Independent Markers of Attentional Shifts

Visuospatial attention has been extensively studied using a wide variety of markers in the brain and in behavior. We can broadly classify these markers into two distinct groups: reactive measures, where the presence and strength of attentional biases are inferred from the reaction to a probe stimulus, and proactive measures, which track attentional biases without (or before) the presentation of a probe stimulus. A difference in either type of measure (between two experimental conditions) is typically interpreted to reflect a difference in attentional resources. However, it is not a given that these various reactive and proactive measures of attention all read out the same construct; they may very well tap into independent processes. Here, we measure some of these proactive and reactive signatures of attending a location (changes in cortical excitability as measured by rapid invisible frequency tagging, alpha lateralization, gaze bias, probe-evoked ERPs, and behavioral reports) in an EEG–eye-tracking task (n = 23). Though all these metrics indeed correlate with our attentional manipulation, we show that they are not direct correlates of each other and that, of all the proactive measures, only the gaze bias predicts subsequent behavioral responses. This demonstrates that existing markers of attention cannot be used interchangeably and likely capture independent subprocesses of attention.