MEDiCINe: Motion Correction for Neural Electrophysiology Recordings

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Electrophysiology recordings using laminar multielectrode arrays capture simultaneous neural activity.
• Electrode movement relative to neural tissue is a common issue in electrophysiology, affecting data quality.
• Accurate motion correction is essential for reliable single-neuron tracking and spike sorting.

Purpose of the Study:

• Introduce MEDiCINe (Motion Estimation by Distributional Contrastive Inference for Neurophysiology), a novel method for estimating motion in electrophysiology recordings.
• Evaluate MEDiCINe's performance against existing motion estimation techniques.
• Provide an open-source software implementation for broader research community adoption.

Main Methods:

• Developed MEDiCINe, a novel motion estimation algorithm leveraging distributional contrastive inference.
• Validated MEDiCINe using extensive simulated neurophysiology recordings.
• Tested MEDiCINe on real-world electrophysiology data from primate and rodent brains.

Main Results:

• MEDiCINe demonstrated superior performance compared to existing methods in simulated recordings.
• The method leads to more accurate spike sorting results.
• MEDiCINe accurately estimated motion across diverse electrophysiology datasets with varying motion characteristics.

Conclusions:

• MEDiCINe offers a robust and accurate solution for motion estimation in electrophysiology.
• The open-source release of MEDiCINe facilitates improved spike sorting and data analysis for researchers.
• This tool enhances the utility of electrophysiology datasets by enabling more precise neural activity analysis.