Item

Abstract

Neurons in the lateral prefrontal cortex exhibit a huge diversity in their activity patterns mediating various cognitive functions ranging from working memory to monitoring of serial order and even visual awareness. In a previous study, we examined the neuronal activity in this region utilizing an ambiguous visual stimulation paradigm called the binocular flash suppression and found that a majority of the feature selective single unit responses were also perceptually modulated. The proportion of units displaying feature selective responses among all the neurons recorded, were but a minority. The present study aimed at characterizing any other task related patterns if any among the remaining single units. In order to do this, we decomposed the matrix of peristimulus time histograms of the remaining neurons utilizing the non-negative matrix factorization procedure, enabling us to characterize five dominant response patterns. Interestingly, the peak amplitudes of the different patterns were distributed across different phases of a trial. Further, a majority of neurons with firing profiles similar to a given response pattern did not display significant differences in their modulation during the monocular and binocular conditions of the task, thus indicating that sequential firing in the PFC is unaffected by sensory visual competition. Next, we aimed at assessing the functional connectivity as measured with noise correlations between pairs of neurons with modulation patterns similar to the five different response patterns obtained. Pairs of single units with firing profiles similar to identical response patterns displayed higher correlations, thus indicating a stronger functional coupling among units that were temporally coincident. However, when neurons were chosen from temporally separated populations, we observed a reduction in correlations. When positive and negative correlations were evaluated separately, such a correlation structure was observed specifically for positive correlations. Surprisingly, the negative correlations were uniformly distributed across the different populations. This possibly suggests a computational network principle mediating a representation of sequential patterns of activity in the lateral prefrontal cortex.