Journal of Abnormal ChildPsychology, Vol. 16, No. 6, 1988, pp. 641-655
Controlled Processing and Vigilance in Hyperactivity: Time Will Tell Jaap van der Meere 1 and J o s e p h Sergeant 2
This paper reviews the concept o f sustained attention, placing it within a theoretical framework in which deficits of attention are conceived of as deficits of controlled information processing. Two types of deficit o f sustained attention are distinguished: perceptual sensitivity and perceptual criterion. These two deficits are linked to a model of human performance that links controlled processes to the energetic pools: arousal and activation. Perceptual sensitivity (d/) deficits are said to reflect arousal deficiencies, especially when observed in the early period o f a vigil. Perceptual criterion deficits are associated with the activation pool and the response criterion measure 3. Despite clear evidence of perceptual deficiency in the hyperactive children to a greater extent than in the control group, and that performance in d ~ declined with time on task, a significant interaction failed to occur between group classification and time on task. Thus, the results failed to support the hypothesis o f a sustained attention deficit in hyperactives, since if hyperactives have a sustained attention deficit, both d ~ and 3 shouM have shown a significantly greater decline in the hyperactive group than in the controls with time on task.
INTRODUCTION There have been a n u m b e r o f attentional deficits thought to be associated with attention deficit disorder with hyperactivity ( A D D H ) children with a SUSManuscript received in final form March 28, 1988. The authors wish to thank L. Leertouwer for making the drawings. This research was supported by grants from the Netherlands Foundation for the Advancement of Pure Research (ZWO) and the Professor Duijker Fund. ~Laboratory of Experimental Clinical Psychology, University of Groningen, Turfsingel 46, 9712 KR, Groningen, The Netherlands. 2Address all correspondence to Dr. Joseph Sergeant, Department of Clinical Psychology, University of Amsterdam, Weesperplein 8, 1018 XA Amsterdam, The Netherlands.
641 00914]627/88/12004)641506.00/0 9 1988 Plenum Publishing Corporation
van der Meere and Sergeant
tained attention deficit as the most prominent (Douglas & Peters, 1979; Douglas, 1983). This conclusion has been reached on the basis of research mainly with various forms of the continuous performance task (CPT). In this paper we first review the concept of sustained attention and distinguish two types of vigilance deficit: perceptual sensitivity and response criterion. The former is associated with limitations in working memory (Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977; Fisk & Schneider, 1981), while the latter is associated with response bias or strategy (Swets & Kristofferson, 1970). For either type of vigilance deficit to be associated with sustained attention, we will argue that a decline in performance over time is required. We will make a number of theoretical distinctions intended to bring greater conceptual clarity to the nature of the attentional deficit(s) associated with ADDH. A link will be made between studies of vigilance and studies using working memory as the locus of limitations in human information processing. A brief review of sustained attention studies in ADDH will be presented. Finally, we report a study of sustained attention of ADDH children using a paradigm derived from the theoretical framework of Fisk and Schneider (1981). The results will be discussed in the context of an energetic model of human information processing (Sanders, 1983). Theoretical Considerations
Most clinical descriptions of ADDH refer to three types of attention deficit: divided attention, focused attention, and sustained attention. All three have a common feature: They reflect deficits of controlled information processing. This assertion is derived from an influential model of attention (Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977; Fisk & Schneider, 1981). This model states that limitations of attention are due to restrictions in the rate at which information can be processed in working memory. It is because information must be processed slowly, serially, and with effort that attention is required to be directed to meet specific, relatively novel task demands. When task demands are well learned, limitations of controlled processing are bypassed through automatic processing. Since it has been empirically demonstrated that sustained attention deficits are more substantial in tasks requiring controlled processing than in tasks using relatively automatic processing (Fisk & Schneider, 1981), this distinction between controlled and automatic processing is crucial in that the establishment of sustained attention deficits in ADDH requires the use of controlled processing tasks. It may appear that there is a considerable gap between sustained attention and tasks involving controlled processing by working memory. This is understandable when memory is viewed as a store of past events. However, research in human cognition and performance has shown that working memory is an active and dynamic system (Broadbent, 1971). The link between vigilance and controlled processing may become clearer when memory operations in the CPT are considered.
Sustained Attention and Hyperactivity
There are two general C P T paradigms. In the first, the subject is given a constant target to retain in working memory and is instructed to press a button when the letter (or number) defined as the target appears in a train o f stimuli. Note that the subject retains the target in working memory, each stimulus is encoded, and a comparison occurs between the presented stimulus and the target. A decision is made, followed by a response. The second C P T paradigm differs from the first by not defining a constant target. The subject is instructed to press the button when the previously appearing stimulus is the same as the current stimulus. The subject encodes a stimulus, retains it in working memory, encodes the succeeding stimulus, and compares the current with the prior stimulus. If a match occurs between the two in working memory, a decision is reached and a response is made. This paradigm requires more controlled processing than the former since the subject is continuously required to update working memory by defining new relations that do not remain constant over time. In the first paradigm, given sufficient time and trials, the subject can learn the task so that relatively automatic processing may develop. How subjects perform over time is the subject o f vigilance research. The concept of sustained attention is an apparently simple one, but it deserves careful definition. Recently, vigilance theorists have defined decrements in sustained attention as a "progressive decline in performance" (Parasuraman, 1984; Warm, 1984). This reminds us that sustained attention is not meant to refer to constant allocation of attention but to the decline in human performance over time. Hence, it is not a group main effect alone in the C P T that allows statements concerning differences in maintenance of attention. Only if groups differ with respect to the amount of decline in performance over time can statements regarding differences in maintenance in attention be justified. Studies by Parasuraman (1979, 1984) and Parasuraman and Davies (1977) suggest that a decline in performance over time is the result either o f a reduction in discriminating between signals and nonsignals or of changes in the response bias in the decision system. Both aspects may be expressed in terms o f the signal detection theory measures perceptual sensitivity (d z) and response criterion (~)3 (Green & Swets, 1966). Which factor is responsible for the performance decline depends on task parameters. A decline in perceptual sensitivity, d z, with no change in response criterion, ~5, over time is obtained if the subject is required to discriminate a target from a nontarget represented in working memory, and at a high signal rate (event rate). In con3dZ(perceptual sensitivity) refers to the distance between a hypothetical noise distribution a n d a signal distribution embedded in noise. The distance between the two means provides a measure of perceptual sensitivity of the sensory system. The nonparametric measure of perceptual sensitivity P(A) is analogous to d ~ but does not require the distribution assumptions of d / (Davies & P a r a s u r a m a n , 1982). The response bias measure ~ is the overall or marginal probability of a "yes" target response. This probability lies between 0 and I and reflects the willingness of the observer to say "yes." To calculate the marginal probability of a "yes" response, the average of the conditional probabilities of a "yes" response given signal and nonsignal trials is weighted by the probability of occurrence of the signal and the nonsignal trials.
van der Meere and Sergeant
trast, if few working memory processes are involved or the event rate is low, the performance decrement is then thought to result from changes in decision criterion/3, with no change in d / over time. Different interpretations have been given to d / and ~. In signal detection theory, a decline in perceptual sensitivity d / and response criterion ~ refers, respectively, to an inability to detect targets and to changes in the response criterion (motivation) as task duration progresses (Swets & Kristofferson, 1970). In the so-called stage approach to attention (Sternberg, 1969; Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977), performance is thought to be constrained by the rate at which input, central, and output processing stages are influenced by different computational demands. The computational demand at the input stage is influenced by stimulus degradation, the central stages are affected by the level of memory...