Top-down modulation of early selective attention processes in children

Transcription

1 Ž. International Journal of Psychophysiology Top-down modulation of early selective attention processes in children M.J. Taylor a,,1, S.C. Khan a,2 a Di ision of Neurology and Research Institute, The Hospital for Sick Children, Uni ersity of Toronto, Toronto, Canada Received 20 July 1999; received in revised form 14 November 1999; accepted 16 November 1999 Abstract It has been demonstrated in adults that attention can modulate very early stages of perceptual cognitive processing, but it has not been determined if this capacity for modulation develops with age. We investigated early attentional processes in parallel and serial visual search tasks in 40 children 7 12 years old, using event-related potentials Ž ERPs.. Two single-feature pop-out tasks were used to assess parallel processing; a conjunction of features task was used to study serial processing. There were significant decreases in latencies with age for the peaks measured posterior P1, N1 and anterior P2 Ž P2a.. P1 and N1 latency effects also varied with trial type and were consistent with top-down modulation of processing, which differed between the parallel tasks. P2a amplitude for the hit trials decreased with increasing age for the colour and serial tasks. Increasing R L hemispheric asymmetries with age in P2a amplitude were seen on non-target trials for the size and serial tasks only, reflecting serial processing. These data demonstrate that developmental changes in visual selective attention for early stages of processing are continuing through childhood, and that there is top-down modulation at these latencies in children Elsevier Science B.V. All rights reserved. Ž. Keywords: Visual search; Event-related potential ERP ; Development; Selective attention Corresponding author. Tel.: ; fax: address: taylor@cerco.tlse ups.tlse Ž M.J. Taylor. 1 Present address: Centre de Recherche Cerveau and Cognition UMR 5549, Faculte de Medecine de Rangueil, Toulouse, France. 2 Present address: Mood Disorders Group, Department of Psychiatry, IWK Grace Health Centre, Halifax, Nova Scotia, Canada $ - see front matter 2000 Elsevier Science B.V. All rights reserved. Ž. PII: S

2 136 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Introduction There are currently several competing models of visual selective attention that distinguish between parallel and serial search Že.g. Treisman, 1988; Wolfe et al., 1989., which place more or less emphasis on the differences between the automatic and intentional aspects of visual search. As behavioural data can be interpreted to fit either model, some researchers have employed eventrelated potentials Ž ERPs. in the investigation of visual search processes, as ERPs can provide fine-grained indices of cognitive processing ŽPic- ton and Hillyard, In an early study utilising a traditional visual search paradigm, Luck and Hillyard Ž found evidence supporting Treisman s Ž model, when only late ERP components were measured. A more recent study using pop-out stimuli Žwhich require parallel processing. showed pop-outs to be associated with enhanced early components, suggesting automatic detection Ž Luck and Hillyard, They concluded that pre-attentive processes guide attention rapidly to a relevant object, but that the process was not entirely automatic, as there was evidence of top-down influence on components from P2 through to P3 and on RTs, consistent with the guided search model Ž Wolfe et al., Eals and Taylor Ž recorded ERPs to parallel and serial processing tasks, with the serial task consisting of a conjunction of features. The data did not divide according to parallel and serial tasks; colour pop-out task had the shortest RT and ERP latencies, followed by the size pop-out task, and finally the serial task, suggesting that parallel and serial processing are part of a continuum, also consistent with the guided search model of visual attention Ž Wolfe et al., There are numerous studies on the development of attention in children, but few have tested differing models of attention Žsee, however, Trick and Enns, The developmental changes in performance on attention tasks may be due to age-related increases in either attentional capacities or utilisation of strategies. In behavioural studies, there are increases in processing speed with age across a wide range of tasks Ž Kail, However, some