., Neshat-Doost, H., Moradi, A., Canterbury, R., & Yule, W. (in press). Patterns of processing bias for emotional information across clinical disorders: An investigation of attention, memory and prospective cognition in children and adolescents with depression, generalized anxiety and Posttraumatic Stress Disorder (PTSD). Journal of Clinical Child and Adolescent Psychology.
DALGLEISH, T., Taghavi, R., Neshat-Doost, H., Moradi, A., Yule, W., & Canterbury, R. (1997). Information processing in clinically depressed and anxious children and adolescents. Journal of Child Psychology and Psychiatry, 38, 535-541.
DALGLEISH, T., Tchanturia, K., Serpell, L., Hems, S., de Silva, P.. & Treasure, J. (2001). Perceived control of events in the world in eating disorders: A preliminary study. Personality and Individual Differences, 31, 453-460.
DALGLEISH, T., Tchanturia, K., YIEND, J., Serpell, L., Hems, S., De Silva, P., & Treasure, J. (submitted). Parental style and autobiographical memory in patients with eating disorders.
DALGLEISH, T., Wood, J., & Yule, W. (submitted). Prospective cognitive processing of high- and low-cost negative events in clinical anxiety.
DALGLEISH, T., YIEND, J., Bramham, J., Malhi, G., Teasdale, J., Ogilvie, A., Checkley, S., & Howard, R. (submitted). Altered decision-making following sub-caudate lesions underlies recovery from major depression.
Diefenbach, G., McCarthy-Larzelere, M., Williamson, D., MATHEWS, A., Manguno-Mire, G., & Bentz, B. (2001). Anxiety, depression, and the content of worries. Depression and Anxiety, 14, 247-250.
Duke, D.J., BARNARD, P.J., Duce, D.A. & May, J. (1998). Syndetic modelling. Human Computer Interaction. Vol 13, 337-393.
Fulcher, E., MATHEWS, A., MACKINTOSH, B., & Law, S. (2001). Evaluative learning and allocation of attention to emotional stimuli. Cognitive Therapy & Research, 25, 261-280.
Gardiner, J.M., RAMPONI, C., & Richardson-Klavehn, A. (1999). Response deadline and subjective awareness in recognition memory. Consciousness and Cognition, 8, 484-496.
Gardiner, J.M., RAMPONI, C., & Richardson-Klavehn, A. (2002). Recognition memory and decision processes: A meta-analysis of remember, know and guess responses. Memory, 10, 83-98.
Goerendt, I.K., LAWRENCE, A.D., Brooks D.J. (submitted). Reward processing in health and Parkinson's disease: neural organization and reorganization.
Goerendt, I.K., Messa C., LAWRENCE, A.D., Grasby, P.M., Piccini P., & Brooks D.J. (in press) . Endogenous dopamine release during sequential movement in healthy volunteers and patients with Parkinson's disease: a PET study. Brain.
Grey, S., & MATHEWS, A. (2000). Effects of training on interpretation of emotional ambiguity. Quarterly Journal of Experimental Psychology, 53, 1143-1162.
Grey, S., Price, G., & MATHEWS, A. (2000). Reduction of anxiety during MR imaging: a controlled trial. Magnetic Resonance Imaging, 18, 351-355.
Hertel, P., MATHEWS, A., Peterson, S., & Kintner, K. (in press). Transfer of training emotionally biased interpretations. Applied Cognitive Psychology.
Hirsch, C., Clark, D.M., MATHEWS, A., & Williams, R (in press). Self-images play a causal role in social phobia. Behaviour Research & Therapy.
Hirsch, C., & MATHEWS, A. (1997). Interpretative inferences when reading about emotional events. Behaviour Research & Therapy, 35, 1123-1132.
Hirsch, C., & MATHEWS, A. (2000). Impaired positive inferential bias in social phobia. Journal of Abnormal Psychology, 109, 705-712.
Hutton, S.B., MURPHY, F.C., Joyce, E.M., Rogers, R.D., Cuthbert, I., Barnes, T.R.E., McKenna, P.J., Sahakian, B.J., & Robbins, T.W. (2002). Decision making deficits in patients with first-episode and chronic schizophrenia. Schizophrenia Research, 55, 249-259.
JOHNSRUDE, I. S., OWEN, A. M., White, N. M., Zhao, W. V., & Bohbot, V. (2000). Impaired preference conditioning after anterior temporal-lobe resection in humans. Journal of Neuroscience, 20, 2649-56.
JOHNSRUDE, I. S., OWEN, A. M., Zhao, W. V. & White, N. M. (1999). Conditioned pattern preference without awareness in humans: A novel experimental approach. Learning and Motivation, 30, 250-264.
Joseph, S., DALGLEISH, T., Thrasher, S., Yule, W., & Williams, R. (1997). Impulsivity in survivors of a major disaster. Personality and Individual Differences, 22, 279-281.
