Abstract
Saliency map theory postulates that visual input is transformed into a topographic representation of visual conspicuity, which interacts with goal-related signals on a behavioral priority map for the control of attention/gaze. We have shown evidence that the superior colliculus (SC) is suited for the role of a saliency and priority mechanism, compartmentalized within its visual (SCs) and visuomotor (SCi) layers, respectively. Here, we examined how visual-salience and behavioral-relevance are coded in SC neuronal oscillations, while monkeys were presented with salient stimuli that were either goal-relevant or -irrelevant. Stimuli consisted of either a single item, or an array of items with a salient oddball (defined by color-orientation difference), which appeared in or opposite the receptive field (RF). In the first experiment, stimuli were goal-irrelevant because gaze was contingent upon a separate stimulus that always stepped orthogonal to the RF. We recorded local field potentials (LFP) using single electrodes and examined LFP power across conditions. Presenting the item or oddball in the RF caused a transient increase in LFP power across a broad (~15-90Hz), and separable narrow (5-12Hz), band that lasted ~200ms. This transient response was not modulated by saliency, and was similar in SCs and SCi. This was followed by prolonged (≥500ms) suppression at narrow beta (15-30 Hz) and theta/alpha (5-12 Hz) frequencies. The beta suppression was evident even when there was no stimulus in the RF, and was similar for the oddball and item. In contrast, the alpha/theta suppression was more pronounced for the oddball relative to the item. This frequency-specific signal was mostly restricted to the SCs. In a second experiment, where the stimulus was made goal-relevant, we observed a sustained increase in beta and low-gamma frequencies. These results may provide insight into how signals related to visual-salience and behavioral-relevance combine to establish a distributed priority map.
Meeting abstract presented at VSS 2016