Abstract
Selective attention enables the brain to efficiently cope with overwhelming amounts of visual information by selecting only that which is relevant to the categorization task at hand while ignoring irrelevant inputs. Here, we study the mechanisms of categorization-dependent attentional selection in an experiment that used 64 images of a realistic city street comprising varying embedded targets: a central face (task 1: “male” vs. “female”; task 2: “happy” vs. “neutral”), left flanked by a pedestrian (task 3: “male” vs. “female”), right flanked by a parked vehicle (task 4: “car” vs. “SUV”). Bubbles randomly sampled each image to generate 768 stimuli. In a within-participant design (N = 10), each performed the four 2-AFC categorization tasks (as listed above) on the same stimulus set. We concurrently recorded their categorization responses and source-localized MEG activity. First, we reconstructed the features each participant used in each task–computed as Mutual Information(Pixel visibility; Correct vs. Incorrect). Then, we traced the dynamic representation of each feature into MEG source responses–computing Mutual Information(Feature visibility; MEG sources)–and examined how different categorization tasks modulates the MEG source representations of the same stimulus features –computing Synergy(Feature visibility; MEG sources; Task). Based on these synergistic interactions, we reconstructed an attentional network that selects task-relevant features and reduces the same features when they are task-irrelevant. Specifically, left dorsal prefrontal and left ventrolateral prefrontal cortex (~80-90ms) interact with ventral and dorsal pathway to change the representation format of the same features depending on task–i.e. opponent format into the amplitude responses of the same source ~ 96-120ms. When task-relevant, features are each selected from occipital to higher cortical regions for categorization; When task-irrelevant, each is quickly reduced into occipital cortex [<170ms].