You are in: Home » Research at the Unit » Rehabilitation research at the Unit
The Sustained Attention to Response Test (SART)
Tom Manly (May 2009)
Background to the SART
An issue in the literature on the effects of Traumatic Brain Injury (TBI) was that, whilst complaints of problems with maintaining concentration were very common, formal tests of sustained attention using vigilance paradigms generally failed to detect obvious problems. Vigilance paradigms had been developed (not least at the CBU – then APRU) to capture real-life situations such as radar monitoring or watching a conveyor belt to spot defects in manufactured goods – in other words, situations in which something relevant came along very rarely and the main job consisted of keeping oneself sufficiently alert/awake to spot a signal when it came. Accordingly, in vigilance tasks, people were typically asked to watch a stream of information during which a target would occasionally appear. The key measure was not the overall level of performance but rather the change in performance from, say, the first to last half of the task – this being a conceptually neat way of controlling for the general demands of the task (e.g. how well you can see or hear were likely to remain constant throughout and so any change could be attributed to attentional drift). Although this logic is sound, a problem with such tasks is if attention does not drop-off in a gradual way but rather lapses and recovers over much shorter time-scales. Performance on vigilance tasks could therefore switch between being relatively good and relatively poor – detectable in overall performance and performance variability, but not in a simple time-on-task decline. Another issue with vigilance tasks is that, in using a single target, performance may tend to become rather automatic. Your own name is a good example. In a crowded party situation you may be paying no attention to the general babble and yet your name - an arbitrary word that has been repeatedly used with reference to you – can cut through and draw your attention. Whilst the target in a vigilance task does not have anything like this level of training, in that detection becomes relatively automated, the demands on sustained attention may actually decline with time.
The SART was designed to get around these problems and, potentially, become a better measure of sustained attention for people with TBI and other brain injuries. The measure, developed for a study by Ian Robertson, Tom Manly, Jackie Andrade, Bart Baddeley (yes, son of...) and Jenny Yiend is a a "go no-go" paradigm, in which responding to the target is placed in competition with, rather than becomes, the automatic response (Robertson et al., 1997).
In the task, people watch single digits 1-9 appearing on the computer screen at a regular, rhythmic rate of one every 1.15 seconds. They are asked to press a the same response key for each number – an action that rapidly becomes rather automatic and 'driven' by the task, essentially just tapping along in time. The catch is that, occasionally, a digit designated as a no-go trial will be presented. The trick is therefore to resist making your frequent responses too automatic so that you stand a chance of withholding your response at the critical moment. That turns out to be rather demanding and most healthy people will let their attention wander and make mistakes on a few no-go trials. When compared with a more conventional vigilance task, the SART proved sensitive to traumatic brain injury severity and, critically, correlated with everyday life problems that patients reported (Robertson et al., 1997). It therefore had the potential to be a relatively brief (about 4.5 mins usually) but sensitive and informative clinical measure.
SART FAQ
1. Is the SART available commercially or by another route for clinical assessment or research?
Whilst data from 109 healthy adults are presented in Manly et al. 2000, this is not necessarily a representative sample in terms of age, IQ etc for making relevant clinical comparisons – although can, of course, be used as the basis for cautious interpretation. We have made versions of the SART available for research projects although this is a rather labour intensive business. Soon we are hoping, with overseas colleagues, to make a version of the SART available for download for research purposes.
2. Is the SART a measure of sustained attention or response inhibition?
The best (although perhaps not the most helpful) answer to this seems to be that the SART aims to use response inhibition to measure sustained attention. As highlighted above, most healthy people accidentally respond to some no-go trials. However, they accurately withhold their responses to the majority of these trials. What is the difference between these accurate and error trials? Why has the person's 'response inhibition' capacity changed. One way of thinking about that is that, when people are actively attending to what they are doing, they are in a better position to withhold their responses. When they are absentmindedly performing that task, even if they notice the no-go digit, it is often too late to prevent the response. This type of argument is useful but clearly, differences between people could be attributable to response inhibition or sustained attention. Someone with poor response inhibition, for example, would need to pay more attention to the task to get the same score as someone with relatively poor attention but who, in theory, had good 'response inhibition' mechanisms. One way in which we have looked at that is to examine the effects of reminding people to pay attention. In some studies we played people occasional 'bleeps' and asked them to use this a reminder to keep focused on the task. The idea was that these bleeps helped them to maintain their attention but – other than through improving attention – would not change basic response inhibition capacity. By looking at the degree of improvement we could estimate the contribution of poorly sustained attention (when no bleeps were presented) to error rates. The merit of this approach – although clearly a noisy measure – is that it has the potential to inform more detailed assessment and rehabilitation (see here, for example).
