Neuro-oscillatory mechanisms of intersensory selective attention and task switching in school-aged children, adolescents and young adults

Abstract

The ability to attend to one among multiple sources of information is central to everyday functioning. Just as central is the ability to switch attention among competing inputs as the task at hand changes. Such processes develop surprisingly slowly, such that even into adolescence, we remain slower and more error prone at switching among tasks compared to young adults. The amplitude of oscillations in the alpha band (~8-14 Hz) tracks the top-down deployment of attention, and there is growing evidence that alpha can act as a suppressive mechanism to bias attention away from distracting sensory input. Moreover, the amplitude of alpha has also been shown to be sensitive to the demands of switching tasks. To understand the neural basis of protracted development of these executive functions, we recorded high-density electrophysiology from school-aged children (8-12 years), adolescents (13-17), and young adults (18-34) as they performed a cued inter-sensory selective attention task. The youngest participants showed increased susceptibility to distracting inputs that was especially evident when switching tasks. Concordantly, they showed weaker and delayed onset of alpha modulation compared to the older groups. Thus the flexible and efficient deployment of alpha to bias competition among attentional sets remains underdeveloped in school-aged children.

Figure 1. The task

From trial to trial, Participants were visually cued to attend to either an auditory stimulus (by the illustration of the headphones) or a visual stimulus (by the illustration of the computer monitor) in a pseudorandom fashion. Cue stimuli (200 ms duration) were 100% valid, meaning a target stimulus (S2) always contained stimuli in the cued modality. The S2 stimuli be presented alone or accompanied by a stimulus from the un-cued modality. The auditory stimulus consisted in two 100 ms tones presented in rapid succession, the first at 1300 ms after the onset of the cue stimulus followed by the second after a 5 ms gap at 1405 ms. The visual S2 consisted of two bilaterally presented Gabors, presented at 1355 ms.

Figure 2. Preliminary wavelet analysis

Preliminary wavelet analysis broken out by age groups. Time-frequency plots reflect the subtraction of all Cue Visual conditions from all Cue Auditory condition. Waveforms depict the time course in the frequency range of 8–14 Hz for the Cue Auditory condition (blue) and Cue Visual condition (red). All units are in decibels. Note the different y-axes on the waveform plots across the three age groups.

Figure 3. D-prime data

(A) D-prime data for all conditions and participant groups. Error bars reflect standard error of the mean (SEM). (B) D-prime data collapsed across the Cue Auditory and Cue Visual conditions. (C) Switch costs (Repeat minus Switch) for the Cue Auditory conditions. (D) Switch costs for the Cue Visual conditions. Here the 8–12 age group shows a marked difference from the other two groups on incongruent trials. Error bars reflect standard error of the mean.

Figure 4. D-prime data collapsed across the age groups

(A) All conditions collapsed across the three age groups. (B) D-prime collapsed across both the three age groups and Switch and Repeat trials. (C) Switch costs, computed as Repeat minus Switch (greater switch costs are more positive), collapsed across the three age groups. Error bars reflect standard error of the mean.

Figure 5. Alpha power modulation and effects of task switching

(A) Topographic representations of Cue modality alpha power modulation (Cue Auditory minus Cue Visual) in decibels for the three age groups, the two time windows, and Switch and Repeat trials. (B) Topographic representations of task-switching related alpha power modulation (Switch minus Repeat) for each of the three age groups, the two time windows, and the two Cue conditions. (C) Error bars depicting alpha power modulation for the three age groups, the two time windows, and Switch and Repeat trials. Power modulation values were computed from the average over the left and right parieto-occipital sensor groups. Error bars reflect the standard error of the mean.

Figure 6. Relationship of alpha power to behavior

Scatter plots depicting the relationship of alpha power modulation (Cue Visual Repeat minus Cue Visual Switch) to the switch cost on Cue Visual Incongruent trials over all age groups. Left and right scatterplot panels reflect the left and right parietal-occipital sensor groups, respectively. Far right: the same relationship computed for each EEG sensor. Red “x”s indicate a significant correlation at the p < 0.05 level. Color in the topographic map codes for the r value of the each correlation.