investigators have not found evidence of developmental differences in parallel processing between children 7 8 years of age and adults Ž see Plude et al., 1994, for a review.. The controversy as to whether attentional processes are mature at a young age is due in part to the difficulty in measuring these processes without the contamination from other behaviours, such as motor responses, which show marked maturational changes. As ERPs assess cognitive processing independently of behavioural measures, they can determine the presence or absence of developmental changes in specifically attentional processes. Developmental studies in selective attention can also contribute to validating modelž. s based exclusively on adult data; if a model can also explain developmental data, then that proffers strong support for the model s validity ŽEnns, Thus, the current study was designed to investigate parallel and serial processing across the age range in which parallel processing is suggested to be mature, and to determine which model of visual selective attention could best account for the developmental findings. 2. Methods 2.1. Subjects Forty children, divided into three age groups: 7 8 years Ž mean years, n 13., 9 10 years Ž mean years, n 14. and years Ž mean years, n 13. were studied. All subjects were in the age-appropriate grade at school, had no history of neurological, behavioural or psychiatric disorder, and all had normal or corrected to normal eyesight. Thirty-seven subjects were right-handed, and one left-handed subject was in each age group. Subjects used their preferred hand for behavioural responses Žbutton presses.. There were eight males in the 7 8-yearold group, six in the 9 10-year-old group, and nine in the year-old group. The children had been recruited via friends and colleagues within the institution and local schools. All children were given the WRAT-R and abbreviated WISC-III Žblock design and vocabu-

3 M.J. Taylor, S.C. Khan International Journal of Psychophysiology lary. to ensure that all were in normal cognitive range for their age. None had to be excluded on this basis. The children were in the high normal to superior range, with the IQs being: 7 8 year olds, ; 9 10 year olds, ; and year olds, There were no significant differences in IQ among the age groups Procedure ERPs were recorded using a NeuroScan system from 27 electrodes ŽFp1, Fpz, Fp2, Fz, F3, F4, F7, F8, Fc1, Fc2, Fc3, Fc4, Cz, C3, C4, Pc3, Pc4, Pz, P3, P4, Oz, O1, O2, T3, T4, T5, T6.Ž see Fig. 2. using a cervical spinal reference. The electrode impedance was below 5 k. EOG was monitored with electrodes placed at the outer canthus and supra-orbital ridge of the left eye. The ERPs were recorded with a bandpass of Hz and a 1-s sweep, starting 50 ms prior to stimulus presentation. The EEG was averaged off-line and trials with artefact Ž 120 V. were rejected. Stimuli were presented centrally on a Macintosh computer. Each task consisted of 120 trials. Two parallel processing tasks were used. For both parallel tasks, the standard stimuli consisted of eight small blue vertical rectangles ŽFig. 1a; from Eals and Taylor Ž 1996., adapted from Luck and Hillyard Ž In the colour task, target trials were arrays where one of the rectangles was red Ž P In the size task, target trials were arrays which contained a large blue square Ž P In both tasks the other pop-out stimuli were included as non-targets i.e. colour pop-out stimuli were presented in the size task Ž P and the size pop-out stimuli were presented within the colour task Ž P 0.17., both as non-targets pop-outs. For the serial task, stimuli were eight multicoloured squares and rectangles which differed in size and colour ŽFig. 1b; from Eals and Taylor Ž The targets were arrays which contained an item with a conjunction of features from the parallel tasks, i.e. large red squares. Targets for the serial task occurred on 35% of trials. During each trial, a white rectangle Ž 4 6. was presented in the centre of the screen 300 ms before the stimulus array appeared. The stimulus array was presented within the rectangle for 400 ms and the rectangle remained on for another 1 s. The interval between trials was 1.5 s. Target standard and non-target pop-out trials were randomly interspersed. Subjects responded to target trials by pressing a button as quickly as possible. The tasks were explained using coloured figures of the types