Real-time, motion compensation in Scanbox

Our colleague Tobias Rose was recently asking about Scanbox’s ability to stabilize the motion of images and signals from regions-of-intrest (ROIs) in real-time.  The goal of such processing is to be able to do experiments in closed-loop, and do very quick analyses on the neuronal responses, such as computing tuning curves on the fly.

Below is a simple demonstration of automatic stabilization in Scanbox, where the microscope is tracking the activity on a few ROIs the size of a single cell body. In the first pass of the ROI signals displayed at the bottom motion compensation is off.  The traces are noisy as the cell bodies themselves move with respect to the defined ROI.  The second pass shows the same image sequence once image stabilization is engaged (when the checkbox on the lower left is clicked), the images appear much more stable in space and, correspondingly, the signals for the ROIs are smoother and the resulting SNR is higher.

Incidentally, the demo also shows another feature of Scanbox — the ability to play back already collected data in a loop, so people can  learn the features of the microscope and try things out without the need of a living, animal subject.