Spike inference from calcium imaging data acquired with GCaMP8 indicators

For neuroscience to be reproducible, its key methods must be quantitative and interpretable. Calcium imaging is such a key method, but it records neuronal activity only indirectly and is therefore difficult to interpret. These difficulties arise from the kinetics, nonlinearity, and sensitivity of calcium indicators, but also from the methods used for signal analysis. Here, we evaluate how methods for spike inference can be optimized to interpret data with the recently developed calcium indicator GCaMP8. We find that the indicator linearity of GCaMP8 enables more accurate recovery of high-frequency spiking events and the detection of single action potentials. Ground truth recordings from mouse neocortex show that the most linear calcium variants, GCaMP8s and GCaMP8m - but not GCaMP6, GCaMP7f, or GCaMP8f - robustly detect isolated spikes in pyramidal neurons under realistic noise conditions. In addition, we fine-tune and benchmark existing algorithms for spike inference (CASCADE, OASIS, MLSpike) with GCaMP8 data, investigate spike inference for interneurons, and demonstrate how the fast rise time of GCaMP8 enables low-latency real-time detection of neuronal activity. Together, our study provides tools and guidelines to optimally process calcium signals with GCaMP8 and highlights the key role of linearity in improving the interpretability of calcium imaging.