The output of interneurons in the primary visual cortex is best reflected by pre-synaptic activity, not somatic activity

Recent advances in 2-photon calcium-imaging in awake mice have made it possible to study the effect of different behavioural states on cortical circuitry. Many studies assume that somatic activity can be used as a measure for neuronal output. We set out to test the validity of this assumption by comparing somatic activity with the pre-synaptic activity of VIP (Vasoactive intestinal peptide)- and SST (Somatostatin)-positive interneurons in layer 2/3 of the primary visual cortex (V1). We used mice expressing genetically encoded calcium indicators in VIP/SST-interneurons across the whole cell (VIP/SST:GCaMP6f) or confined to pre-synapses (VIP/SST:SyGCaMP5). Mice were exposed to a full-field visual stimulation protocol consisting of 60-second-long presentations of moving Gabor gratings (0.04 cpd, 2 Hz) alternated by 30 seconds of grey screen. During imaging, mice were placed on an air-suspended Styrofoam ball, allowing them to run voluntarily. We compared neural activity during three 4-second time-windows: Before visual stimulation (-4 to 0 sec), during the initial onset (1 to 5 sec) and at the end of the stimulation (56 to 60 sec.). These were further compared while the mice were stationary and while they were voluntarily locomoting. Unlike VIP-somas, VIP-pre-synapses showed strong suppressive responses to the visual stimulus. Furthermore, VIP-somas were positively correlated with locomotion, whereas in VIP-synapses we observed a split between positive and negative correlations. In addition, a similar but weaker distinction was found between SST-somas and pre-synapses. The excitatory effect of locomotion in VIP-somas increased over the course of the visual stimulus but this property was only shared with the positively correlated VIP-pre-synapses. The remaining negatively correlated pre-synapses showed no relation to the overall activity of the Soma. Our results suggest that when making statements about the involvement of interneurons in V1 layer 2/3 circuitry it is crucial to measure from synaptic terminals as well as from somas.