Abstract
Largely on the basis of its interesting anatomy, the pulvinar has been hypothesized to play a useful role in visual attention. The pulvinar is the largest nucleus of the thalamus and is notable for its hub-like connectivity: it has extensive reciprocal connections with the cortex, including occipital, parietal, temporal, and frontal areas. This architecture might facilitate tasks in which one cortical area must modulate the activity of another. Furthermore, gradients of connectivity have been found within the non-human primate (NHP) pulvinar, where nearby cortical areas project to nearby areas of the pulvinar. Of most interest here is a gradient running roughly from anterior to posterior pulvinar. Highly spatiotopic areas, like striate and extrastriate cortex, connect more anteriorly; areas with less spatial orgranisation, like IT, connect more posteriorally.
So how might this detailed anatomical picture relate to attentional function in humans? Some previous human lesion studies have tended to group diverse pulvinar lesions together. Here we studied spatial and temporal feature binding in three patients, focusing on the location of the lesion and the connectivity that was likely to be disrupted. Based on NHP anatomy, we expected that anterior compared to posterior lesions would be more likely to disrupt connections with highly spatiotopic cortex. Consistent with this hypothesis, we found that lesion to the anterior pulvinar mainly disrupted the ability to bind visual features to a location in space, whereas lesion to the posterior pulvinar produced deficits in binding to a point in time. These results suggest correspondence with the neural anatomy of the pulvinar, and its gradient of connectivity to spatially organized cortex. We discuss our findings within the scope of neural network models of visual attention, in which the pulvinar facilitates communication between different brain areas.
This work was funded by grant C501417 from the Biotechnology and Biological Sciences Research Council (BBSRC).