Neuro-cognitive models of single-trial EEG measures describe latent effects of spatial attention during perceptual decision making

Abstract

Visual perceptual decision-making involves multiple components including visual encoding, attention, accumulation of evidence, and motor execution. Recent research suggests that EEG signals can identify the time of encoding and the onset of evidence accumulation during perceptual decision-making. Although scientists show that spatial attention improves participant performance in decision making, little is known about how spatial attention influences the individual cognitive components that give rise to that improvement in performance. We found evidence in this work that both visual encoding time (VET) before evidence accumulation and other non-decision time processes after or during evidence accumulation are influenced by spatial top-down attention. Specifically, we used an open-source dataset in which participants were informed about the location of a target stimulus in the visual field on some trials during a face-car perceptual decision-making task. Fitting neural drift–diffusion models to response time, accuracy, and single-trial N200 latencies ($\sim$ 125 to 225 ms post-stimulus) of EEG allowed us to separate the processes of visual encoding and the decision process from other non-decision time processes such as motor execution. These models were fitted in a single step in a hierarchical Bayesian framework. Quantitative model comparison to simulation-based theories reveals that spatial attention manipulates both VET and other non-decision time processes. We discuss why spatial attention may affect other non-evidence accumulation processes, such as motor execution time (MET), and why this may seem unexpected given the literature. We provide recommendations for future work to deal with this topic by a combination of neuro-cognitive models and model simulations at the single-trial level.