Breakstone, M.; Chen, S.; Vadapalli, S.; Chavez, E.; Parsonnet, L. S.; Gross, R. E.; Tescarollo, F.; Barker, D. J.; Sun, H.

Simultaneous fiber photometry and optogenetics is a powerful emerging technique for precisely studying the interactions of neuronal brain networks. However, spectral overlap between photometry and optogenetic components has severely limited the application of an all-optical approach. Due to spectral overlap, light from optogenetic stimulation saturates the photosensor and occludes photometry fluorescence, which is especially problematic in physically smaller model organism brains like mice. Here, we demonstrate the Multi-Frequency Interpolation X-talk removal algorithm (MuFIX) for recovering crosstalk-contaminated photometry responses recorded with lock-in amplification. MuFIX exploits multi-frequency lock-in amplification by modeling the remaining uncontaminated data to interpolate across crosstalk-affected segments (R2 ~ 1.0); we found that this approach accurately recovers the original photometry response after demodulation (Pearson\'s r ~ 1.0). When applied to crosstalk-contaminated data, MuFIX recovered a photometry response closely resembling the dynamics of non-crosstalk photometry recorded simultaneously. Upon further verification using simulated and empirical data, we demonstrated that MuFIX reproduces any signal that underwent simulated crosstalk contamination (r ~ 1.0). We believe adopting MuFIX will enable experimental designs using simultaneous fiber photometry and optogenetics that were previously not feasible due to crosstalk.