Linking behavioral and neural measures of visual search

Abstract

Abstract

Visual search is a ubiquitous behavior, yet its neural mechanisms remain incompletely understood. Here, we combine behavioral modeling and fMRI to identify neural signatures of core visual search computations. Participants performed visual search tasks in which targets differed from distractors along color, shape, or both dimensions, while attentional focus was experimentally manipulated. We show that behavioral search slopes – quantifying how response time scales with set size – are mirrored by corresponding fMRI-derived slopes, such that neural responses increase systematically with set size and scale with target-distractor similarity. This correspondence identifies a neural index of search difficulty and implicates regions in mid- and higher-level visual cortex in target-template matching processes. In addition, we replicate behavioral evidence for flexible attentional weighting of feature dimensions and demonstrate that this prioritization is reflected in modulation of neural activity in a subset of higher-level visual regions. Finally, we show that fMRI responses in bidimensional search can be predicted from activity measured during unidimensional search, demonstrating that computational models of behavior generalize to neural data. Together, these findings establish a quantitative link between behavioral and neural measures of visual search and provide a framework for studying attentional guidance and feature integration in the human brain.