of stimuli, and practice trials were not used. Subjects made few errors and these did not vary with age Žmean accuracy: 7 8 years, 92.2%; 9 10 years, 92.4%; years, 89.4%. or task Žmean accuracy: colour, 87.5%; size, 92.8%; serial, 93.7%.; trials with incorrect behavioural responses were rejected from average files. Grand averages were compiled for target and non-target standard trials for each task and for the non-target pop-outs for the two parallel tasks, for each age group. These eight grand averages per age group then served as templates for peak detection within each child s data. The P1 and N1 peaks were measured at posterior electrodes ŽOz, O1, O2, Pz, P3, P4, T5, T6, Pc3, Pc4.; P2a was measured at the anterior electrodes ŽFpz, Fp1, Fp2, Fc1, Fc2, Fc3, Fc4, F7, F3, Fz, F4, F8, Cz, C3, C4, T3, T4.. All peaks were measured within 50 ms of the peak s latency in the appropriate grand average. Peak latencies and amplitudes Ž from the pre-stimulus baseline. were measured for targets and for the standard stimuli of the three tasks and for the non-target pop-outs in the two parallel tasks. The data were submitted to repeated measures ANOVAs Žusing Greenhouse Geisser corrections. and analysed as a function of age, task and electrodes and then as a function of hemisphere Žusing the lateral electrodes.. The amplitudes were analysed first for main effects and then the transformed amplitudes Ž transformed to values between 0 and 1. were analysed for interactions ŽMcCarthy and Wood, Results Fig. 2 shows an example of the target and non-target standard ERPs recorded across electrode sites for the 7 8-year-olds for the colour task. The peaks measured from the anterior elec-

4 138 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Fig. 1. Ž. a Examples of stimuli used for the two parallel tasks, showing a standard stimulus on the left and an example of a colour pop-out Ž target for the colour task. on the upper right and an example of the size pop-out Ž target for the size task. on the lower right. Ž b. An example standard Ž left. and target Ž right. stimuli for the serial task. Stimulus arrays were presented within a white horizontal rectangle on a black background. The spatial distribution, and colours of items in the arrays in the serial task, varied from trial to trial.

5 Ž. Ž. Fig. 2. Grand averaged ERPs recorded from the 7 8 year olds showing the ERPs to the target dark grey lines and standard light grey lines stimuli for the colour task, for the 27 electrodes. The peaks measured are indicated: P1 and N1 on Oz; P2a on Fz. P1 and N1 were measured at the 10 posterior electrodes sites, while the P2a was measured at the 17 anterior electrodes. M.J. Taylor, S.C. Khan International Journal of Psychophysiology 37 ( 2000 )

6 140 M.J. Taylor, S.C. Khan International Journal of Psychophysiology trodes Ž P2a. and the posterior electrodes ŽP1 and N1. are indicated P1 For P1 latency there were significant age ŽF 2, , P , trial type ŽF7, , P 0.015, , electrode ŽF9, , P , and age electrode ŽF 18, , P 0.005, effects. P1 latency decreased with age Ž Fig. 3. was shortest for the colour targets, and longest for the serial standard stimuli Ž Table 1, Fig. 4.. The colour and serial task targets had shorter P1 latencies than their respective standard non-target stimuli ŽF1, , P 0.01; F1, , P 0.05, respectively.. When the pop-out stimuli were compared when they were or were not targets, there was a significant effect for the colour pop-out stimuli ŽF 1, , P , due to shorter latencies for those stimuli when they were targets Ž Fig. 4.. For P1 amplitude there were significant age Ž F 4.47, P , trial type Ž 2,34 F7, , P 0.027, and electrode ŽF9, , P , main effects. The age effect was due to larger P1s in the Table 1 Mean peak latencies for targets and non-targets for all three tasks, and non-target pop-outs for the colour and size tasks in ms Ž S.D.. Stimulus type Peak Task 7 8 years 9 10 years years Targets P1 Colour 140 Ž Ž Ž Serial 148 Ž Ž Ž Size 144 Ž Ž Ž N1 Colour 200 Ž Ž Ž Serial 212 Ž Ž Ž Size 212 Ž Ž Ž P2a Colour 232 Ž Ž Ž Serial 233 Ž Ž Ž Size 235 Ž Ž Ž Non-target P1 Colour 149 Ž Ž Ž standard Serial 155 Ž Ž Ž stimuli Size 146 Ž Ž Ž N1 Colour 208 Ž Ž Ž Serial 223 Ž Ž Ž Size 221 Ž Ž Ž P2a Colour 232 Ž Ž Ž Serial 243 Ž Ž Ž Size 234 Ž Ž Ž Non-target P1 Colour 149 Ž Ž Ž pop-outs Size 147 Ž Ž Ž N1 Colour 210 Ž Ž Ž Size 224 Ž Ž Ž P2a Colour 221 Ž Ž Ž Size 223 Ž Ž Ž 13.8.

