Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HuiduRep: A Robust Self-Supervised Framework for Learning Neural Representations from Extracellular Recordings

Feng Cao, Zishuo Feng, Jicong Zhang, Wei Shi

TL;DR

HuiduRep addresses spike-sorting challenges under low SNR and electrode drift by marrying contrastive self-supervision with a denoising autoencoder in a MoCo-v3–style framework. A transformer-based encoder processes physiologically augmented views, and a Gaussian Mixture Model clusters learned spike representations to yield a label-free sorting pipeline. Across diverse datasets, including synthetic drift and real ground-truth recordings, HuiduRep demonstrates robust performance and often outperforms state-of-the-art tools like KiloSort4 and MountainSort5, with ablations underscoring the DAE’s role in cross-domain robustness. The work suggests a promising direction for generalizable extracellular signal processing and could extend to other bioelectrical signals such as EMG, ECoG, and EEG.

Abstract

Extracellular recordings are transient voltage fluctuations in the vicinity of neurons, serving as a fundamental modality in neuroscience for decoding brain activity at single-neuron resolution. Spike sorting, the process of attributing each detected spike to its corresponding neuron, is a pivotal step in brain sensing pipelines. However, it remains challenging under low signal-to-noise ratio (SNR), electrode drift and cross-session variability. In this paper, we propose HuiduRep, a robust self-supervised representation learning framework that extracts discriminative and generalizable features from extracellular recordings. By integrating contrastive learning with a denoising autoencoder, HuiduRep learns latent representations that are robust to noise and drift. With HuiduRep, we develop a spike sorting pipeline that clusters spike representations without ground truth labels. Experiments on hybrid and real-world datasets demonstrate that HuiduRep achieves strong robustness. Furthermore, the pipeline outperforms state-of-the-art tools such as KiloSort4 and MountainSort5. These findings demonstrate the potential of self-supervised spike representation learning as a foundational tool for robust and generalizable processing of extracellular recordings. Code is available at: https://github.com/IgarashiAkatuki/HuiduRep

HuiduRep: A Robust Self-Supervised Framework for Learning Neural Representations from Extracellular Recordings

TL;DR

HuiduRep addresses spike-sorting challenges under low SNR and electrode drift by marrying contrastive self-supervision with a denoising autoencoder in a MoCo-v3–style framework. A transformer-based encoder processes physiologically augmented views, and a Gaussian Mixture Model clusters learned spike representations to yield a label-free sorting pipeline. Across diverse datasets, including synthetic drift and real ground-truth recordings, HuiduRep demonstrates robust performance and often outperforms state-of-the-art tools like KiloSort4 and MountainSort5, with ablations underscoring the DAE’s role in cross-domain robustness. The work suggests a promising direction for generalizable extracellular signal processing and could extend to other bioelectrical signals such as EMG, ECoG, and EEG.

Abstract

Extracellular recordings are transient voltage fluctuations in the vicinity of neurons, serving as a fundamental modality in neuroscience for decoding brain activity at single-neuron resolution. Spike sorting, the process of attributing each detected spike to its corresponding neuron, is a pivotal step in brain sensing pipelines. However, it remains challenging under low signal-to-noise ratio (SNR), electrode drift and cross-session variability. In this paper, we propose HuiduRep, a robust self-supervised representation learning framework that extracts discriminative and generalizable features from extracellular recordings. By integrating contrastive learning with a denoising autoencoder, HuiduRep learns latent representations that are robust to noise and drift. With HuiduRep, we develop a spike sorting pipeline that clusters spike representations without ground truth labels. Experiments on hybrid and real-world datasets demonstrate that HuiduRep achieves strong robustness. Furthermore, the pipeline outperforms state-of-the-art tools such as KiloSort4 and MountainSort5. These findings demonstrate the potential of self-supervised spike representation learning as a foundational tool for robust and generalizable processing of extracellular recordings. Code is available at: https://github.com/IgarashiAkatuki/HuiduRep

Paper Structure

This paper contains 18 sections, 3 equations, 3 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Template-based Spike Sorters
  4. Representation Learning Models
  5. Method
  6. Architecture of HuiduRep
  7. Spike Sorting Pipeline
  8. Datasets
  9. International Brain Laboratory (IBL) Dataset
  10. Hybrid Janelia Dataset
  11. Paired MEA64C Yger Dataset
  12. Experiments
  13. Implementation Details
  14. Performance Evaluation
  15. Ablation Study
  16. ...and 3 more sections

Figures (3)

  • Figure 1: Overall architecture of HuiduRep and the pipeline. During training, the contrastive learning branch adapts the MoCo v3 style framework, where the query is compared with key and other in-batch samples (not shown in the figure due to the limited space). Only the View 1 is passed to the DAE branch for reconstruction. During inference, only the transformer encoder and DAE module are used to extract representations.
  • Figure 2: (a): Boxplot of HuiduRep and other models. (b): Clustering results, visualized after reduction via PCA.
  • Figure 3: Reduced feature space of IBL training dataset and other datasets. The centroid of each dataset is marked with a black X. The features of each dataset are reduced to 2 dimensions using PCA.