Bypassing spike sorting: Density-based decoding using spike localization from dense multielectrode probes.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Machine Learning

Background:

• Neural decoding links brain activity to behavior, crucial for brain-computer interfaces (BCI).
• Spike sorting, assigning action potentials to neurons, is a key but often inaccurate step.
• Existing methods discard information by not modeling spike assignment uncertainty.

Purpose of the Study:

• To develop a spike sorting-free decoding method for improved BCI performance.
• To directly decode behavior from neural features while accounting for spike assignment uncertainty.
• To overcome limitations of current spike sorting inaccuracies.

Main Methods:

• Proposed a novel decoding method using a mixture of Gaussians (MoG) to model spike feature distributions.
• Integrated time-varying mixing proportions in the MoG to adapt to behavioral changes.
• Employed variational inference for model fitting and decoding, bypassing explicit spike clustering.

Main Results:

• The proposed spike sorting-free decoder consistently outperformed thresholding (multi-unit activity) and traditional spike sorting methods.
• Demonstrated robust performance across diverse animal recordings and probe configurations.
• The method effectively utilizes rich spike features from high-density probes.

Conclusions:

• A new spike sorting-free decoding framework offers superior performance in neural decoding.
• Directly modeling spike feature distributions and uncertainty enhances BCI decoding accuracy.
• This approach provides a more informative alternative to traditional spike sorting for BCI applications.