7 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Fig. 3. The main effect of age on the latencies of the three components measured; all three decrease significantly in latency with increasing age. Ž F 12.11, P Ž Fig. 3. 2,35, due to age effects in the target and non-target standard trials ŽF2, , P 0.001; F2, , P , respectively; see Table 2.. Significant overall trial type effects Ž F 4.26, P 0.001, ,245 were seen due to the longer latencies for the standard trials in the serial and size tasks ŽTable 1.. P2a amplitude showed age ŽF2, , P and trial type ŽF7, , P 0.002, effects. The age effect was due to marginally decreasing P2a amplitudes with age across trial types, but when trial types were analysed separately, age was only significant for target trials Ž F 3.57, P ,35. The trial type effect was due to lower P2a amplitudes in the non-target trials compared to the target trials year-olds. The trial type effect showed P1 amplitude was larger for targets Ž 7.2 V. and non-target pop-outs Ž 6.6 V. than standard Ž 4.9 V. trials Ž F 16.38, P Ž Fig. 5. 1,34. The pop-out stimuli had comparable P1 amplitudes regardless of whether they were targets or non-targets. There were no hemispheric effects N1 For the posterior N1 component there were significant decreases in latency with age ŽF2, , P Ž Fig. 3. and across trial types Ž F 4.62, P , ,238. The latter effect was due to shorter latencies for the colour pop-outs, both when they were targets compared to other target stimuli Ž190 ms vs. 199 ms; F2, , P , , and when they were non-targets, compared to size non-target pop-outs Ž 192 ms vs. 202 ms; F 5.08, P ,35. When the pop-out trials were compared, there was an age task interaction for the size pop-outs ŽF 2, , P , due to a steeper slope for the age-related decreases in latency for the non-target size pop-out stimuli. There were no significant amplitude effects Ž Table 2. nor hemispheric asymmetries in latencies or amplitudes P2a P2a latency decreased significantly with age Fig. 4. Mean latencies of the P1 component across the eight trial types. As there were no interactions with age, these latencies are averaged across age groups; an asterisk marks significant differences between colour pop-outs, whether they were targets Ž far left. or non-targets, and between the targets and standard non-targets in the colour and serial tasks.

