Reverse-engineering sensory-evoked signal flow in rat barrel cortex

We present a novel reverse-engineering approach that allows investigating sensory-evoked signal flow through individual neurons within the context of their surrounding neural networks. To do so, spontaneous and sensory-evoked activity is recorded from individual neurons in vivo. In addition, the complete 3D dendrite and axon projection patterns of these neurons are reconstructed and registered into an anatomically realistic model of rat barrel cortex. This model allows estimating the number and cell type-specific subcellular distribution of synapses on these neurons. Next, the neurons are “wired” into the network by connecting the synapses to presynaptic neurons based on cell type-specific connection probabilities. The number of functional synapses on this neuron is determined by including measurements of cell type-specific ongoing and sensory-evoked spiking probabilities for all presynaptic cell types. Finally, this neuron is turned into a compartmental model and constrained by comparing model responses to ongoing and sensory-evoked synaptic inputs to in vivo measured responses. For the first time, this allows investigating in vivo measured sensory-evoked responses of single neurons in anatomically realistic computer models, complementing previous biophysically detailed models of single neurons based on in vitro experiments. To investigate the mechanistic principles underlying sensory responses in different cell types, we use a Monte Carlo method for sampling the parameter space of the network-embedded neuron model. For example, by varying the functional connectivity or timing of sensory input, we identify model configurations that match the in vivo observed responses of these neurons. As a first demonstration of the feasibility of this approach, we investigate sensory-evoked responses in two different pathways in rat barrel cortex