Neural Determinants of Task Performance during Feature-Based Attention in Human Cortex

Abstract

Studies of feature-based attention have associated activity in a dorsal frontoparietal network with putative attentional priority signals. Yet, how this neural activity mediates attentional selection and whether it guides behavior are fundamental questions that require investigation. We reasoned that endogenous fluctuations in the quality of attentional priority should influence task performance. Human subjects detected a speed increment while viewing clockwise or counter-clockwise motion (baseline task) or while attending to either direction amid distracters (attention task). In an fMRI experiment, direction-specific neural pattern similarity between the baseline task and the attention task revealed a higher level of similarity for correct than incorrect trials in frontoparietal regions. Using TMS, we disrupted posterior parietal cortex and found a selective deficit in the attention task, but not in the baseline task, demonstrating the necessity of this cortical area during feature-based attention. These results reveal that frontoparietal areas maintain attentional priority that facilitates successful behavioral selection.

Significance Statement To cope with the computational limits of visual processing, the brain selectively prioritizes a subset of visual input. The selection for visual features, such as color and motion, has been associated with activity in a frontoparietal cortical network. Yet, the role this activity plays in mediating selection and influencing behavior is not clear. Using fMRI, we show that neural activity patterns in several frontoparietal areas correlated with task performance. Furthermore, neurodisruption of the posterior parietal cortex using transcranial magnetic stimulation selectively impaired feature selection. These results provide the first evidence that the neural representation of prioritized features in frontoparietal areas play a causal role in selecting visual features.