Abstract
We previously demonstrated that attention to a motion direction inhibits nearby directions using behavioural measures (Yoo et al., VSS 2017), suggesting feature-based surround suppression (Tsotsos, 2011). In the present study, we extended this finding by recording responses of direction selective neurons in area MT and MST of macaques while they viewed different configurations of two random dot kinematograms (RDKs) presented within a RF. One RDK always moved in the neurons preferred direction (preferred pattern); the other could moved in one of 12 different directions (tuning pattern, 30° step). The animals were cued to attend to the preferred (attend-preferred condition) or tuning pattern (attend-tuning condition) and detected a direction change in the attended pattern while ignoring changes in the unattended pattern. In a third condition (fixation), the animals attended to the fixation point and detected colour change while ignoring both patterns. We measured neural responses as a function of the tuning pattern's direction. In all conditions, neurons maximally responded when the tuning pattern moved in the preferred direction. For the attend-tuning and fixation conditions, neuronal responses monotonically decreased as the direction of the tuning pattern became dissimilar from the preferred direction. However, for the attend-preferred condition, the minimal response was observed when the direction of the tuning pattern was 90° away from the preferred direction (p < .001), and then the response increased again as the directional difference became greater (p = .013). We fitted data from all conditions with three models: Single Gaussian (G), sum of two Gaussians (2G), and Mexican hat (MH). For the attend-preferred condition, the 2G and Mexican Hat functions fitted the data better than the G (p = .035). These results demonstrate that feature-based attention produces a modulation of direction tuning curves compatible with a suppressive surround around the attended stimulus feature.
Meeting abstract presented at VSS 2018