The development of attentional control mechanisms in multisensory environments

Abstract

Outside the laboratory, people need to pay attention to relevant objects that are typically multisensory, but it remains poorly understood how the underlying neurocognitive mechanisms develop. We investigated when adult-like mechanisms controlling one's attentional selection of visual and multisensory objects emerge across childhood. Five-, 7-, and 9-year-olds were compared with adults in their performance on a computer game-like multisensory spatial cueing task, while 129-channel EEG was simultaneously recorded. Markers of attentional control were behavioural spatial cueing effects and the N2pc ERP component (analysed traditionally and using a multivariate electrical neuroimaging framework). In behaviour, adult-like visual attentional control was present from age 7 onwards, whereas multisensory control was absent in all children groups. In EEG, multivariate analyses of the activity over the N2pc time-window revealed stable brain activity patterns in children. Adult-like visual-attentional control EEG patterns were present age 7 onwards, while multisensory control activity patterns were found in 9-year-olds (albeit behavioural measures showed no effects). By combining rigorous yet naturalistic paradigms with multivariate signal analyses, we demonstrated that visual attentional control seems to reach an adult-like state at ∼7 years, before adult-like multisensory control, emerging at ∼9 years. These results enrich our understanding of how attention in naturalistic settings develops.

Keywords: Attentional control; Development; Electrical neuroimaging; Multisensory; N2pc; Visual attention.

Fig. 1

Experimental trial sequence for our paradigm. In this example, the blue target diamond is preceded by a nontarget-colour (NCC), i.e., red, ‘cue’, both highlighted here by white circles (that did not appear in the experimental task). A spatially diffuse sound was presented together with the onset of the colour change cue (on 50 % of all trials), creating an audiovisual nontarget colour distractor (NCCAV) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

Fig. 2

Mean reaction times shown for each of the 4 age groups on trials where Cue-Target Location was the same versus different, shown separately for target colour-cue (TCC) and nontarget colour-cue (NCC) trials, as well as visual (V) and audiovisual (AV) trials. Line graphs show the mean RTs, bar graphs show error rates (in percentages), and error bars represent the standard error of the mean. The RT ranges that best display the Spatial Cueing effects variability in the data are displayed. Thus, each age group’s scale has a different range, but the range lengths are the same (200 ms), save for 5-year-olds where the variability was too large to maintain this range length.

Fig. 3

Bars coloured according to the figure legend in the image represent behavioural attentional capture indexed by mean RT spatial cueing effects, and error bars represent the standard error of the mean. Adults, 9-year-olds, and 7-year-olds all showed presence of top-down visual attentional control, exemplified by TAC. Specifically, all 3 age groups showed reliable attentional capture effects for target colour-cues, but not for nontarget colour-cues. In contrast, only in adults, attentional capture showed MSE.

Fig. 4

N2pc waveform results. Mean amplitude values are shown at contralateral and ipsilateral electrode sites, indicated in orange and black, per the head model and legend on the figure. The N2pc time-window of 180-300 ms is highlighted in light orange, where the contra-ipsi difference is significant, and light grey where it is not. Significance levels are denoted as follows: ** < .01, *** < .001. Adults show significant contra-ipsi differences, that is reliable N2pc’s, for target-colour cues (TCC) but not nontarget colour-cues (NCC). In children, there was no reliable N2pc in any of the four conditions.

Fig. 5

Scalp topography of the 4 lateralised difference template maps elicited over the N2pc time-window as a function of cue condition and observer age group. The four template maps resulting from the segmentation of the adult lateralised ‘mirrored’ difference ERP data are shown in the upper row. The bar graphs below represent each difference template map’s relative duration (% ms) over the N2pc time window, shown separately for the adults and the 3 younger groups, and for each of the V and AV cue conditions separately. Bars in the graphs are coloured according to their map’s backgrounds in the top row, and error bars denote the standard error of the mean. As visible in the lower graphs, Map 4 was the most dominant in adults, 9-year-olds, and 7-year-olds, while 5-year-olds did not have a clear map dominance pattern. Only in adults’ duration of Map 4 was modulated by cue type that is whether cue colour matched that of the target-colour.