Non-reciprocal callosal projections and input gradients underlie interhemispheric communication in binocular visual cortex

How the brain combines information received independently by the two hemispheres is not fully understood. Here we describe a non-reciprocal circuit for interhemispheric communication in mouse binocular visual cortex via anatomically segregated neuronal populations. Callosal projecting neurons (CPNs) received only weak or no callosal input, whereas callosal receiving neurons (CRNs) made only weak or no callosal projections. Both populations received direct input from the thalamus. At the cellular level, CRNs had reduced excitability compared to non-CRNs (putative CPNs) due to elevated Kv1 potassium channel expression (encoded by the KCNA2 gene), with excitability of CRNs correlated with the magnitude of callosal input. Functionally, CRNs were predominately binocular, with binocularity correlated with callosal input, whereas non-CRNs (putative CPNs) were predominately monocular. In summary, we find that non-reciprocal callosal projections between CPNs and CRNs together with differences in excitability shaped by callosal input underlie interhemispheric communication in binocular visual cortex.