Joseph, S., DALGLEISH, T., Williams, R., Thrasher, S., & Yule, W, (1997). Attitudes towards emotional expression and post-traumatic stress disorder at five years following the Herald of Free Enterprise disaster. British Journal of Clinical Psychology, 133-138.
KEANE, J., CALDER, A. J., HODGES, J.R., & Young, A.W. (2002). Face and emotion processing in frontal variant frontotemporal dementia. Neuropsychologia, 40, 655-665.
Kelvin, R.G., Goodyer, I.M., TEASDALE, J.D., & Brechin, D. (1999). Latent negative self schema and high emotionality in well adolescents at risk for psychopathology. Journal of Child Psychology and Psychiatry, 40, 959-968.
Kumari, V., Mitterschiffthaler, M. T., TEASDALE, J.D., Malhi, G.S., Brown, R.G., Giampietro, V., Brammer, M.J., Poon, L., Simmons, A., Williams, S.C.R., Checkley, S.A., & Sharma, T. (in press). Functional MRI of cognitive generation of affect in depression. Biological Psychiatry.
LAWRENCE, A.D. (2000). Error correction and the basal ganglia: similar computations for action, cognition and emotion? Trends in Cognitive Science, 4, 365-367.
LAWRENCE, A.D., CALDER, A.J., McGowan, S.M., & Grasby, P.M. (2002). Selective disruption of the recognition of facial expressions of anger. NeuroReport, 13, 881-4.
LAWRENCE, A.D., Dowson, J., Foxall, G.L., Summerfield, R., Robbins, T.W., & Sahakian, B.J. (in press). Impaired visual discrimination learning in anorexia nervosa. Appetite.
LAWRENCE, A.D., Watkins L.H.A., Sahakian, B.J., HODGES, J.R., & Robbins, T.W. (2000). Visual object and visuospatial cognition in Huntington's disease: implications for information processing in corticostriatal circuits. Brain, 123, 1349-1364.
MA, S.H. & TEASDALE, J.D. (submitted). Mindfulness-based cognitive therapy for depression: Replication and exploration of differential relapse prevention effects.
MACKINTOSH, B., & MATHEWS, A. (in press). Don't look now: Attentional avoidance of emotionally-valenced cues. Cognition and Emotion.
MACKINTOSH, B., MATHEWS, A., & Holden, E. (in press). Bigger than a breadbox? Attention to distracters may not enhance negative priming. Journal of Experimental Psychology: Human Perception and Performance.
MATHEWS, A., Fox, E., YIEND, J., & CALDER, A. (submitted). The face of fear: effects of eye gaze and emotion on attentional engagement.
MATHEWS, A., & MACKINTOSH, B. (1998). A cognitive model of selective processing in anxiety. Cognitive Therapy and Research, 22, 539-560.
MATHEWS, A., & MACKINTOSH, B. (2000). Induced emotional interpretation bias and anxiety. Journal of Abnormal Psychology, 109, 602-615.
MATHEWS, A., & MACKINTOSH, B. (submitted). Take a closer look: effects of emotion on memory for scenes.
MATHEWS, A., MACKINTOSH, B., & Fulcher, E. (1997). Cognitive biases in anxiety and attention to threat. Trends in Cognitive Science, 1, 340-345.
MATHEWS, A,. & MacLeod, C. (2002). Induced processing biases have causal effects on anxiety. Cognition and Emotion, 16, 310-315.
MATHEWS, A., RIDGEWAY, V., Warren, R., & Britton, P. (2002). Predicting worry following a diagnosis of breast cancer. Psycho-Oncology, 10, 1-4.
May, J., Dean, M, & BARNARD, P. (submitted). Using film cutting techniques in interface design.
Moradi, A., Neshat-Doost, H., Taghavi, R., Yule, W., & DALGLEISH, T. (1999). Everyday memory performance deficits in children and adolescents with PTSD: Performance on the Rivermead Behavioural Memory Test. Journal of Child Psychology and Psychiatry, 40, 357-362.
Moradi, A., Neshat-Doost, H., Taghavi, R., Yule, W., & DALGLEISH, T. (1999). Performance of children of adults with PTSD on the Stroop color-naming task. Journal of Traumatic Stress, 12, 663-672.
Moradi, A., Taghavi, M.R., Neshat-Doost, H., Yule, W., & DALGLEISH, T. (1999). The performance of children and adolescents with PTSD on the Stroop colour naming task. Psychological Medicine, 29, 415-419.
Moradi, A., Taghavi, R., Neshat-Doost, H., Yule, W., & DALGLEISH, T. (2000). Memory bias for emotional information in children and adolescents with PTSD: A preliminary study. Journal of Anxiety Disorders, 14, 521-534.
Morris, J.S., Friston, K.J., Buchel, C., Frith, C.D., Young, A.W., CALDER, A.J., & Dolan, R.J. (1998). A neuromodulatory role for the human amygdala in processing emotional facial expressions. Brain, 121, 47-57.
MURPHY, F.C., NIMMO-SMITH, I., & LAWRENCE, A.D. (submitted). Maps of emotion space: a view from functional neuroimaging.
