Multi-day Neuron Tracking in High Density Electrophysiology Recordings using EMD.

Area of Science:

• Neuroscience
• Computational Neuroscience

Background:

• Accurate long-term neuronal tracking is essential for studying brain plasticity, learning, and adaptation.
• High-density electrophysiology, like Neuropixels, offers potential but faces challenges due to tissue drift and signal loss.
• Existing neuron tracking methods often rely on firing statistics, limiting their applicability.

Approach:

• Developed a novel neuron tracking method based on non-rigid alignment of spike-sorted clusters across multiple days.
• This approach identifies the same cells independent of their firing patterns, overcoming limitations of current techniques.
• Verified cell identity using visual receptive field mapping in mice.

Key Points:

• The proposed method successfully tracks neurons across recording sessions separated by up to 47 days.
• Achieved an average recovery rate of 84% for the same cells.
• Demonstrates robustness against tissue drift and signal variability inherent in long-term electrophysiology.

Conclusions:

• This novel alignment-based method significantly improves the accuracy and reliability of long-term neuronal tracking.
• Enables more robust investigation of neuronal dynamics during learning and adaptation.
• Provides a valuable tool for analyzing longitudinal neural data from high-density probes.