Signal formation and validity

Use this category before interpreting fluorescence as neural activity. It answers whether samples share a usable clock, whether correction is identifiable, and whether optical or stimulation artifacts invalidate a biological claim.

Choose the workflow

Need	Workflow	Status	Important boundary
Inspect sampling, dropouts, flat steps, and channel health	Preprocessing and QC	Supported	QC can refuse analysis; it does not silently delete observations
Make irregular recordings safe for downstream methods	Irregular sampling	Supported	gaps remain gaps; resampling is explicit
Mask declared stimulation and assess sensor/reference validity	Optogenetic artifacts and sensor validity	Experimental	masking is not artifact correction
Account for indicator response dynamics	Sensor-kinetic modeling	Experimental	deconvolution is conditional on a calibrated model
Separate known fluorophores measured in overlapping channels	Optical unmixing	Experimental	the mixing matrix must be independently identified

Coverage gaps this category exposes

• validated automatic motion-artifact correction beyond declared reference fits;
• nonlinear and time-varying spectral mixing;
• preparation-specific, independently measured sensor and optical calibrations;
• prospective ground-truth benchmarks for control-free baselines; and
• validated correction of stimulation transients rather than conservative masks.

These are product gaps, not invitations to hide uncertainty. Until they are validated, outputs should expose the unresolved condition and narrow the claim.

Continue to

After the signal contract is fixed, choose an event-locked, continuous, or multi-signal analysis. Finish with population inference and robustness.