Rapid learning of neural circuitry from holographic ensemble stimulation enabled by model-based compressed sensing.

Discovering how computations are implemented in the brain at the level of monosynaptic connectivity requires probing for connections from potentially thousands of presynaptic candidate neurons. Two-photon optogenetics is a promising technology for mapping such connectivity via sequential stimulation of individual neurons while recording postsynaptic responses intracellularly. However, this technique is currently not scalable because stimulating neurons one by one requires prohibitively long experiments. Here we developed novel computational tools that, when combined, enable learning of monosynaptic connectivity from high-speed holographic ensemble stimulation. First, we developed a model-based compressed sensing algorithm that identifies connections from postsynaptic responses evoked by stimulating many neurons at once, greatly increasing mapping efficiency. Second, we developed a deep-learning method that isolates the postsynaptic response to each stimulus, allowing stimulation to rapidly switch between ensembles without waiting for the postsynaptic response to return to baseline. Together, our system increases the throughput of connectivity mapping by an order of magnitude, facilitating discovery of the circuitry underlying neural computations.