MRC Cognition and Brain Sciences Unit

You are in:  Home » History of the Unit

7. COGNITIVE SKILLS

General Notes. This material has been scanned from the original typescript, while we have done our best to remove errors, some may well remain. You can access other parts of this particular Progress Report either from the menu at the bottom of this entry or by navigating back to the Unit history timeline. References for this report are indexed by number and these can be found in a dedicated section also accessible from the menu at the bottom of this entry.

7.1 Arithmetic (Hitch, Young)

A series of experiments by Hitch (107; 108) explored the role of information storage in working memory in performing mental arithmetic. They showed that the forgetting of both initial and interim information was a potent source of errors. A memory decay model was shown to provide a good fit to the data.

Using an approach based on "Production Systems" (i.e. rule-based information-processing systems) Young has been concerned with the errors in children's written subtraction, showing that they can largely be accounted for by the omission of individual rules from a correct pro­duction system (307). This work is continuing in collaboration with T. O'Shea of the Open University. At present, data on the real-time performance of subtraction are being gathered in the classroom, the aim being to provide information about changes in the skill at the time that a new rule is being acquired. A future direction for the theoretical analysis is to model the way in which children's experience with various concrete representations of arithmetic—such as Dienes blocks, abacuses, or counters—affects their skill at written arithmetic and their more abstract understanding of number.

7.2 Language

7.2.1 Word comprehension (Marcel)

The backward-masking technique developed by Marcel has been extended to investigate how polysemous (multi-meaning) words are under­stood. These studies (130; 131) suggest that both meanings of such words (e.g. PALM) are automatically accessed unconsciously and that verbal context serves to select that meaning which will be apprehended consciously.

7.2.2 Word retrieval in aphasia (Morton, K. Patterson, Purell)

The studies of acquired dyslexia and dysgraphia described above are proving enormously fruitful in the framework of theoretical cognitive psychology. From a more practical or applied point of view, we have been assessing the efficacy of certain techniques used by speech therapists to assist word retrieval in aphasic patients. When a patient cannot find the name of something which he wants to say (e.g. "sugar"), speech therapists commonly attempt to improve the patient's performance in one of the following three ways:

(a) The therapist says the word and asks the patient to repeat it.

(b) The therapist gives a partial phonemic cue for the word, e.g. "sh" or "shu".

(c) The therapist gives a sentence context for the word, e.g. "I like my tea sweet; please may I have some ........... ?"

In a research project funded by the DHSS, we have designed tests to evaluate these three techniques. Studies of the first two are complete (176). The results are quite clear, if a little disheartening: we have confirmed that repetition and phonemic cueing are successful at the time of application, but we have found no evidence for any long-term benefit from these procedures. The third study on contextual cueing, plus a few individual case studies involving intensive exposure to repetition and phonemic cueing, are planned for next year.

7.2.3 Philosophy of psychology (Morton)

Many theories of psychology and language appear to be confused in their aims. In particular there is a tendency to try to find biological accounts of psychological phenomena. Morton, together with Jacques Mehler (of C.N.R.S., Paris) has critically examined some of the current trends. They see little reason to suppose that there are significant biological constraints on psychological theories (143 U; 154; 155). Attempts have also been made to explore, from a phylogenetic as well as an ontogenetic viewpoint, possible relationships between language-specific mechanisms and more general cognitive functions (153). The possible relationships between models and data have also been studied.

7.3 Reading

7.3.1 Psychological units in reading (McKenna)

Letter detection tasks can be used to indicate the way in which words are psychologically represented. A series of experiments by McKenna suggests that the stem of a word such as "waiting" (i.e. wait), is represented in a different way from its suffix (ing). This finding relates to certain neurological reading disorders in which patients show preserved reading of word stems but impaired reading of affixes (171; 214).

By using tasks in which it is necessary to process every letter (typing and handwriting) McKenna has shown that the suffix is processed separately from the stem, but as a unit. Future research will attempt to determine whether the unitisation proceeds directly from the word or whether the context directs this process.