3. What is the relationship between speed and accuracy in the SART?.
Speed-accuracy trade-off, as the name implies, is a well known phenomenon by which, if people are rushing they tend to make more errors. This is particularly true in go no-go tasks in which, like the SART, you have to respond to some or most trials but withhold a response to others, or 'stop-signal paradigms' (Logan, Schachar, & Tannock, 1997) in which people are asked to respond as fast as possible to a go-signal which may or may-not be shortly followed by a stop-signal. If you respond quickly to go-signals, you tend to make mistakes on more no-go or stop-signal trials. This is certainly true in the SART with the correlation between go-trial reaction times (RT) and no-go trial error rates being typically in the region of 0.5. So one argument is that some people, for whatever reason, are fast go responders and they have more trouble stopping this response (the horse has bolted before the stable door is shut).
However, this relationship could also be mediated by something like attention. In the SART, if you are paying close attention to the identity of each digit in order to choose your response, you are likely to be slower on go trials than if you have absentmindedly allowed your responses to be driven by the task (even by anticipating the onset of the next trial). This would still produce a 'speed-accuracy trade-off' but with rather different implications to the 'fast vs. slow' person model. One elegant approach to this has been described by Jonathan Smallwood and colleagues (see reference list below) who measure the degree to which SART participants are engaged in task unrelated thoughts (TUTs!) and changes in speed-accuracy trade-offs. Another approach that we took was to try and reduce inter-person go-trial speed variability. To do this we put a click in each trial and asked people to respond on the click rather than "as fast as possible" to each digit. The idea was that the automaticity of the task would be similar – you can absentmindedly press-press-press in time with the click-click-click but that reaction times relative to the crucial digit onsets would be more fixed. Interestingly, while this did reduce between-people variations in RT, error rates were not changed and were highly correlated with the standard (no click) SART (Manly, Davison, Heutink, Galloway, & Robertson, 2000) suggests that, to a first approximation at least, people with relatively poor sustained attention are the ones who will tend to speed-up/show more errors in the standard SART. research/rehabilitation/Manly%20et%20al%202000%20SART%20speed%20accuracy.pdf
References
Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity And Inhibitory Control. Psychological Science, 8(1), 60-64.
Manly, T., Davison, B., Gaynord, B., Greenfield, E., Parr, A., Ridgeway, V., et al. (2004). An electronic knot in the handkerchief: "Content free cueing" and the maintenance of attentive control. Neuropsychological Rehabilitation, 14(1-2), 89-116.
Manly, T., Davison, B., Heutink, J., Galloway, M., & Robertson, I. (2000). Not enough time or not enough attention?: Speed, error and self-maintained control in the Sustained Attention to Response Test (SART). Clinical Neuropsychological Assessment, 3, 167-177.
Manly, T., Lewis, G. H., Robertson, I. H., Watson, P. C., & Datta, A. K. (2002). Coffee in the cornflakes: time-of-day as a modulator of executive response control. Neuropsychologia, 40(1), 1-6.
Manly, T., Owen, A. M., McAvinue, L., Datta, A., Lewis, G. H., Scott, S. K., et al. (2003). Enhancing the sensitivity of a sustained attention task to frontal damage: Convergent clinical and functional imaging evidence. Neurocase, 9(4), 340-349.
Manly, T., Robertson, I. H., Galloway, M., & Hawkins, K. (1999). The absent mind: Further investigations of sustained attention to response. Neuropsychologia, 37, 661-670.
O'Connor, C., Robertson, I. H., Manly, T., Hevenor, S. J., & Levine, B. (2003). Endogenous versus exogenous engagement of sustained attention: An fMRI study. Clinical Neuropsychologist, 17(1 - Supplement), 117.
Robertson, I. H., Manly, T., Andrade, J., Baddeley, B. T., & Yiend, J. (1997). 'Oops!': Performance correlates of everyday attentional failures in traumatic brain injured and normal subjects. Neuropsychologia, 35(6), 747-758.
Smallwood, J., Davies, J. B., Heim, D., Finnigan, F., Sudberry, M., O'Connor, R., et al. (2004). Subjective experience and the attentional lapse:Task engagement and disengagement during sustained attention. Consciousness and Cognition, 13, 657-690.