MURPHY, F.C., Michael, A., Robbins, T.W., Sahakian, B.J. (in press). Cognitive impairment in patients with major depressive disorder: the effects of feedback on task performance, Psychological Medicine.
MURPHY, F.C., & Sahakian, B.J. (2001). Neuropsychology of bipolar disorder. British Journal of Psychiatry, 178 (suppl. 41), 120-127.
MURPHY, F.C., Sahakian, B.J., Rubinsztein, J.S., Michael, A., Rogers, R.D., Robbins, T.W., & Paykel, E.S. (2001). Decision-making cognition in mania and depression. Psychological Medicine, 31, 679-693.
MURPHY, F.C., Smith, K.A., Cowen, P.J., Robbins, T.W., & Sahakian, B.J. (2002) The effects of tryptophan depletion on cognitive and affective processing in healthy volunteers. Psychopharmacology, 163, 42-53.
Neshat-Doost, H., Moradi, A., Taghavi, R., Yule, W., & DALGLEISH, T. (1999). The development of a corpus of emotional words produced by children and adolescents. Personality and Individual Differences, 27, 433-451.
Neshat-Doost, H., Moradi, A., Taghavi, R., Yule, W., & DALGLEISH, T. (2000). Lack of attentional bias for emotional information in clinically depressed children and adolescents on the dot-probe task. Journal of Child Psychology and Psychiatry, 41, 363-368.
Neshat-Doost, H., Taghavi, R., Moradi, A., Yule, W., & DALGLEISH, T. (1998). Memory for emotional trait adjectives in clinically depressed youth. Journal of Abnormal Psychology, 107, 642-650.
Neshat-Doost, N., Taghavi, R., Moradi, Yule, W., DALGLEISH, T. (1997). The performance of clinically depressed children and adolescents on the modified Stroop paradigm. Personality and Individual Differences, 23, 753-759.
Nunn, J., MATHEWS, A., & Trower, P. (1997). Selective processing of concern-related information in depression. British Journal of Clinical Psychology, 36, 489-503.
Papps, B.P., CALDER, A.J., Young, A.W., & O'Carroll, R. (in press) Dissociation of affective modulation of recollective and perceptual experience following amygdala damage. Journal of Neurology, Neurosurgery, and Psychiatry.
Park, R.J., Goodyer, I.M., & TEASDALE, J.D. (2002). Overgeneral autobiographical memory in adolescents with depressive disorder. Psychological Medicine, 32, 267-276.
Park, R.J., Goodyer, I.M., & TEASDALE, J.D. (submitted). Effects of induced rumination and distraction on mood and overgeneral autobiographical memory in adolescent major depressive disorder and controls.
Paykel, E.S., Scott, J., TEASDALE, J.D., Johnson, A.L., Garland, A., Moore, R., Jenaway, A., Cornwall, P.L., Hayhurst, H., Abbott, R., & Pope, M. (1999) Prevention of relapse in residual depression by cognitive therapy: A controlled trial. Archives of General Psychiatry, 56, 829-835.
Phillips, M.L., Young, A.W., Scott, S.K., CALDER, A.J., Andrew, C., Giampietro, V., Williams, S.C.R., Bullmore, E.T., Brammer, M., & Gray, J. A. (1998). Neural responses to facial and vocal expressions of fear and disgust. Proceedings of the Royal Society of London. Series B-Biological Sciences, 265, 1809-1817.
Phillips, M.L., Young, A.W., Senior, C., Brammer, M., Andrew, C., CALDER, A.J., Bullmore, E.T., Perrett, D.I., Rowland, D., Williams, S.C.R., Gray, J.A., & David, A.S. (1997). A specific neural substrate for perceiving facial expressions of disgust. Nature, 389, 495-498.
Power, M.J., & DALGLEISH, T. (1998). Cognition in the context of emotion: The case of depression. Analise Psicologica, 381-413.
Power, M.J., & DALGLEISH, T. (1999). Two routes to emotion: Some implications of multi-level theories of emotion for therapeutic practice. Cognitive and Behavioural Psychotherapy, 27, 129-142.
Power, M.J., DALGLEISH, T., & Claudio, V. (2000). The directed forgetting task: Application to emotionally valent material. Journal of Affective Disorders, 57, 147-157.
Rabiner, E.A., Messa, C., Sargent, P.A., Husted-Kjaer, K., Montgomery, A., LAWRENCE, A.D., Bench, C.J., Gunn, R.N., Cowen, P., & Grasby, P.M. (2002a). A database of normal [11C]WAY-100635 binding to 5HT1a receptors in normal male volunteers: normative data and relationship to methodological, demographic, physiological, and behavioral variables. NeuroImage, 15, 620-632.
RAMPONI, C. & BARNARD, P. (submitted). Recollection deficits in dysphoric mood: an effect of schematic models and executive mode?.
Scott, S.K., BARNARD, P.J. & May, J. (2001). Specifying executive representations and processes in number generation tasks. Quarterly Journal of Experimental Psychology. 54A. 641-664.