8 142 M.J. Taylor, S.C. Khan International Journal of Psychophysiology older age groups ŽF4, , P 0.042, For the reaction times, significant age ŽF2, , P and task ŽF2, , P main effects were found. The RTs decreased with increasing age, and across tasks from colour size serial, but the rate of change with age did not vary with the task. 4. Discussion Fig. 5. Mean amplitudes of the P1 component across the eight trial types. The target trials Ž on the left. had significantly larger amplitudes than the non-target standard trials Žin the middle.. There was no difference between the pop-out trials, as a function of target non-target status. Ž Table 2.. For the target trials there was also an age task interaction ŽF4, , P 0.029, , due to decreasing amplitudes with age in the colour and serial tasks, but little change as a function of age in the size task Ž Fig. 6.. When the seven lateral electrodes over each anterior hemisphere were analysed ŽLH-elec- trodes: Fp1, F7, F3, Fc1, Fc3, T3, C3; RH-electrodes: Fp2, F8, F4, Fc2, Fc4, T4, C4., there was a significant task hemisphere interaction for the non-target standard trials in the three tasks ŽF 2, , P 0.007, which remained significant with transformed amplitudes ŽF2, , P 0.003, This asymmetry was due to a larger right than left P2a for the nontarget trials in the size and serial tasks only ŽFig. 7.. This asymmetry was more marked in the two This study found significant age and task effects in the early ERP components recorded in children, performing parallel and serial visual search tasks. The age-related decreases in latency are consistent with those studies in the literature that have reported developmental improvement in attentional processes, but the finding of effects as early as the P1 is particularly interesting. P1 is reported as the earliest index of endogenous processing of visual stimuli ŽMangun, It is well established in the ERP literature that this component is sensitive to attention, although only amplitude effects are reported with manipulations in spatial attention in adults ŽMangun and Hillyard, 1988; Hillyard and Anllo- Vento, The significant latency decreases in this component with increasing age demonstrate that this early stage of processing is still maturing, with shorter latencies interpreted to index more efficient processing Ž Cerella and Hale, Also, these data confirm the classification of the P1 as an endogenous component, in contrast to the slightly earlier exogenous P100 component, which does not show maturational change at these ages Ž Taylor and McCulloch, More importantly, the shorter latency for P1 to colour pop-out stimuli when they were targets compared to exactly the same stimuli when they were non-targets, strongly suggests the presence of top-down modulation at this early stage of processing. As this did not interact with age, it demonstrates the presence of this top-down influence as young as 7 8 years of age. Having a sample size of 40 may also have been critical in seeing this effect, as it has not been reported in

9 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Table 2 Mean peak amplitudes for target and non-targets for all three tasks, and non-target pop-out stimuli for the colour and size tasks in V Ž S.D.. Stimulus type Peak Task 7 8 years 9 10 years years Targets P1 Colour 6.36 Ž Ž Ž 8.5. Serial 8.3 Ž Ž Ž 8.0. Size 5.62 Ž Ž Ž 9.3. N1 Colour 5.49 Ž Ž Ž 7.1. Serial 6.11 Ž Ž Ž 7.5. Size 7.69 Ž Ž Ž 8.5. P2a Colour Ž Ž Ž Serial Ž Ž Ž 7.3. Size 9.72 Ž Ž Ž 7.5. Non-target P1 Colour 4.24 Ž Ž Ž 6.7. standard Serial 4.6 Ž Ž Ž 8.7. stimuli Size 2.05 Ž Ž Ž 8.2. N1 Colour 5.54 Ž Ž Ž 6.4. Serial 4.29 Ž Ž Ž 7.3. Size 6.35 Ž Ž Ž 6.3. P2a Colour 6.97 Ž Ž Ž 5.3. Serial 7.4 Ž Ž Ž 6.4. Size 4.07 Ž Ž Ž 5.7. Non-target P1 Colour 5.97 Ž Ž Ž 8.5. pop-outs Size 4.25 Ž Ž Ž 9.6. N1 Colour 7.14 Ž Ž Ž 7.9. Size 9.03 Ž Ž Ž 7.6. P2a Colour 6.24 Ž Ž Ž Size 5.26 Ž Ž Ž 7.8. adult studies, where the number of subjects is typically subjects. When the children were watching for and responding to a red pop-out target stimulus within an array, they processed the information faster, than when the same red stimuli appeared when they were watching for and responding to a large blue square Žthe size task.. This is, to our knowledge, the first report of top-down modulation at such an early latency in children. P1 amplitude was larger for target than standard stimuli, as seen in the adult ERP literature Ž. Ž. e.g. Han et al., Eimer 1997 found that in sustained attention tasks, attention to colour enhanced early positive peaks, but these were seen in difference potentials and measured anteriorly. Many other studies have shown posterior colour effects. For example, Lange et al. Ž found a small early parietal effect of attending to colour, which was distinct from the larger effects of location selection. Allison et al. Ž found colourspecific effects in a colour adaptation paradigm in intracranial recordings, beginning on the initial positive component Ž at approx. 120 ms. at medial locations on the posterior fusiform gyrus. In a colour selective attention paradigm, the earliest

