Ultrafast optical brain imaging via blurred matrix completion

Summary: Neural computation is a global phenomenon driven by a myriad of local, granular biophysical phenomena at millisecond time-scales. Advancing our knowledge of neural function and dysfunction requires new technologies that can capture granular information over large areas at fast framerates. We propose a new computational imaging approach that combines easily disseminable optics with algorithmic reconstruction algorithms that will advance the capabilities of both in vivo two-photon and light-sheet imaging, reaching the >100 Hz framerate range over large areas and volumes.