The balance of excitation and inhibition in sensory cortex – BASCO

Visual cortex (V1) is the site at which dramatic transformations in neuronal receptive field properties - and thus the representation of the visual world - occur. One of the major transformations is the emergence of orientation selectivity. The functional organization of orientation selectivity in V1, however, takes different forms across species. In primates and carnivores it is topographically organized across cortex but in rodents no apparent organization is observed, yet rodents still exhibit orientation selectivity. Models that describe the emergence of orientation selectivity have relied on the functional organization found in primates to guide connectivity between neurons that share selectivity. Two different hypotheses have been proposed to explain the emergence of orientation selectivity without functional organization in rodent V1. In one hypothesis, a specific synaptic connectivity between neurons with shared orientation preference may nonetheless exist without topographic organization of cortex. Alternatively, a computational study has now demonstrated that orientation selectivity may arise from non-specific network connectivity, with the constraint that the excitatory and inhibitory inputs are balanced ("balanced network model"). These two hypotheses are not mutually exclusive, and evidence for both hypotheses currently exists, but the degree to which each of these hypotheses reflects the actual connectivity underlying orientation selectivity in rodent V1 is unclear. The goal of our proposal is to address the relative contributions of the balanced network and specific cortical connectivity to the generation of V1 orientation selectivity using experimental and computational studies. The proposed research is divided into three specific Aims that will be carried out collaboratively and will integrate theory and experiment.
Aim 1 What is the nature of the LGN input into layer 4 of V1 and how does layer 4 transform this input? In species with an orientation map the LGN neurons afferent inputs are precisely arranged. Is this also true for species without orientation map, and how does subcortical selectivity impact cortical selectivity? Can we explain the mechanism for orientation selectivity using a balanced network?
Aim 2 Is the cortical connectivity specific? If V1 operates in the balanced state, strong orientation selectivity will arise in layer 2/3, whether or not the connectivity is feature dependent. We will measure the orientation dependence of input correlations and integrate any specific connectivity into a balanced model.
Aim 3 How does disturbing the balanced state affect the cortical response? Our hypothesis is that the V1 operates in balanced excitation and inhibition regime. Perturbing this balance will be investigated theoretically and experimentally. Despite decades of study as the prime example of sensory processing, how V1 transforms incoming visual information is not well understood. It is not clear for example, whether feature specific connectivity is required to perform its function. In species with an orientation map, feature specific connectivity is not easily distinguished from connectivity that is solely dependent on anatomical distance because the anatomical and functional map are linked. The lack of an anatomical organization for orientation selectivity in rodent V1 therefore presents us with an opportunity to study circuitry in a system in which the functional selectivities of neurons are independent of their location within the cortical network.
Our proposal represents an integrative collaboration between theoreticians and experimentalists that will create an environment for students and postdoctoral fellows from different background to work side-by-side, gaining access to distinct expertise and perspectives. The collaboration represents a major effort for scientists to work in partnership between France and the US.