Short-line-illumination high-speed two-photon fluorescence microscopy

Two-photon fluorescence microscopy is a powerful technique for imaging neuronal activity, but its speed is traditionally constrained by the point-by-point scanning mechanism. To overcome this limitation, we present a novel approach that combines short-line illumination to boost imaging speed with computational techniques to recover spatial resolution. Our system excites the sample using a short-line point spread function and employs a single-pixel detector to capture the signal from the entire line simultaneously. A deep neural network then reconstructs the high-resolution images from the line-sampled data. This method substantially accelerates imaging by reducing the number of scanning rows required, while maintaining cellular resolution. We validate the fidelity of the reconstructed videos using simulated datasets. This approach offers a promising pathway toward high-throughput, high-speed neural activity imaging.