Attentional Selection Determines Saccade Endpoint

The premotor theory of attention postulates that spatial attention arises from the activation of saccade areas and that the deployment of attention is the consequence of motor programming. Indeed, oculomotor and attentional processes share neural circuitries (e.g. frontal eye fields and superior colliculi). Nevertheless, within the same areas, these processes can be dissociated in a covert attention task, eliciting activation of visual neurons without concurrent activation of motor neurons. This dissociation contradicts the premotor theory of attention and suggests that motor preparation and visual attention can be separated at the neuronal level. Can we find evidence of such dissociation at the behavioral level? Here, we instructed participants to make a saccade towards one of two competing saccade cues and measured attention using oriented targets presented either at the saccade cues, in between them or at several other equidistant locations. When saccades ended at one of the saccade cues, we found an expected pre-saccadic shift of attention, with improved visual sensitivity at the location that became the endpoint of a saccade over the other. As the two saccade cues were presented in close proximity we could dissociate the final saccade endpoint from the intended saccade goal. Indeed, this spatial arrangement led to a substantial proportion of averaging saccades landing in between the two saccade cues. Interestingly, when saccades landed in between the two cues, attention was equally distributed across them and not deployed at the position in between. This shows that attention is not strictly coupled to the motor program of a saccade, ruling out the premotor theory of attention at the behavioral level. Moreover, our results suggest that saccade vectors are determined by the state of attentional selection before the eyes start to move and that averaging saccades arise from an unresolved attentional selection. Meeting abstract presented at VSS 2017