Many cortical network models use recurrent coupling strong enough to require inhibition for stabilization. Yet it has been experimentally unclear whether inhibition-stabilized network (ISN) models describe cortical function well across areas and states. Here, we test several ISN predictions, including the counterintuitive (paradoxical) suppression of inhibitory firing in response to optogenetic inhibitory stimulation. We find clear evidence for ISN operation in mouse visual, somatosensory, and motor cortex. Simple two-population ISN models describe the data well and let us quantify coupling strength. Although some models predict a non-ISN to ISN transition with increasingly strong sensory stimuli, we find ISN effects without sensory stimulation and even during light anesthesia. Additionally, average paradoxical effects result only with transgenic, not viral, opsin expression in parvalbumin (PV)-positive neurons; theory and expression data show this is consistent with ISN operation. Taken together, these results show strong coupling and inhibition stabilization are common features of the cortex.

Many cortical network models use recurrent coupling strong enough to require inhibition for stabilization. Yet it has been experimentally unclear whether inhibition-stabilized network (ISN) models describe cortical function well across areas and states. Here we test several ISN predictions, including the counterintuitive (paradoxical) suppression of inhibitory firing in response to optogenetic inhibitory stimulation. We find clear evidence for ISN operation in mouse visual, somatosensory, and motor cortex. Simple two-population ISN models describe the data well and let us quantify coupling strength. Though some models predict a non-ISN to ISN transition with increasingly strong sensory stimuli, we find ISN effects without sensory stimulation and even during light anesthesia. Additionally, average paradoxical effects result only with transgenic, not viral, opsin expression in parvalbumin (PV)-positive neurons; theory and expression data show this is consistent with ISN operation. Taken together, these results show strong coupling and inhibition stabilization are common features of cortex.[149 words] I. INTRODUCTION 16Extensive recurrent connectivity between nearby neurons is an ubiquitous feature of the cerebral cortex [1; 17 2; 3; 4]. Theoretical work has shown that the strength of recurrent coupling has a major impact on several 18 computational properties of networks of excitatory (E) and inhibitory (I) neurons, including the speed of 19 the response to external stimuli [5; 6], the ability of a network to sustain persistent activity [7], the capacity 20 and robustness of memory storage [8], and the amplification of various input modes [9]. 21 Strong excitatory recurrent coupling, however, can lead to unstable dynamics unless stabilized by inhi-22 bition. When recurrent connectivity is weak, excitatory cells can show stable firing rates independent of the 23 activity of inhibitory cells. In networks with strong recurrent connections, excitatory-to-excitatory (E-E) 24 connections amplify responses so that the excitatory network is unstable if the firing rates of inhibitory 25 neurons are kept fixed. Stable excitatory network operation across a range of firing rates can be restored 26 2 if inhibitory recurrent connections are sufficiently strong, allowing inhibition to track and balance excita-27 tion [5; 6; 7; 10; 11]. Such network models, with strong recurrent connections rendering the excitatory 28 cells self-amplifying and thus unstable, and requiring inhibition for stability, are called inhibitory-stabilized 29 networks (ISNs) [12]. 30 Whether cortical networks function in the ISN regime, in which conditions they do so, and which cortical 31 areas may operate as ISNs has been the subject of debate. Cat primary visual cortex (V1) shows behavior 32 consistent with the ISN regime [12]. But since that work used sensory stimuli, it could not determine whether 33 cat V1 operates as an ISN in the absence of visual stimuli (i.e. at rest). Based on these data and others, 34 Miller and co-authors later d...

Primary visual cortex (V1) in the mouse projects to numerous brain areas, including several secondary visual areas, frontal cortex, and basal ganglia. While it has been demonstrated that optogenetic silencing of V1 strongly impairs visually-guided behavior, it is not known which downstream areas are required for visual behaviors. Here we trained mice to perform a contrast-increment change detection task, for which substantial stimulus information is present in V1. Optogenetic silencing of visual responses in secondary visual areas revealed that their activity is required for even this simple visual task. In vivo electrophysiology showed that, although inhibiting secondary visual areas could produce some feedback effects in V1, the principal effect was profound suppression at the location of the optogenetic light. The results show that pathways through secondary visual areas are necessary for even simple visual behaviors.

AbstractPrimary visual cortex (V1) in the mouse projects to numerous brain areas, including several secondary visual areas, frontal cortex, and basal ganglia. While it has been demonstrated that optogenetic silencing of V1 strongly impairs visually-guided behavior, it is not known which downstream areas are required for visual behaviors. Here we trained mice to perform a contrast-increment change detection task, for which substantial stimulus information is present in V1. Optogenetic silencing of visual responses in secondary visual areas revealed that their activity is required for even this simple visual task. In vivo electrophysiology showed that, although inhibiting secondary visual areas could produce some feedback effects in V1, the principal effect was profound suppression at the location of the optogenetic light. The results show that pathways through secondary visual areas are necessary for even simple visual behaviors.