7.3.2 Translation of print to phonology (Marcel, Kay)

It has commonly been supposed that phonological representation of alphabetic written language may be achieved by non-lexical grapheme-phoneme rules. An examination of the oral reading of single words by Surface Dyslexic patients and normal beginning readers by Marcel (129) casts doubc on this. A new theory has been proposed whereby the orthography of both words and nonwords is segmented lexically and orthographic segments are pronounced by analogy with their occurrence in known words. A series of experiments on normal adults supports this view and under­mines the classic "two-process" model (116).

7.3.3 Phonological awareness and literacy (Marcel)

An important factor in learning to read and write alphabetic scripts is the reflexive awareness that our own speech can be segmented into phonemes. Marcel (126) has suggested its importance in both the initial teaching and clinical breakdown of reading. In a joint project with Halliday of Manchester University, he is investigating both the relationship of phonetic segmentation with reading problems and efficacious methods of initial teaching and remediation.

In addition, the way in which phonological awareness is related to spelling has been explored in groups of children, adult illiterates and aphasic patients whose speech perception and production appears to rely on an inadequate acoustic code (128).

7.3.4 Phonological coding in fluent reading (Baddeley, Eldridge, Lewis)

The rate of subvocalisation in fluent adult reading was studied by requiring subjects to produce an irrelevant utterance, e.g. the word "the", continuously while reading and comprehending sentences (4; 18) and prose (14). We found our subjects able to read and comprehend simple sentences with no apparent impairment. With more complex passages they read equally rapidly, but are more prone to miss errors in the text, indicating that subvocalisation may operate as an optional checking mechanism in fluent reading. Evidence that sub­vocalisation is not essential for phonemic coding comes from studies in which subjects were required to judge whether pairs of words rhymed or not. They were able to do this while suppressing articulation, with no marked loss in either speed or accuracy. It is suggested on the basis of this result that subvocalisation may be supplemented by an auditory image which may play an important role in reading.

7.3.5 Acquired disorders of reading

Deep dyslexia (Baddeley, Lewis, Marcel, Morton, K. Patterson, Shallice)

Morton and Patterson (158) have attempted to provide a comprehensive theoretical account of the pattern of deficits observed in deep dyslexia, based on Morton's logogen model of the processes of recognition, com­prehension and pronunciation of written words by normal readers. Shallice in collaboration with Dr. E.K. Warrington of the National Hospital, has also developed a general model of the central acquired dyslexias, in which deep dyslexia is viewed as a mixed syndrome which can arise from varying combinations of more basic syndromes (214). For example, Shallice and A.K. Coughlan (of the Wolfson Rehabilitation Centre) have conducted a case study on deep dyslexia which has established that the deficit can arise from a modality-specific deficit in attaining the meaning of abstract words together with a deficit of the phonological reading route (213).

In the past, major interest in this syndrome has centred on factors such as semantically related errors in reading (e.g. protein » carbo­hydrate") and the major influence of word imageability or concreteness on reading performance. Recently, however, Morton and Patterson have turned to analysis of some of the more syntactic/grammatical dimensions of the syndrome.

(1) They have investigated whether grammatical function words (like of, on, with, her etc.), which the patients have grave difficulty in reading, are comprehended (157 U). The answer is:

(a) At least the more meaningful function words (like on and her) are well understood by a deep dyslexic patient. Therefore the reading deficit does not 'reduce' to a comprehension deficit, since even these rather meaningful function words are poorly read.

(b) Such words may be reasonably well understood in isolation; but the patient is unable to use word order or relational information to construct the meaning of a whole message. Thus he knows that on means something to do with top or above; but given the sentence The magazine is on the book, he cannot work out which object is on cop.

(2) Patterson has also been investigating effects of the morphemic structure of words (mostly involving affixes like -ed, -er, -tion, etc.) (171). Like function words, these affixes cause grave difficulties for deep dyslexic patients in reading aloud. Experimental questions have addressed patients' knowledge of correct inflectional and derivational forms (e.g. ability to judge chat fearest and quickly are legitimate words whereas fearest and passly—while composed of real root morphemes and real suffixes—are not) and their ability to select correct derivational forms of words to fit in sentences.

