The Miniscope uses wide-field fluorescence imaging to record neural activity in awake, freely moving mice. It has a mass of 3 grams and uses a single, flexible coaxial cable for power, control signals, and imaging data. You can buy for just under two grand as well as other parts on labmaker.org.
The website can be a little tricky to use, and much of the information you want is actually accessed by clicking on the “Sub-Systems header”. However if you get things built, then much like OpenSPIM the software to operate these devices is open source and designed for Micro-Manager which is nice and easy.
Maya Anjur-Dietrich gave me an old model to play with. Included was a low mag lens, and a high mag lens with condenser. There was an LED light source, and a magnetic mounter for your cell phone to take pictures. There’s also a way to mount the Foldscope to your phone so that you can use the light from your phone as a projector which I didn’t try.
The lenses have a strong aberrations, and the Foldscope is not the easiest thing to build or to use. However I was working with an older model, and you can read a detailed review of the newer model here. I think that the actual tiny lenses themselves are the most valuable component. Combined with a few old phones or webcams, 20 of these lenses provide a powerful way to take time-lapses on a bunch of different samples all at once.
*Post-edit: Just a week after writing this I came upon an experiment where using a collar was the difference between pristine data and no data at all. So I’ve been avoiding correction collars less now.