10 144 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Fig. 6. Grand averaged ERPs showing the frontal P2a components to the target trials for the three tasks, for each age group. Note the decreasing amplitude with age for the colour Ž thin line. and serial Ž dotted line. task targets, but no significant changes with age for P2a amplitude to targets for the size task Ž thick line.. reported amplitude attention effects were seen at 130 ms Ž Anllo-Vento et al., 1998., with the suggested source in medial occipital gyrus, consistent with Corbetta et al. Ž Thus, our P1 amplitude data are in accordance with these the findings of early attention modulation with colour stimuli in posterior brain regions, and show the presence of this effect in children. In our data, however, the amplitude did not differ between Fig. 7. The task hemisphere interaction for P2a amplitudes for the non-target, standard stimuli, showing decreasing amplitudes over the left hemisphere ŽLH-electrodes: Fp1, F7, F3, Fc1, Fc3, T3, C3. vs. the right hemisphere ŽRH-electrodes: Fp2, F8, F4, Fc2, Fc4, T4, C4. for the serial and size tasks. target and non-target pop-out stimuli, suggesting that the non-target pop-outs capture comparable attention as do the targets in our young subjects. As we did not manipulate directed spatial attention in our tasks, those amplitude effects well established in the adult literature could not be studied. The shorter latencies for processing the attended colour stimuli were also visible in the N1, as the colour pop-outs had shorter N1 latencies than either other targets or the size pop-out nontargets. For the size pop-out stimuli, there was also a significant target non-target latency effect which occurred at N1. As the size pop-out stimuli were again the same set and only their status as targets changed between the two tasks, it also suggests that top-down modulation during task performance produced this interaction. Thus, for the size task, it appears discriminative processes started at the N1 latencies, later than was seen for the processing associated with the colour task, at P1 latencies. Our data show that the use of size as a single-feature task is not comparable to the single feature of colour in children; these two parallel tasks are processed differently by children, with colour appearing more salient as it yielded the shortest latencies. These task differences are consistent with our interpretation of the target non-target effects seen in the P2a. The anterior P2 peak Ž P2a. is associated with

11 M.J. Taylor, S.C. Khan International Journal of Psychophysiology stimulus evaluation and attentional demands in adults Ž e.g. Luck and Hillyard, The latency decrease with age is consistent with other cognitive developmental ERP studies which show that with increasing age children become faster or use more efficient strategies for evaluating visual stimuli, with some of the more pronounced effects being seen at anterior sites ŽStauder et al., 1993; Taylor and Smith, The shorter P2a latencies for the colour pop-out target stimuli seen in adults Ž Luck and Hillyard, 1994., were not seen in children, despite shorter P1 and N1 latencies to these stimuli. The longer latencies for the standard non-targets in both the serial and size tasks suggest that these stimuli were processed slower, serially, and in distinction to the colour task standard stimuli. As the standard stimuli were identical in the colour and size tasks, this is consistent with the salience of the singleton feature of colour; it also argues that both the size and serial tasks were processed as conjunction of features in children Ž Muller et al., P2a amplitude showed a slight decrease with age and higher amplitudes in the target compared to the non-target trials. For the target trials there was also an age task interaction, due to decreasing