Scott, J., Palmer, S., Paykel, E.S., TEASDALE, J.D. & Hayhurst, H. (submitted). Is prevention of relapse in depression by cognitive therapy cost-effective?
Scott, J., TEASDALE, J.D., Paykel, E.S., Johnson, A.L., Abbott, R., Hayhurst, H., Moore, R.G., & Garland, A. (2000). Effects of cognitive therapy on psychological symptoms and social functioning in residual depression. British Journal of Psychiatry, 177, 440-446.
Scott, S.K., Young, A.W., CALDER, A.J., Hellawell, D.J., Aggleton, J.P., & Johnson, M. (1997). Impaired auditory recognition of fear and anger following bilateral amygdala lesions. Nature, 385, 254-257.
Segal, Z.V., TEASDALE, J.D., Williams, J.M.G., & Gemar, M.C. (2002). The mindfulness-based cognitive therapy adherence scale: Inter-rater reliability, adherence to Protocol, and treatment distinctiveness. Clinical Psychology and Psychotherapy, 9, 131-138.
Serpell, L., Treasure, J., TEASDALE, J.D., & Sullivan, V.(1999). Anorexia friend or foe: A qualitative analysis of the themes expressed in letters written by anorexia nervosa patients. International Journal of Eating Disorders, 25, 177-186.
SHEPPARD, L.C., & TEASDALE, J.D. (2000). Dysfunctional thinking in major depressive disorder: A deficit in metacognitive monitoring? Journal of Abnormal Psychology, 109, 768-776.
SHEPPARD, L.C., & TEASDALE, J.D. (submitted). How does dysfunctional thinking decrease during recovery from major depression?
Spinks, H., & DALGLEISH, T. (2001). Attentional processing and levels of symptomatology in SAD: A preliminary longitudinal study. Journal of Affective Disorders, 62, 229-232.
Sprengelmeyer, R., Young, A.W., Karnat, A., CALDER, A.J., Lange, H., Rowland, D., & Perrett, D.I. (1997a). Recognition of facial expressions of basic emotions in Huntington's disease. Cognitive Neuropsychology, 14, 839-879.
Sprengelmeyer, R., Young, A.W., Pundt, I., Sprengelmeyer, A., CALDER, A.J., Berrios, G., Winkel, R., Vollmoeller, W., Kuhn, W., Sartory, G., & Przuntek, H. (1997b). Disgust implicated in obsessive-compulsive disorder. Proceedings of the Royal Society of London. Series B-Biological Sciences, 264, 1767-1773.
Taghavi, R., Moradi, A., Neshat-Doost, H., Yule, W., & DALGLEISH, T. (2000). The interpretation of ambiguous emotional information in clinically anxious children and adolescents. Cognition and Emotion, 14, 809-822.
Taghavi, R., Neshat-Doost, H.T., Moradi, A.R., Yule, W., & DALGLEISH, T. (1999). Biases in visual attention in children and adolescents with clinical anxiety and mixed depression-anxiety disorder. Journal of Abnormal Child Psychology, 27, 215-223.
Taghavi, R., DALGLEISH, T., Neshat-Doost, H.T., Moradi, A.R., & Yule, W. (in press). Performance on the Stroop colour naming task in clinically anxious children and adolescents. British Journal of Clinical Psychology.
TEASDALE, J.D. (1999b). Metacognition, mindfulness, and the modification of mood disorders. Clinical Psychology and Psychotherapy, 6, 146-155.
TEASDALE, J.D. (1999c). Emotional processing, three modes of mind, and the prevention of relapse in depression. Behaviour Research and Therapy, 37, S53-S77.
TEASDALE, J.D., & COX, S.G. (2001). Dysphoria: Self-devaluative and affective components in recovered depressed patients and never depressed controls. Psychological Medicine, 31, 1311-1316.
TEASDALE, J.D., & GREEN, H.A.C. (submitted). Ruminative self-focus and autobiographical memory.
TEASDALE, J.D., Howard, R., COX, S.G., Ha, Y., Brammer, M.J., Williams, S.C.R., & Checkley, S.A. (1999). Functional MRI study of the cognitive generation of affect. American Journal of Psychiatry, 156, 209-215.
TEASDALE, J.D., Lloyd, C-A., & Hutton, J.M. (1998). Depressive thinking and dysfunctional schematic mental models. British Journal of Clinical Psychology, 37, 247-258.
TEASDALE, J.D., Moore, R.G., Hayhurst, H., Pope, M., Williams, S., & Segal, Z.V. (2002). Metacognitive awareness and the prevention of relapse in depression: Empirical evidence. Journal of Consulting and Clinical Psychology, 70, 275-287.
TEASDALE, J.D., Scott, J., Moore, R.G., Hayhurst, H., Pope, M., & Paykel, E.S. (2001). How does cognitive therapy prevent relapse in residual depression? Evidence from a controlled trial. Journal of Consulting and Clinical Psychology, 69, 347-357.
TEASDALE, J.D., Segal, Z.V., & Williams, J.M.G. (in press). Mindfulness and problem formulation. Clinical Psychology: Science and Practice.
