Inherent Biases in Spontaneous Cortical Dynamics

Does visual perception begin with a 'blank slate'? Is the activity of visual cortex 'at rest' composed of a random pattern of noisy neurons? Or is there structure inherent in the spontaneous activity of the cortex? Previous studies have suggested that the pattern of spontaneous activity in cortex consists of random or stochastic transitions from one active population (network) to another. Here, as part of an effort to understand interactions between neurons that encode surface and border, we have recorded from pairs of full-field luminance-modulated (surface) and oriented (border) cells in cat areas seventeen and eighteen, and examined the relationships in their spike firing pattern using cross-correlation analysis. Surprisingly, under spontaneous conditions, we find shifted correlation peaks between these two cell classes, suggesting a directional, non-random interaction. Furthermore, the peak positions under spontaneous conditions do not predict those during visual stimulation. The directional nature of these spontaneous interactions indicates that 'at rest' there are inherent biases in cortical dynamics and suggests a temporally structured baseline on which visually driven cortical activity is superimposed.