amplitudes with age only in the colour and serial tasks Ži.e. those tasks using colour as a defining feature of targets.; for these simple target stimuli, this may reflect the increasing ease of identification. For the non-target trials there was a task hemisphere interaction, when the lateral electrodes over the anterior hemispheres were analysed. This asymmetry was due to a larger right than left P2a for the size and serial tasks only Ž Fig. 7., and was more marked in the two older age groups. Our finding of decreased anterior positive peaks with age for non-target trials in the size and serial tasks, suggest increasing frontal processing negativity ŽKarayanidis and Michie, with age for visual information which requires serial processing. These data would argue for increasing use of left frontal areas for serial attentional processes. Several models of attention converge on the issues of frontal lobe involvement in attentional processing Ži.e. Stuss et al., 1995., but in children this has not been as well studied Ž Smith et al., Physiological measures such as P2a contribute to our understanding of the role and development of anterior attention functions, by demonstrating, systematic changes with age depending upon the attentional demands. The data in the current study are consistent with the Guided Search model of attention ŽWolfe et al., 1989., which explains visual selective attention as an interaction between top-down and bottom-up processes. Our data showed evidence of top-down processing in both parallel and serial tasks in children, as the latencies were faster for stimuli when those stimuli were designated targets, than when the same stimuli were presented as non-targets. The relative ease with which the children performed the serial task is also explained by the Guided Search model: the large red square was visually salient and bottom-up perceptual processes can help guide attention quickly to its location, reducing the need for serial search. In behavioural studies the feature of colour appears more effective at separating targets from non-targets Ž Cave and Wolfe, 1990., and when subjects have prior knowledge of which dimension is critical for target detection, considerable top-down control can be exercised Ž Muller et al., In the present data, targets discriminated by colour had shorter P1 latencies, suggesting top-down influences for this salient feature. Serial search was requisite, however, for non-target trials Ž Chun and Wolfe, in the size and serial tasks, and these showed a distinct pattern of P2a results: longer P2a latencies, decreased amplitudes with age and frontal asymmetries in the two older groups of children. Behavioural measures showed only the expected age and task main effects: shorter latencies in the older children, and longer latencies for the serial task, shortest for the colour task ŽTaylor et al., Behavioural studies in the literature have also found the ubiquitous behavioural latency decreases with age and task ŽShepp and Barrett, 1991; Trick and Enns, The interest of this study is the sensitivity of the ERPs, compared to behavioural measures, in assessing early stages of cognitive processing in visual search tasks. The ERPs demonstrated that even in the very simple colour parallel task, developmental

12 146 M.J. Taylor, S.C. Khan International Journal of Psychophysiology changes in the early cortical processing of the stimuli are continuing. The age-related decreases in latency are consistent with those studies in the literature that have reported developmental improvement in attentional processes, but the finding of significant effects as early as the P1 is important to theories of attention and its development. 5. Conclusions These ERP data show that: Ž. 1 early attentional processes for parallel or serial processing are not mature by the early school-age years; Ž. 2 as seen in adults Ž Muller et al., 1995., considerable top-down modulation occurs in children in both parallel and serial processing, and the latency of the effects increase with task difficulty Žat P1 latency for the colour pop-outs, at N1 latency for the size pop-outs and at P2a latency for serial processing.; and Ž 3. that the Guided Search model of attention can best explain these findings. Acknowledgements This research was supported by Medical Research Council of Canada. References Allison, T., Begleiter, A., McCarthy, G., Roessler, E., Nobre, A.C., Spencer, D.D., Electrophysiological studies of color processing in human visual cortex. Electroencephalogr. Clin. Neurophysiol. 