Deep Learning for Patch Clamp Electrophysiology
With the recent advances in deep learning enabling rapid and accurate identification of complex structures, it’s application to in vitro patch-clamp electrophysiology was inevitable. We are exploring the application of these techniques to enhance the capabilities of our fully-automatic patch clamp robots. For example, we have demonstrated fully automatic, real-time detection of healthy neurons within traditional DIC images and vision-based techniques to correct for hardware error stack-up during pipette localization.
Gonzalez, M. M., Lewallen, C. F., Yip, M. C., & Forest, C. R. (2021). Machine Learning-Based Pipette Positional Correction for Automatic Patch Clamp In Vitro. ENeuro, 8 (August), 1–8. doi: https://doi.org/10.1523/ENEURO.0051-21.2021[PDF]
M.C. Yip, M.M. Gonzalez, C.R. Valenta, M.J.M. Rowan, C.R. Forest. Deep learning-based real-time detection of neurons in brain slices for in vitro physiology. Sci Rep 11, 6065 (2021). https://doi.org/10.1038/s41598-021-85695-4[PDF]
M.M. Gonzalez, M.C. Yip, C.F. Lewallen, M.J. Rowan, C.R. Forest, Machine learning-based pipette correction for automated patch clamp in vitro, SfN Global Connectome, Virtual Conference, Jan 11-13, 2021.