Dynamic neural reconstructions of attended object location and features using EEG

Abstract

We live in a dynamic world and our environment contains far more stimuli than our brain can process at one time. Visual attention allows us to select relevant information and ignore irrelevant information. What happens when attention is shifted from one item to another? To answer this question, it is critical to have tools that can accurately recover neural representations of both feature and location information, and do so across shifts of attention with high temporal resolution. In the current study, we used human electroencephalography (EEG) and machine learning methods to explore how neural representations of object features and locations update across dynamic shifts of attention. We demonstrate that EEG can be used to create simultaneous timecourses of the neural representations of attended features (timepoint-by-timepoint inverted encoding model reconstructions) and attended location (timepoint-by-timepoint decoding) during both stable periods and across dynamic spatial shifts of attention. On each trial two oriented gratings were presented that flickered at the same frequency but had different orientations; participants were cued to attend to one of them, and on half of the trials received a shift cue in the middle of the trial. We trained the models on a stable period from Hold attention trials, and then reconstructed/decoded the attended orientation/location information at each timepoint on Shift attention trials. Our results showed that both feature reconstructions and location decoding dynamically track the shift of attention, and that there may be timepoints during the shifting of attention when (1) feature and location representations become uncoupled, and (2) both the previously-attended and currently-attended orientations are represented with roughly equal strength. The results offer insight into our understanding of attentional shifts, and the noninvasive techniques developed in the current study lend themselves well to a wide variety of future applications.