Signal separability in integrated neurophotonics

Photonic probes record fluorescent signals by using arrays of light emitters and detectors embedded in neural tissue. Neither the emitted nor collected light fields are focused. Instead, in proposed configurations, hundreds of emitters will form rapid sequences of structured illumination patterns--providing sufficient spatial and temporal differentiation of neural signals for computational demixing. Here we define criteria for evaluating probe designs for achieving better signal separability. We find that probe geometry has profound, often unintuitive, effects on the separability of neural signals, providing initial design guidelines to achieve separation of individual cells in densely labeled populations.