Attentional shifts in parietal cortex and behavior

During natural behavior attention is rapidly shifted among elements of a scene in accordance with evolving behavioral demands. Here we present data collected from two monkeys in a novel attention-switch task that allowed us to measure with temporal precision the neurophysiologic and behavioral time course of shifting spatial attention in response to an endogenous cue. The task consisted of detecting a near-threshold transient speed change (the "motion pulse") at one of two peripheral motion stimuli. The monkey was cued as to which stimulus was more likely to contain the change (85% valid cues, 15% invalid cues), and the cued stimulus could switch mid-trial. Surprisingly, the animals' accuracy at the newly cued location was better than at the previously cued location for motion pulses that occurred simultaneously with--or even before--the cue switch. Since the animals could not predict the timing of the cue switch, this time course must be due to the attentional switch interacting with the neuronal response to the motion pulse during the delay between when the motion pulse occurs and when the animals signal their response. In fact, the timing of the behavioral attentional shift can be reasonably well explained by a simple, data-based model comparing the time course of the neurophysiologic attentional switch with the time course of the neural response to the motion pulse. This supports the hypothesis that attentional effects on behavior arise from an interaction between neural attentional modulation and incoming sensory responses. We also show that the magnitude of the response to the motion pulse in both middle temporal (MT) and lateral intraparietal (LIP) cortex correlated with the animals' detection performance. Such results have been interpreted previously as variability in sensory responses that lead to variable perception. However, we suggest that, at least for our data set, a more parsimonious explanation is that shifts of spatial attention in response to the motion pulse are largely responsible for the variability. Lastly, we show that attentional modulation appears substantially earlier in area LIP than in MT, an anatomically-connected lower visual area. This temporal sequence of attentional latencies demonstrates that endogenous changes of state can occur in higher visual areas before lower visual areas, and supports a role for LIP as a source of top-down attentional signals to early visual cortex.