TEASDALE, J.D., Segal, Z.V., Williams, J.M.G., RIDGEWAY, V.A., Soulsby, J.M., & Lau, M.A. (2000). Prevention of relapse/recurrence in major depression by mindfulness-based cognitive therapy. Journal of Consulting and Clinical Psychology, 68, 615-623.
Vasey, M., DALGLEISH, T., & Silverman, W. (in press). Research on information-processing factors in child and adolescent psychopathology: A critical commentary. Journal of Child & Adolescent Clinical Psychology.
Watkins, L.H.A., Rogers, R.D., LAWRENCE, A.D., Sahakian, B.J., Rosser, A.E., & Robbins T.W. (2000) Impaired planning but intact decision making in early Huntington's disease: implications for specific fronto-striatal pathology. Neuropsychologia, 38, 1112-1125.
Watkins, E., & TEASDALE, J.D. (2001). Rumination and overgeneral memory in depression: Effects of self-focus and analytical thinking. Journal of Abnormal Psychology, 110, 353-357.
Watkins, E., & TEASDALE, J.D. (submitted). Adaptive and maladaptive self-focus in depression.
Watkins, E., TEASDALE, J.D., & Williams, R. (2000). Decentering and distraction reduce overgeneral autobiographical memory in depression. Psychological Medicine, 30, 911-920.
Watkins, E., TEASDALE, J.D., & Williams, R. (in press). Contextual questions prevent mood primes from maintaining experimentally induced dysphoria. Cognition and Emotion.
Williams, J.M.G., TEASDALE, J.D., Segal, Z.V., & Soulsby, J. (2000). Mindfulness-based cognitive therapy reduces overgeneral autobiographical memory in formerly depressed patients. Journal of Abnormal Psychology, 109, 150-155.
WOOD, J., MATHEWS, A., & DALGLEISH, T. (2001). Anxiety and cognitive inhibition. Emotion, 1, 166-181.
YIEND, J., & MATHEWS, A. (2001a). Anxiety and attention to threatening pictures. Quarterly Journal of Experimental Psychology, 54, 665-681.
Young, A.W., Rowland, D., CALDER, A.J., Etcoff, N.L., Seth, A., & Perrett, D.I. (1997). Megamixing facial expressions. Cognition, 63, 271-313.
Books
DALGLEISH, T. & Power, M.J. (Eds.) (1999). Handbook of Cognition and Emotion. Chichester: Wiley.
Davies, G. & DALGLEISH, T. (Eds.) (2001). Recovered Memory Experiences: Seeking the Middle Ground. Chichester: Wiley.
Power, M.J., & DALGLEISH, T. (1997). Cognition and Emotion: From Order to Disorder. Hove: Psychology Press.
Segal, Z.V., Williams, J.M.G., & TEASDALE, J.D. (2002). Mindfulness-Based Cognitive Therapy for Depression: A new approach to preventing relapse. New York: Guilford Press.
Williams, J.M.G., Watts, F.N., MacLeod, C., & MATHEWS, A. (1997). Cognitive Psychology and Emotional Disorders. Chichester, Wiley.
Theses
Bishop, S. (2001). The cognitive processing of emotional information in childhood anxiety, depression and post-traumatic stress disorder.
Ma, H. (2002). Prevention of relapse/recurrence in recurrent major depression by mindfulness-based cognitive therapy .
Dunn, B. (2002). Exploring the Interaction of Mind and Body in Depression
Potts, J. (2002). The effect of mood and modes of cognitive processing on overgeneral memory and social problem-solving.
Rafter, A. (2002). Preattentive bias in anxiety.
Sheppard, L. (2000). Processing of the depressive schema in major depressive disorder
Watkins, E. (1999). Contextual awareness and the maintenance of depressed mood.
Wood, J. (2000). Inhibitory processing and anxiety.
Book Chapters & Contributions
BARNARD, P.J. (1999). Interacting Cognitive Subsystems: modelling working memory phenomena within a multi-processor architecture. In: Miyake, A & Shah, P. (Eds) Models of Working Memory: Mechanisms of Active Maintenance and Executive Control, (pp. 298-339). Cambridge: Cambridge University Press.
BARNARD, P., (in press). Asynchrony, implicational meaning and the experience of self in schizophrenia. In A. David & T. Kircher (Eds). The Self in Neuroscience and Psychiatry. Cambridge: Cambridge University Press.
BARNARD, P. & May, J. (2000) Towards a theory-based form of cognitive task analysis of broad scope and applicability. In. J.M.C. Schraagen, S.F. Chipman, & V.L. Shalin, (Eds.) Cognitive Task Analysis, (pp.147-163). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
BARNARD, P., May, J., Duke, D. & Duce, D. (2001). Macro-theory for Systems of Interactors. In: J.M. Carroll (Ed.) Human-Computer Interaction in the New Millennium, Ch2, (pp.31-52). Boston: Addison-Wesley.
