Acute aerobic exercise benefits allocation of neural resources related to selective attention

Abstract

A single session of aerobic exercise has been shown to potentially benefit subsequent performance in a wide range of cognitive tasks, but the underlying mechanisms are still not fully understood. In this study, we investigated the effects of exercise on selective attention, a cognitive process that involves prioritized processing of a subset of available inputs over others. Twenty-four healthy participants (12 women) underwent two experimental interventions in a random, crossover, and counterbalanced design: a vigorous-intensity exercise (60-65% HRR) and a seated rest (control) condition. Before and after each protocol, participants performed a modified selective attention task that demanded attending stimuli of different spatial frequencies. Event-related magnetic fields were concurrently recorded using magnetoencephalography. The results showed that exercise, relative to the seated rest condition, reduced neural processing of unattended stimuli and increased processing of attended stimuli. The findings suggest that changes in neural processing related to selective attention may be one of the mechanisms underlying exercise-induced improvements in cognition.

Figure 1

Schematic illustration of the selective attention task used in the study. Participants were presented with high and low spatial frequency (SF) Gabor gratings in each of the two runs. In an "attend high SF run," participants detected targets of slightly higher SF than corresponding high SF standards. In an "attend low SF run," participants detected targets of slightly lower SF than corresponding low SF standards. Detecting targets of particular SF was assumed to demand directing more attentional resources to standards of corresponding SF (attended standards) than for standards of distinct SF (unattended standards). The difference between attended and unattended standards was the main contrast of interest in the neural data analyses. Each trial started with a fixation dot followed by the stimulus and an instruction to respond. Participants responded with one finger to targets and another to non-target stimuli.

Figure 2

The study diagram.

Figure 3

Time course and topographies of the spatiotemporal cluster showing a significant main effect of TASK (attended vs. unattended stimuli) at both pre-test runs, i.e., before the experimental EX or REST protocols. As expected, attended stimuli evoked significantly stronger neural activation in a widely distributed spatiotemporal cluster starting around 90 ms after stimulus onset. There were no significant clusters with relatively stronger neural activation for unattended stimuli.

Figure 4

Left Top: Topography of strongest neural difference activities between attended and unattended stimuli in both baseline runs (i.e., before exercise and before rest). An occipito-parietal region of interest was defined by estimated neural sources showing at least 50% of the maximal difference. Effects remained qualitatively identical when regions of interest with 60%, 70%, or 80% of the maximal difference were chosen (small topography shows 80% region of interest) and when taking into consideration the overall difference between attended and unattended stimuli (i.e., before and after exercise and rest). Bottom: Time courses of the global power of estimated neural activities within the bilateral occipito-parietal regions of interest as defined above (difference > 50%) for the Exercise (left) and Rest (right) conditions. Based on a sliding window analysis of neural activity within these regions of interest, attended stimuli (blueish colors) evoked stronger neural activity than unattended stimuli (reddish colors) before (bright colors) and after (dark colors) the protocols, starting around 90 ms after stimulus onset in the left hemisphere. Right Top: Relative change of neural activity from pre- to post-protocol between 90 and 600 ms in the 50% region of interest. While the pre-test (“normal”) relation of neural activity (i.e., attended > unattended) relatively increased after exercise (i.e., a relatively stronger increase of attended vs. unattended), this relation decreased after rest. Error bars denote 95% confidence intervals.

Figure 5

left: Topography (back view) of the spatiotemporal cluster showing a significant 3-way PROTOCOL*TIME*TASK interaction at bilateral occipito-parietal cortex regions between 380 and 590 ms. Center: In this cluster, independent of TIME (pre/post) or PROTOCOL (exercise/rest), attended stimuli evoked stronger neural activity than unattended stimuli (i.e., the main effect of TASK). Right: While the pre-test (“normal”) relation of neural activity (i.e., attended > unattended) became relatively strengthened after exercise (i.e., a relatively stronger increase of attended compared to unattended), this relation decreased after rest. Error bars denote 95% confidence intervals.