This is a review of recent work in the automation of patch clamp. The reason I discuss it in a journal of theoretical neuroscience is that I have the (slightly naïve) hope that automating patch clamp experiments could be a way for theoretical neuroscientists (specifically those working on the neurophysiology of neurons) to do experiments themselves while still developing theory. Indeed, patch clamp is a notoriously tedious technique with a steep learning curve. In the last few years, there has been considerable progress in automating it.

The review covers two broad types of automation, with many interesting references. The first kind applies only to dense cell suspension cultures, and therefore has relatively limited application for neurophysiology. There are for example planar arrays, where cells are placed in tiny orifices where the recording is done. The quality of recordings is however not very good (but sufficient for testing some drugs, for example). The second kind uses robotics. It has been first developed in vivo, for blind patch clamp (simpler because it uses a single axis and no camera), and later in slices for targeted recordings (which requires a camera and three axes). The latest developments (last year) use two-photon microscopy for targeted patch clamp in vivo, using both cell and pipette fluorescence for visually guiding the pipette.

The review also identifies some of the challenges for future progress, in particular: implementing automation with the existing hardware of a lab is difficult, and some key parts of the experiment are still not automated, for example surgery, slicing, pipette pulling, internal solution making.