88, Anllo-Vento, L., Luck, S.J., Hillyard, S.A., Spatio-temporal dynamics of attention to color: Evidence from human electrophysiology. Hum. Brain Mapping 6, Cave, K.R., Wolfe, J.M., Modeling the role of parallel processing in visual search. Cogn. Psychol. 22, Cerella, J., Hale, S., The rise and fall in informationprocessing rates over the life span. Acta Psychol. 86, Chun, M.M., Wolfe, J.M., Just say no: how are visual searches terminated when there is no target present? Cogn. Psychol. 30, Corbetta, M., Miezin, F.M., Dobmeyer, S., Shulman, G.L., Petersen, S.E., Selective and divided attention during visual discriminations of shape, color, and speed: functional anatomy by positron emission tomography. J. Neurosci. 11, Eals, M., Taylor, M.J., ERPs reflecting parallel and serial processing of colour arrays. Electroencephalogr. Clin. Neurophysiol. 46 Ž Suppl.., Eimer, M., An event-related potential Ž ERP. study of transient and sustained visual attention to color and form. Biol. Psychol. 44, Enns, J.T., What can be learned about attention from studying its development? Can. Psychol. 34, Han, S., Fan, S., Chen, L., Zhuo, Y., On the different processing of wholes and parts: a psychophysiological analysis. J. Cogn. Psychol. 9, Hillyard, S.A., Anllo-Vento, L., Event-related brain potentials in the study of visual selective attention. Proc. Natl. Acad. Sci. 95, Kail, R., Processing time decreases globally at an exponential rate during childhood and adolescence. J. Exp. Child Psychol. 56, Karayanidis, F., Michie, P.T., Frontal processing negativity in a visual selective attention task. Electroencephalogr. Clin. Neurophysiol. 99, Lange, J.J., Wijers, A.A., Mulder, L.J.M., Mulder, G., Color selection and location selection in ERPs: differences, similarities and neural specificity. Biol. Psychol. 48, Luck, S.J., Hillyard, S.A., Electrophysiological evidence for parallel and serial processing during visual search. Percept. Psychophys. 48, Luck, S.J., Hillyard, S.A., Electrophysiological correlates of feature detection during visual search. Psychophysiology 31, Mangun, G.R., Neural mechanisms of visual selective attention. Psychophysiology 32, Mangun, G.R., Hillyard, S.A., Spatial gradients of visual attention: behavioural and electrophysio-logical evidence. Electroencephalogr. Clin. Neurophysiol. 70, McCarthy, G., Wood, C.C., Scalp distributions of eventrelated potentials: an ambiguity associated with analysis of variance models. Electroencephalogr. Clin. Neurophysiol. 62, Muller, H.J., Heller, D., Ziegler, J., Visual search for singleton feature targets within and across feature dimensions. Percept. Psychophys. 57, Picton, T.W., Hillyard, S.A., Endogenous event-related potentials. In: Picton, T.W. Ž Ed.., EEG Handbook, vol. 3. Elsevier Science Publishers, Amsterdam, pp Plude, D.J., Enns, J.T., Brodeur, D., The development of selective attention: a life-span overview. Acta Psychol. 86, Shepp, B.E., Barrett, S.E., The development of perceived structure and attention: Evidence from divided and selective attention tasks. J. Exp. Child Psychol. 51,

13 M.J. Taylor, S.C. Khan International Journal of Psychophysiology Smith, M.L., Kates, M.H., Vriezen, E.R., The development of frontal-lobe functions. In: Segalowitz, S.J., Rapin, I. Ž Eds.., Handbook of Neuropsychology. Elsevier Science Publishers, Amsterdam, pp Stauder, J.E.A., Molenaar, P.C.M., van der Molen, M.W., Scalp topography of event-related brain potentials and cognitive transition during childhood. Child Dev. 64, Stuss, D.T., Shallice, T., Alexander, M.P., Picton, T.W., A multidisciplinary approach to anterior attentional functions. Structure and functions of the human prefrontal cortex. Ann. N.Y. Acad. Sci. 769, Taylor, M.J., Malone, M.A., Khan, S.C., Parallel and serial attentional processes: a developmental ERP study. Dev. Neuropsychol. 15, Taylor, M.J., McCulloch, D.L., Visual evoked potentials in infants and children. J. Clin. Neurophysiol. 9, Taylor, M.J., Smith, M.L., Maturational changes in the ERPs to verbal and nonverbal memory tasks. J. Psychophysiol. 9, Treisman, A., Features and objects: the fourteenth Bartlett memorial lecture. Q. J. Exp. Psychol. A 40, Trick, L.M., Enns, J.T., Lifespan changes in attention: the visual search task. Cogn. Dev. 13, Wolfe, J.M., Cave, K.R., Franzel, S.L., Guided search: an alternative to the feature integration model for visual search. J. Exp. Psychol. 15,