An Unsupervised Compressed Sensing Algorithm for Multi-Channel Neural Recording and Spike Sorting

Tao Xiong, Jie Zhang, Clarissa Martinez-Rubio, Chetan S. Thakur, Emad N. Eskandar, Sang Peter Chin, Ralph Etienne-Cummings, Trac D. Tran

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

We propose an unsupervised compressed sensing (CS)-based framework to compress, recover, and cluster neural action potentials. This framework can be easily integrated into high-density multi-electrode neural recording VLSI systems. Embedding spectral clustering and group structures in dictionary learning, we extend the proposed framework to unsupervised spike sorting without prior label information. Additionally, we incorporate group sparsity concepts in the dictionary learning to enable the framework for multi-channel neural recordings, as in tetrodes. To further improve spike sorting success rates in the CS framework, we embed template matching in sparse coding to jointly predict clusters of spikes. Our experimental results demonstrate that the proposed CS-based framework can achieve a high compression ratio (8:1 to 20:1), with a high quality reconstruction performance (>8 dB) and a high spike sorting accuracy (>90%).

Original languageEnglish (US)
JournalIEEE Transactions on Neural Systems and Rehabilitation Engineering
DOIs
StateAccepted/In press - Apr 26 2018
Externally publishedYes

Fingerprint

Compressed sensing
Sorting
Learning
Glossaries
Group Structure
Tetrodes
Action Potentials
Cluster Analysis
Electrodes
Template matching
Labels

Keywords

  • Compressed sensing
  • compressed sensing
  • Dictionaries
  • dictionary learning
  • Electrodes
  • Machine learning
  • multi-channel
  • neural recording
  • Neurons
  • Sensors
  • Sorting
  • spike sorting
  • unsupervised

ASJC Scopus subject areas

  • Neuroscience(all)
  • Biomedical Engineering
  • Computer Science Applications

Cite this

An Unsupervised Compressed Sensing Algorithm for Multi-Channel Neural Recording and Spike Sorting. / Xiong, Tao; Zhang, Jie; Martinez-Rubio, Clarissa; Thakur, Chetan S.; Eskandar, Emad N.; Chin, Sang Peter; Etienne-Cummings, Ralph; Tran, Trac D.

In: IEEE Transactions on Neural Systems and Rehabilitation Engineering, 26.04.2018.

Research output: Contribution to journalArticle

Xiong, Tao ; Zhang, Jie ; Martinez-Rubio, Clarissa ; Thakur, Chetan S. ; Eskandar, Emad N. ; Chin, Sang Peter ; Etienne-Cummings, Ralph ; Tran, Trac D. / An Unsupervised Compressed Sensing Algorithm for Multi-Channel Neural Recording and Spike Sorting. In: IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2018.
@article{649779ceb385461bbc13e3c6cbeedf76,
title = "An Unsupervised Compressed Sensing Algorithm for Multi-Channel Neural Recording and Spike Sorting",
abstract = "We propose an unsupervised compressed sensing (CS)-based framework to compress, recover, and cluster neural action potentials. This framework can be easily integrated into high-density multi-electrode neural recording VLSI systems. Embedding spectral clustering and group structures in dictionary learning, we extend the proposed framework to unsupervised spike sorting without prior label information. Additionally, we incorporate group sparsity concepts in the dictionary learning to enable the framework for multi-channel neural recordings, as in tetrodes. To further improve spike sorting success rates in the CS framework, we embed template matching in sparse coding to jointly predict clusters of spikes. Our experimental results demonstrate that the proposed CS-based framework can achieve a high compression ratio (8:1 to 20:1), with a high quality reconstruction performance (>8 dB) and a high spike sorting accuracy (>90{\%}).",
keywords = "Compressed sensing, compressed sensing, Dictionaries, dictionary learning, Electrodes, Machine learning, multi-channel, neural recording, Neurons, Sensors, Sorting, spike sorting, unsupervised",
author = "Tao Xiong and Jie Zhang and Clarissa Martinez-Rubio and Thakur, {Chetan S.} and Eskandar, {Emad N.} and Chin, {Sang Peter} and Ralph Etienne-Cummings and Tran, {Trac D.}",
year = "2018",
month = "4",
day = "26",
doi = "10.1109/TNSRE.2018.2830354",
language = "English (US)",
journal = "IEEE Transactions on Neural Systems and Rehabilitation Engineering",
issn = "1534-4320",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - An Unsupervised Compressed Sensing Algorithm for Multi-Channel Neural Recording and Spike Sorting

AU - Xiong, Tao

AU - Zhang, Jie

AU - Martinez-Rubio, Clarissa

AU - Thakur, Chetan S.

AU - Eskandar, Emad N.

AU - Chin, Sang Peter

AU - Etienne-Cummings, Ralph

AU - Tran, Trac D.

PY - 2018/4/26

Y1 - 2018/4/26

N2 - We propose an unsupervised compressed sensing (CS)-based framework to compress, recover, and cluster neural action potentials. This framework can be easily integrated into high-density multi-electrode neural recording VLSI systems. Embedding spectral clustering and group structures in dictionary learning, we extend the proposed framework to unsupervised spike sorting without prior label information. Additionally, we incorporate group sparsity concepts in the dictionary learning to enable the framework for multi-channel neural recordings, as in tetrodes. To further improve spike sorting success rates in the CS framework, we embed template matching in sparse coding to jointly predict clusters of spikes. Our experimental results demonstrate that the proposed CS-based framework can achieve a high compression ratio (8:1 to 20:1), with a high quality reconstruction performance (>8 dB) and a high spike sorting accuracy (>90%).

AB - We propose an unsupervised compressed sensing (CS)-based framework to compress, recover, and cluster neural action potentials. This framework can be easily integrated into high-density multi-electrode neural recording VLSI systems. Embedding spectral clustering and group structures in dictionary learning, we extend the proposed framework to unsupervised spike sorting without prior label information. Additionally, we incorporate group sparsity concepts in the dictionary learning to enable the framework for multi-channel neural recordings, as in tetrodes. To further improve spike sorting success rates in the CS framework, we embed template matching in sparse coding to jointly predict clusters of spikes. Our experimental results demonstrate that the proposed CS-based framework can achieve a high compression ratio (8:1 to 20:1), with a high quality reconstruction performance (>8 dB) and a high spike sorting accuracy (>90%).

KW - Compressed sensing

KW - compressed sensing

KW - Dictionaries

KW - dictionary learning

KW - Electrodes

KW - Machine learning

KW - multi-channel

KW - neural recording

KW - Neurons

KW - Sensors

KW - Sorting

KW - spike sorting

KW - unsupervised

UR - http://www.scopus.com/inward/record.url?scp=85046368780&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046368780&partnerID=8YFLogxK

U2 - 10.1109/TNSRE.2018.2830354

DO - 10.1109/TNSRE.2018.2830354

M3 - Article

C2 - 29877836

AN - SCOPUS:85046368780

JO - IEEE Transactions on Neural Systems and Rehabilitation Engineering

JF - IEEE Transactions on Neural Systems and Rehabilitation Engineering

SN - 1534-4320

ER -