Spike Estimation from Fluorescence Signals Using High-Resolution Property of Group Delay.

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Accurate spike estimation from calcium (Ca2+) fluorescence signals is essential for understanding neural activity.
• Existing data-driven and physiological models struggle with precise spike localization.
• Current methods often lack the accuracy needed for detailed neural circuit analysis.

Purpose of the Study:

• To develop a novel signal processing approach for accurate spike estimation from Ca2+ fluorescence signals.
• To leverage the inherent properties of fluorescence responses using minimum-phase group delay (GD) functions.
• To provide a robust and efficient algorithm for spike deconvolution.

Main Methods:

• Interpreted Ca2+ signals as resonator responses to neural impulses.
• Utilized minimum-phase group delay (GD) analysis for signal filtering.
• Developed the GDSpike algorithm based on GD functions for spike deconvolution.
• Evaluated performance on diverse datasets with various indicators and sampling rates.

Main Results:

• The GDSpike algorithm demonstrated superior performance compared to the Vogelstein deconvolution algorithm.
• GDSpike achieved performance comparable to the MLSpike method.
• The algorithm showed effectiveness across nine diverse datasets from different mouse brain regions.
• Post-processing MLSpike output with GDSpike further enhanced spike estimation accuracy.

Conclusions:

• GDSpike offers a robust signal processing-based alternative for accurate neural spike estimation.
• The method effectively resolves spike locations by exploiting the physical properties of Ca2+ indicators.
• GDSpike provides a valuable tool for neuroscientists, enhancing the analysis of neural dynamics.