Timezone: »

 
Spotlight Poster
Bypassing spike sorting: Density-based decoding using spike localization from dense multielectrode probes
Yizi Zhang · Tianxiao He · Julien Boussard · Charles Windolf · Olivier Winter · Eric Trautmann · Noam Roth · Hailey Barrell · Mark Churchland · Nicholas A Steinmetz · Erdem Varol · Cole Hurwitz · Liam Paninski

Thu Dec 14 03:00 PM -- 05:00 PM (PST) @ Great Hall & Hall B1+B2 #538
Event URL: https://github.com/yzhang511/density_decoding »

Neural decoding and its applications to brain computer interfaces (BCI) are essential for understanding the association between neural activity and behavior. A prerequisite for many decoding approaches is spike sorting, the assignment of action potentials (spikes) to individual neurons. Current spike sorting algorithms, however, can be inaccurate and do not properly model uncertainty of spike assignments, therefore discarding information that could potentially improve decoding performance. Recent advances in high-density probes (e.g., Neuropixels) and computational methods now allow for extracting a rich set of spike features from unsorted data; these features can in turn be used to directly decode behavioral correlates. To this end, we propose a spike sorting-free decoding method that directly models the distribution of extracted spike features using a mixture of Gaussians (MoG) encoding the uncertainty of spike assignments, without aiming to solve the spike clustering problem explicitly. We allow the mixing proportion of the MoG to change over time in response to the behavior and develop variational inference methods to fit the resulting model and to perform decoding. We benchmark our method with an extensive suite of recordings from different animals and probe geometries, demonstrating that our proposed decoder can consistently outperform current methods based on thresholding (i.e. multi-unit activity) and spike sorting. Open source code is available at https://github.com/yzhang511/density_decoding.

Author Information

Yizi Zhang (Columbia University)
Yizi Zhang

I’m currently in my third year of the Statistics PhD program at Columbia University, fortunate to have the opportunity to work with Prof. Liam Paninski. My research interests lie at the intersection of computational neuroscience and machine learning where I focus on improving neural decoding methods for brain computer interfaces. I am broadly interested in machine learning and statistics with recent interests in: - Probabilistic Machine Learning: Latent variable models, deep generative models, variational inference - Computational Neuroscience: Neural decoding, brain-computer interfaces, population neural dynamics - Data Valuation: Influence functions for black-box models, time series data valuation

Tianxiao He (New York University)
Julien Boussard (Columbia University)
Charles Windolf (Flatiron Institute)
Olivier Winter (Institut de Physique du Globe)
Eric Trautmann (Columbia University)
Noam Roth (University of Washington)
Hailey Barrell (University of Washington)
Mark Churchland (Columbia University)
Nicholas A Steinmetz (UCL)
Erdem Varol (New York University)
Cole Hurwitz (Zuckerman Institute, Columbia University)
Liam Paninski (Columbia University)

More from the Same Authors