Alpha-band fluctuations represent behaviorally relevant excitability changes as a consequence of top-down guided spatial attention in a probabilistic spatial cueing design.

Spatial attention is a key function enabling the selection of relevantinformation and meaningful behavioral responses and is likely implemented bydifferent neural mechanisms. In previous work, attention led to robust butuncorrelated modulations of Steady-State-Visual-Evoked-Potentials (SSVEPs) as amarker of early sensory gain and visual as well as motor alpha-band activity. Weprobed the behavioral relevance of attention-modulated trial-by-trialfluctuations of these measures. For this purpose, in an experiment with aclassical probabilistic visuospatial attention cueing task, ato-be-discriminated target stimulus was validly, neutrally, or invalidly cued,while behavioral responses and EEG were recorded. Single-trial flicker-drivenSSVEPs, visual and motor alpha-band activity were measured and the relationshipbetween their amplitudes and reaction times was modeled via Bayesian regressionmodels, respectively. We replicated previous findings that these neural measuresand behavioral responses were overall modulated by the attentional cue. Beyondthat, SSVEP amplitudes were not associated with behavior, while single-trialalpha-band amplitudes were predictive of reaction times: For trials with a validor neutral cue, lower visual and motor alpha-band amplitudes measuredcontralateral to the target in the cue-target interval were associatedwith faster responses (and for valid cues also higher amplitudes ipsilateral tothe target). For invalid cues, which required attentional reallocating to theuncued side, no such relationship was found. We argue that behavioral relevanceof alpha-band modulations is a consequence but not a mechanism oftop-down guided spatial attention, representing neural excitability incortical areas activated by the attentional shift.