Spike inference for genetically encoded calcium indicators with models of multistep binding kinetics

Seminar | March 19 | 4-5:30 p.m. | 177 Life Sciences Addition

 Dr. David Greenberg, Center of Advanced European Studies and Research, Bonn, Germany

 Neuroscience Institute, Helen Wills

Multiphoton imaging of genetically encoded calcium indicators can detect action potential (AP) evoked fluorescence changes from populations of spatially resolved neurons, but the nonlinear dependence of fluorescence on AP counts and variable indicator expression across neurons make quantitative inference problematic. We developed a biophysical model of GCaMP6s in neurons based on the 4-step binding kinetics of calcium to GCaMP6s that accurately describes AP-evoked fluorescence changes in vivo. Because of the model’s physical interpretation, rate constants obtained from in vitro binding assays can be applied to predict AP-evoked fluorescence responses in vivo. We used this model to infer APs from fluorescence data, and measured accuracy using simultaneous cell-attached recording and multiphoton imaging. This approach for AP inference from fluorescence data consistently outperformed alternative methods and demonstrates the utility of biophysically grounded models for quantitative analysis of complex biological data.