Functionally distinct alpha components are differentially modulated by attention and affect behavior

Alpha oscillations (~10 Hz) have been hypothesized to gate information flow in the brain by inhibiting task-irrelevant and disinhibiting task-relevant processing. Recent work using independent component analysis (ICA) showed that functionally distinct alpha components co-exist during a working-memory task. Here, we recorded magnetoencephalography (MEG) while participants performed a visual delayed match-to-sample task with pre- and retro-cueing to investigate the gating role of alpha and the impact of top-down attention on these dynamics. Consistent with prior findings, we observed that alpha power temporally tracked the presentation of the informative cue. Using ICA, we replicate previous results regarding the co-existence of two distinct alpha components (Alpha1 and Alpha2), and extend these findings by showing that these alpha components are differentially modulated during the decision delay with opposite relationship to behavioral outcomes. Furthermore, we report on preparatory mechanisms involving Alpha1 and Alpha2 and show that their subsequent impact on task performance varies with cue informativeness and the task relevance of upcoming input. When task relevance was unknown (i.e., in the retro-cue condition), increases in Alpha1 power before task-relevant input were associated with poorer performance. In contrast, when task relevance was known (i.e., in the pre-cue condition), pre-stimulus increases in Alpha2 power before task-irrelevant (i.e., to-be unattended) input was associated with better task performance. Combined, these findings are in line with our hypothesis that Alpha1 rhythms modulate visual processing while Alpha2 rhythms modulate the short-term storage of visual information. Crucially, these distinct (anticipatory) alpha dynamics only became evident using a decomposition approach, emphasizing the necessity of analysis methods that can disentangle overlapping rhythmic dynamics that stem from different sources.