Abstract
- 1. Visual attention describes the cognitive mechanism that selectively filters behaviorally relevant from irrelevant information. Neurophysiological studies using single unit recordings have uncovered a network of fronto-parietal brain areas involved in the allocation of attention. However, important details regarding how ensembles of simultaneously active neurons support this behavior remain poorly documented. In this doctoral thesis, I investigated how neuronal ensembles within the primate lateral prefrontal cortex (LPFC) underlie attentional filtering under varying behavioral conditions. I first showed that ensemble activity correlates with behavior in a task during which the task difficulty was manipulated by changing the distance between ordinally- ranked stimuli. Behavioral performance improved with decreasing task difficulty and similarly, neuronal ensembles encoded the spatial location of the attended stimulus more efficiently for easier trials. Furthermore, interactions between similarly and dissimilarly tuned neurons also co-varied with task difficulty. However, these patterns of spike count correlations had no measurable effect on information coding in the neuronal population. Instead, I postulate a model that accurately describes all measured ensemble properties, in which these spike count correlations reflect anatomical and functional connectivity patterns of the underlying neuronal circuits. Second, I investigated how LPFC neurons encode different task-relevant stimulus parameters. I found evidence that neuronal ensembles encode both spatial and non- spatial features of attended stimuli more efficiently than single units. Those results indicate that LPFC may be relevant for both types of attention but also highlight that the
- 2. information contained in such maps cannot be predicted from the responses of individual units. Finally, I examined how the spatial arrangement of stimuli modulates attention. When two stimuli were presented within the same hemifield, behavioral performance as well as neuronal attentional filtering performance was impaired. I offer evidence that a change in the underlying correlational structure of the ensemble may affect information coding. These results give credence to the hypothesis that attentional capacities are hemifield-independent. Taken together, these results provide important new insight into the neural ensemble correlates of visual attention. Ensemble activity in the lateral prefrontal cortex of primates appears to be critically involved in attentional filtering under a variety of behavioral conditions.