DALGLEISH, T. (1998). Emotions. In M.W. Eysenck (Ed.), Introduction to Psychology: A European perspective. London: Longman.
DALGLEISH, T. (1999). Cognitive theories of PTSD. In W. Yule (Ed.), Post Traumatic Stress Disorders. Chichester: Wiley.
DALGLEISH, T. (2002). Posttraumatic Stress Disorder. In D. Miel (Ed.), Exploring Psychology (Vol. 3). Milton Keynes: Open University Books.
DALGLEISH, T. (in press). Information-processing approaches to emotion. In R.J. Davidson, K. Scherer & H. Hill Goldsmith (Eds.), Handbook of Affective Sciences. Oxford: Oxford University Press.
DALGLEISH, T. & Bramham, J. (1999). Cognitive Perspective. In D. Levinson, J. Ponzetti & P. Jorgensen (Eds), Encyclopedia of Human Emotions, (pp. 118-121). New York: Macmillan.
DALGLEISH, T., & COX, S. (2000). Mood and memory. In G.E. Berrios & J.R. Hodges (Eds.), Memory Disorders in Psychiatric Practice. Cambridge: Cambridge University Press.
DALGLEISH, T., & COX, S. (in press). Mood, memory and emotional disorder. In A. Baddeley, B. Wilson & M. Kopelman (Eds.), Handbook of memory disorders (2nd Ed.). Chichester: Wiley.
DALGLEISH, T., & Morant, N. (2001). Representations of child sexual abuse: A brief psychosocial history and commentary. In G. Davies & T. DALGLEISH (Eds.), Recovered Memory Experiences: Seeking the middle ground. Chichester: Wiley.
DALGLEISH, T., & Power, M.J. (1999). Cognition and emotion: Future directions. In T. DALGLEISH & M. Power (Eds.), Handbook of Cognition and Emotion. Chichester: Wiley.
Gardiner, J.M., Richardson-Klavehn, A., RAMPONI, C. & Brooks, B.M. (2002.) Involuntary Level-of-Processing In M.Naven-Benjamin, M.Moskovitch, & H.L. Roediger III (Eds).Effects in Perceptual and Conceptual Priming Perspective on Human Memory and Cognitive Aging: Essays in Honor of Fergus Craik. Philadelphia: Psychology Press.
LAWRENCE, A.D. (2002). Huntington's disease. In J.E. Harrison & A.M. Owen (Eds.). Cognitive Deficits In Brain Disorders. London: Martin Dunitz.
LAWRENCE, A.D, & Grasby P.M. The functional neuroanatomy of emotional disorders. In G. Gainotti (Ed.), Handbook of Neuropsychology, 2nd edition. Volume 5, (pp. 235-262). Amsterdam: Elsevier. 2001.
LAWRENCE, A.D, & Sahakian, B.J. Cognition. (2001). In J.W. Fawcett, A.E. Rosser, S.B. Dunnett (Eds). Brain Damage, Brain Repair. Oxford: Oxford University Press.
LAWRENCE, A.D., & Sahakian, B.J. (2001). Outcome variables in dementia trials: conceptual and practical issues. In R. Guiloff (Ed), Clinical Trials In Neurology. London: Springer.
MATHEWS, A. (2002). Emotional processing biases: nature and modification. In L. Backman & C. von Hofsten (Eds), Psychology at the Turn of the Milennium, Volume 1: Cognitive, Biological and Health Perspectives. Hove: Psychology Press.
May, J. & BARNARD, P. (In Press). Cognitive Task Analysis. In Diaper, D. & Stanton, N, (Eds.) The Handbook of Task Analysis for HCI. Hillsdale: Lawrence Erlbaum Associates.
TEASDALE, J.D. (1997a). The relationship between cognition and emotion: The mind-in-place in mood disorders. In D.M. Clark & C.G. Fairburn (Eds.). Cognitive-Behaviour Therapy: Science And Practice. Oxford: Oxford University Press.
TEASDALE, J.D. (1997b). The transformation of meaning: The Interacting Cognitive Subsystems approach. In M.J. Power & C.R. Brewin (Eds.), The Transformation Of Meaning In Psychological Therapies: Integrating Theory And Practice. Chichester: John Wiley.
TEASDALE, J.D. (1999a). Multi-level theories of cognition and emotion. In T. DALGLEISH & M.J. Power (Eds.), The Handbook of Cognition and Emotion. Chichester & New York: John Wiley.
TEASDALE, J.D. (2000). Mindfulness-Based Cognitive Therapy in the Prevention of Relapse and Recurrence in Major Depression. In K.T. Kaku (Ed.). Meditation as Health Promotion: A Lifestyle Modification Approach. Delft, Holland: Eburon
Thrasher, S., & DALGLEISH, T. (1999). Information processing research in PTSD. In W. Yule (Ed.), Post Traumatic Stress Disorders. Chichester: Wiley.
YIEND, J. & DALGLEISH, T. (2002). Grief. In J.Firth (Ed.), Oxford Textbook of Medicine (4th Ed.). Oxford: Oxford University Press.
YIEND, J. & DALGLEISH, T. (2002). Stress. In J.Firth (Ed), Oxford Textbook of Medicine (4th Ed.). Oxford: Oxford University Press.
YIEND, J. & MATHEWS, A. (2002). Induced biases in the processing of emotional information. Advances in Psychology Research: Volume 13, Nova Science.

