Joseph Emerson¹, Ryan Holland^2,3, Gordon Smith^1,2,3, Audrey Sederberg^4,5, Cheryl Olman^1,6; ¹Graduate Program in Neuroscience, University of Minnesota, Twin Cities, ²Department of Neuroscience, University of Minnesota, Twin Cities, ³Optical Imaging and Brain Sciences Medical Discovery Team, University of Minnesota, Twin Cities, ⁴School of Physics, Georgia Institute of Technology, Atlanta, GA, ⁵School of Psychology, Georgia Institute of Technology, Atlanta, GA, ⁶Department of Psychology, University of Minnesota, Twin Cities

The tuning properties of primary visual cortex (V1) display a large degree of heterogeneity across neurons which help support efficient coding of stimulus features. However, the mechanisms supporting heterogeneity in V1 are not well understood. We ask how orientation tuning heterogeneity is impacted by recurrent processing, which is known to shape orientation preference of cells in V1. To this end, we examine 2-photon calcium-imaging data from two anesthetized ferrets that viewed drifting grating stimuli at varying contrasts. We find that in both animals there is substantial variability in orientation selectivity across neurons. Interestingly, the variance of orientation selectivity across the population decreased with increasing contrast, suggesting a relationship between the strength of feedforward inputs and the population heterogeneity of orientation tuning. This trend is not due to contrast-dependent changes in measurement signal-to-noise ratios as a control analysis shows no significant trend in population heterogeneity across contrast when orientation labels are scrambled. We hypothesize that the change in tuning heterogeneity is brought on by a network transition characterized by a shift from feedforward-dominated inputs to a recurrence-dominated regime. To better understand the mechanistic underpinnings of this transition, we are using a stabilized supralinear network (SSN) model to investigate the contributions of recurrent connections in V1 to the population heterogeneity in the tuning properties of neurons. As part of ongoing work, we are testing how spatial and functional tuning properties of recurrent connectivity within the network affect orientation tuning variability across the population in a manner consistent with trends in ferret calcium-imaging data.

Acknowledgements: NIH R01 NS123482