Modality-specific predictive templates in pre-stimulus EEG activity

Abstract

Perceptual decision-making is a combination of sensory information and prior beliefs. In order to perform actions in a timely fashion, it is necessary to anticipate the timing at which events occur, but also what event is more likely than the other. While EEG signatures of anticipation have been identified it is less clear whether classifiers trained on informative (cued) trials generalize to uncued trials and whether such decoded templates predict trial-by-trial shifts in decision strategy (e.g., drift-rate or starting-point changes in a Diffusion Decision Model). This study aimed to determine whether human participants anticipated a visual or auditory stimulus at the single-trial level in both cued and uncued trials. We found that pre-stimulus brain activity contains information about the expected upcoming stimulus and that this information can be successfully extracted from single-trial brain activity. Behavioral analyses revealed a connection between correct anticipation and shifts in decision strategy, while also validating the classification of uncued trials. Importantly, the classification of uncued trials confirms that expectations build even in the absence of triggers. These findings highlight the presence of single-trial, stimulus-specific neural signatures of anticipation, offering new insights into trial-to-trial variability in decision-making and advancing our understanding of cognitive processes. HighlightsO_LIDistributed low-frequency pre-stimulus EEG encodes modality-specific expectations. C_LIO_LIAnticipatory neural states are evoked even in the absence of explicit cues. C_LIO_LIDecoded anticipation predict response speed and accuracy. C_LI