Conference Proceedings, Published Abstracts and Other Reports
BATTYE, G. & BARNARD, P. (2002). The Anterior Cingulate Cortex and Conflict Monitoring: fMRI studies from a range of paradigms implicate inhibitory mechanisms in top down attentional control. NeuroImage, 16, 657.
Bowman, H., & BARNARD, P. J. (2001). Computational Modelling of Distributed Executive Control Technical Report , 12-01: Computing Laboratory, University of Kent at Canterbury, www.cs.ukc.ac.uk/people/staff/hb5/pubs.local.
CALDER, A.J., Burton, A.M., Miller, P. and Akamatsu, S. (1999). It's written all over your face: An image-based analysis of facial expressions. Abstracts of the Psychonomic Society, 4, 33.
CALDER, A.J., KEANE, J., Manes, F., Antoun, N. and Young, A.W. (2001).Impaired recognition and experience of disgust following damage to the insula and basal ganglia.. Journal of Cognitive Neuroscience, 13, Suppl, 30.
CALDER, A.J., KEANE, J. and Young, A.W. (2000). Loss of disgust: Impaired perception of disgust following brain injury. Abstracts of the Psychonomic Society, 5, 24.
CALDER, A.J., LAWRENCE, A.D., KEANE, J., Scott, S.K., Owen, A., Christoffels, I. and Young, A.W. (2002). Reading the mind from eye gaze. Journal of Cognitive Neuroscience,14, Suppl, 74.
CALDER, A.J., Lyons, M., Christoffels, I., Akamatsu, S. (1999). Recognition of orientation-filtered facial expressions. Perception, 28(Suppl): 112a.
CALDER, A.J., Young A.W., Karnat A., Sprengelmeyer R., Perrett D.I., Rowland D.(1997). Selective deficits in emotional expression recognition following brain damage. International Meeting of the Psychophysiology Society, Finland, 1997.
CALDER, A.J., Young, A.W. and KEANE, J. Configural encoding of facial expression. European Conference on Visual Perception, Oxford, 1998. Perception, 27(Suppl), 9a.
CALDER, A.J., Young, A.W., Rowland, D. & Perrett, D.I. (1997). Perception of facial expressions: Differentially severe impairment of fear. Journal of the International Neuropsychological Society, 3, 224.
CALDER, A.J., Young, A.W., Sprengelmeyer, R., Perrett, D.I., and Rowland, D. (1999). Selective deficits in recognising basic emotions following brain damage. International Neuropsychology Society, Durban, South Africa, 1999. Journal of the International Neuropsychological Society, 5, 276.
Cottrell, G.W., Branson, K.M., & CALDER, A.J. (2002). Do expression and identity need separate representations. In Proceedings of the 24th Cognitive Science Conference. Mahwah: Lawrence Erblaum.
Duke, D.J., Duce, D.A., BARNARD, P.J. & May, J. (2001). Human-Computer Protocols. In: C. Stephanidis (Ed), Universal Access In HCI, Volume 3 of the Proceedings of HCI International 2001, (pp. 296-300). Hillsdale, N.J.: Lawrence Erlbaum Associates.
Goerendt, I.K., LAWRENCE, A.D., Brooks, D.J. (2000). Neural correlates of apathy in Parkinson's disease. Movement Disorders, 15, Suppl.
Goerendt, I.K., LAWRENCE, A.D., & Brooks, D.J. (2000). Adaptive neural circuitry in Parkinson's disease. An H2[15]O PET study. Neurology, 54 (suppl 3), A329.
Goerendt, I.K., LAWRENCE, A.D., Stern, J.S., Schimke, N., Gerstner. A., Oertel, W., Odin, P. & Brooks, D.J. (2002). Differential neuromodulation of reward processing pathways in Parkinson's disease following dopaminergic medication and subthalamic nucleus deep brain stimulation. Neurology 58, (Suppl 3), A55-A56.
Goerendt, I.K., Messa, C., Lawrence, A.D., Grasby, P.M., Piccini, P. (2002). Endogenous dopamine release during sequential movement in healthy volunteers and patients with Parkinson's disease: a PET study. Neuroimage, 13, S1172.
Goerendt, I.K., Messa, C., Lawrence, A.D., Grasby, P.M., Piccini, P., Brooks, D.J. (2001). Endogenous dopamine release during movement in healthy volunteers and patients with Parkinson's disease: A C-11-raclopride PET study. Neurology 56, (Suppl 3), A271.
Joyce, E.M., Hutton, S.B., MURPHY, F.C., McKenna, P.J., Barnes, T.R., Sahakian, B.J., & Robbins, T.W. (2001). Decision making deficits in patients with first episode and chronic schizophrenia. Schizophrenia Research, 49 (suppl.),110.
LAWRENCE, A.D., CALDER, A.J., McGowan, S.W. & Grasby, P.M. (2001). Systemic supiride administration impairs anger recognition in healthy volunteers. Journal of Cognitive Neuroscience, 13, 26.
LAWRENCE, A.D., Koepp, M.J., Messa, C., Rabiner, E., Gunn, R.N., & Grasby, P.M.(1999) Investigating dopaminergic and serotonergic correlates of temperament: PET studies. Journal of Psychopharmacology, 13 (suppl), A40.
LAWRENCE, A.D., & MURPHY, F.C. (2001). Frontal asymmetry in positive and negative emotion: Evidence from functional neuroimaging. NeuroImage, 13, S438.
McGowan, S.W., LAWRENCE, A.D., Grasby, P.M. (2000a). Do neuroleptics modulate prefrontal function? A pet study of spatial working memory and sulpiride. Neuroimage 12 (suppl).
McGowan, S.W., LAWRENCE, A.D., Grasby P.M. (2000b). A pilot study of the effect of a dopaminergic antagonist on prefrontal activation in healthy volunteers. Schizophrenia Research, 41 (suppl), A248.
McGowan, S.W., LAWRENCE, A., Sales, T., Quested, D., & Grasby, P.M. (2001). Presynaptic dopaminergic dysfunction in medicated schizophrenic patients. Schizophrenia Research 49, 195-195 Sp. Iss. SI Suppl. S APR 15 2001.
Mehta, M.A., McGowan, S., LAWRENCE, A.D., Aitken, M., Montgomery, A.J., Grasby, P.M. (2002). The effects of systemic sulpiride on regional cerebral blood flow during working memory and planning in man. Journal of Psychopharmacology, 16, (suppl), A27.
MURPHY, F.C., Smith, K.A., Cowen, P.J., Robbins, T.W., & Sahakian, B.J. (2000). The effects of acute tryptophan depletion of cognitive and affective processing in healthy volunteers. Journal of Psychopharmacology, 14, (suppl), A25.
Paterson, H., Pollick, F.E., CALDER, A.J., & Sanford, A.J. (1999). The perception of affect from point-light displays of simple arm movements. Perception, 28(Suppl), 53a.
Pollick, F.E., CALDER, A.J., Hill, H. and Paterson, H. (2002). Recognizing facial expression from spatially and temporally modified movements. Journal of Cognitive Neuroscience, 14, 126.
Pollick, F.E., Paterson, H., CALDER, A.J., Bruderlin, A. and Sanford, A.J. (2000). Perceiving affect from point-light displays of arm movements. Abstracts of the Psychonomic Society, 5, 39.
Rabiner E.A., LAWRENCE, A.D., Bantick, R.A., Bhagwagar, Z., Cowen, P.J., & Grasby, P.M. (2002b). 5HT1a receptor binding correlates negatively with EPQ-L. Journal of Psychopharmacology, 16 (suppl). A51.
Richards, A., French, C., Webb, Fox, Young, A.W. & CALDER, A.J. (1999). Anxiety-related bias in interpretation of emotionally ambiguous facial expressions. Sixth European Congress of Psychology, Rome, July 1999.
Scott, S.K., CALDER, A.J. & Young, A.W. (1997). Impaired perception of auditory emotions after bilateral amygdala lesions Loss of fear and anger. Journal of the International Neuropsychological Society, 3, 224.
Uher R., Murphy T., Ng V., Phillips M., DALGLEISH T, & Treasure J (2001). Perception of food and emotional stimuli in anorexia nervosa. Neuroimage, 13, S1022.
YIEND, J., & MATHEWS, A. (2001b). Anxiety and attention to threatening pictures. Abstracts of the Psychonomic Society 42nd Annual Meeting, Orlando, Florida, Psychonomic Society.