Baddeley and Lewis, in collaboration with T.R. Miles and N. Ellis of the University of Wales, Bangor, have tested the recent claim that developmental dyslexic children reveal a syndrome closely related to deep dyslexia. The reading of developmental dyslexics resembled that c: normal but younger children rather than that of deep dyslexic ra:ients (15).

Phonological dyslexia (K. Patterson, Shallice)

A major question about normal adult reading, to which the study of deep dyslexia has been considered relevant, is the following: are printed words generally (or indeed ever) translated into a phonological representation in order to be recognised and understood? A syndrome of acquired dyslexia described for the first time in 1978, phonological dyslexia, addresses this question even more directly. A normal reader will easily know how to pronounce a new word, such as a surname he has not seen before (e.g. Doob). A phonological dyslexic patient cannot do this (173; 214). Oral reading of familiar words, however, is almost normal in such patients. There are three important implications to be drawn from this research:

(a) If a severe impairment to phonological coding engenders such minimal disruption in a patient's reading, then it is very unlikely that such coding plays a critical role in normal reading.

(b) Theoretical accounts of deep dyslexia have sometimes proposed that symptoms in that syndrome (such as semantically related errors like anxiety* "nervous") might be attributable to the deep dyslexic's loss of phonological coding ability. If phonological dyslexic patients show the phonological impairment but not the other symptoms, however, then the hypothesised causal relation­ship becomes untenable. Thus we are advancing in our notions of how symptoms relate to functional components of the system.

(c) There is one deficit in reading individual words shared by phonological and deep dyslexic patients: a difficulty with grammatical morphemes, especially bound morphemes (affixes) and possibly also free morphemes (function words like of, at, with, etc.). Patterson (173) has suggested that this feature might be attributable to the phonological deficit.

Letter-by-letter reading (Kay, K. Patterson, Shallice)

Patients with this disorder, as the designation implies, read a word by identifying each of its constituent letters in succession. Time required to read a word is thus enormously slow and a monotonic function of word length. In the framework of our developing process models, we ask the following question: is the actual mechanism or process which allows normal readers to recognise words, deficient in these patients, as argued by Warrington and Shallice (225), or is this mechanism intact but forced to accept abnormal input from a preceding Stage of visual analysis of the written word, as argued by Patterson and Kay (175 U)?

Of interest in many interrelated disciplines is the extent to which representation of language in the brain is typically (for right-handed people) lateralised exclusively to the left hemisphere. Although speech production is widely thought to be the sole province of the left hemisphere, there is more controversy regarding comprehension. Patients showing letter-by-letter reading usually have well-localised left-occipital lesions; yet recent extensive investigation of at least six such patients reveals almost no evidence of comprehension of written words despite an apparently intact right hemisphere (175 U; 225).

Surface Dyslexia (Marcel, Shallice)

Work at the National Hospital by Shallice and McCarthy in collaboration with Dr. E.K. Warrington has led to the first detailed study of a "pure" case of surface (or semantic) dyslexia; it produced support for a particular model of the operation of the phonological route in reading. Their analysis fitted well with a more general 2-route model of the central acquired dyslexias for which they had previously provided evidence (214).

Marcel has re-analysed the oral reading of surface dyslexic patients. Contrary to standard accounts, it was shown that patients show several pronounced lexical effects in their reading. This has led to a new account of oral reading in normal people (129) which has gained support from further experiments (116). This model eschews non-lexical grapheme-phoneme rules and relies on the lexicon to segment orthography and retrieve phonology by analogy.

Acquired Dysgraphia (Shallice, K. Patterson)

Work on writing and spelling disorders has shown that a 2-route framework analogous to that developed in recent years for analysis of the acquired dyslexias can be applied to the dysgraphias. The syndrome of phonological dysgraphia, involving a specific deficit to the phonological route, has been described by Shallice (212) and the syndrome of surface dysgraphia, involving reliance on the phonological route, by Patterson and Hatfield (174 U).