A multi-channel optical imaging and stimulation system includes a light source to deliver light beams into a light guide. Different light beams are coupled into different spatial modes supported by the light guide. The light guide includes multiple segments, each of which defines a window to couple a specified group of spatial modes out of the light guide to illuminate or stimulate a target. Light reflected, scattered, or emitted by the target is also collected by the windows in the light guide. The light collected by different windows is detected by different pixels of a detector, thereby creating a correspondence between the pixel location and the spatial location of site at which the light is collected. An image of the target is then reconstructed based on this correspondence.