TRANSFER TO HEALTH SERVICE

The trials demonstrating the efficacy of mindfulness-based cognitive therapy (MBCT) carried out by Dr Teasdale and colleagues have led to the establishment of some MBCT programmes in the NHS.

EXTERNAL GRANTS

Calder, A. (1998-99): Award from ATR Human Information Processing Research Laboratories, Kyoto, Japan. to facilitate collaborative research with members of the ATR laboratories. $20,000.

Lawrence, A. (2002-2004). Investigation of dopamine dysregulation in Parkinson's disease. Parkinson's Disease Society. £50, 000.
Mathews, A. (1998-2000). Prediction of anxiety in breast cancer patients. Cancer Research Campaign.£60,000.

Teasdale, J.D. (1993-1998). Joint Grant holder, Medical Research Council Special Project Grant Cognitive therapy for residual depressive chronicity - a controlled trial. £713,000

Teasdale, J.D. (1995-1998). Joint Grant Holder, NHS Wales Research and Development (Reducing relapse after recovery on antidepressant medication: an evaluation of Attentional Control Training. £97, 850

COLLABORATORS

Dr L. Andrews, University of Essex.
Professor D. Brooks, MRC Clinical Sciences Centre, London
Dr H. Bowman, Dept. of Computer Science, University of Canterbury
Professor A. M. Burton, University of Glasgow.
Dr. D. Duke, Dept of Computer Science, University of Bath
Professor E. Foa, University of Pennsylvania
Professor E. Fox. University of Essex
Professor A.L. Johnson MRC Biostatistics Unit
Dr S. Joseph, University of Warwick.
Professor I.M Goodyer, Section of Child and Adolescent Psychiatry, Cambridge University
Professor P. Grasby, MRC Clinical Sciences Centre, London
Professor J. Herbert, Department of Anatomy, Cambridge University
Professor A. Lees, University College London
Professor C. MacLeod, University of Western Australia
Dr. J. May, Dept of Psychology, University of Sheffield
Dr A. Palmer, Hellesden Hospital, Norwich
Dr R.J. Park, Section of Child and Adolescent Psychiatry, Cambridge University
Professor E.S. Paykel, Department of Psychiatry, Cambridge
Professor D.I. Perrett, St. Andrews University
Professor M.L. Phillips, Institute of Psychiatry
Professor M. Power, University of Edinburgh
Professor TW Robbins, Experimental Psychology, Cambridge
Dr B. Sahakian, Dept. of Psychiatry, University of Cambridge
Professor J. Scott, University of Newcastle (now at Institute of Psychiatry, London)
Professor Z.V. Segal, University of Toronto
Dr R. Sprengelmeyer, St. Andrews University
Professor J. Treasure, Institute of Psychiatry
Dr E. Watkins, Institute of Psychiatry, London
Professor J.M.G. Williams, University of Wales
Professor A.W. Young, University of York
Professor W Yule, Institute of Psychiatry

Other sections in the 1998-2002 report

1. SUMMARY

2. ATTENTION GROUP

3. COGNITION AND EMOTION GROUP

4. LANGUAGE AND COMMUNICATION GROUP

5. MEMORY AND KNOWLEDGE GROUP: DMS SECTION

6. MEMORY AND KNOWLEDGE GROUP: REHABILITATION SECTION

7. METHODS RESEARCH AND INFRASTRUCTURE GROUP