Identifying multiscale hidden states to decode behavior

Hamidreza Abbaspourazad, Yan Wong, Bijan Pesaran, Maryam M. Shanechi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

A key element needed in a brain-machine interface (BMI) decoder is the encoding model, which relates the neural activity to intended movement. The vast majority of work have used a representational encoding model, which assumes movement parameters are directly encoded in neural activity. Recent work have in turn suggested the existence of neural dynamics that represent behavior. This recent evidence motivates developing dynamical encoding models with hidden states that encode movement. Regardless of their type, encoding models have vastly characterized a single scale of activity, e.g., either spikes or local field potentials (LFP). In our recent work we developed a multiscale representational encoding model to simultaneously characterize and decode discrete spikes and continuous field activity. However, learning a multiscale dynamical model from simultaneous spike-field recordings in the presence of hidden states is challenging. Here we present an unsupervised learning algorithm for estimating a multiscale state-space model with hidden states and validate it using spike-LFP activity during a reaching movement. We use the learned multiscale statespace model and a corresponding decoder to identify hidden states from spike-LFP activity. We then decode the movement trajectories using these hidden states. We find that the identified states can accurately decode the trajectories. Moreover, we demonstrate that adding LFP to spikes improves the decoding accuracy, suggesting that our unsupervised learning algorithm incorporates information across scales. This learning algorithm could serve as a new tool to study encoding across scales and to enhance future BMI systems.

Original languageEnglish (US)
Title of host publication40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3778-3781
Number of pages4
Volume2018-July
ISBN (Electronic)9781538636466
DOIs
StatePublished - Oct 26 2018
Event40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018 - Honolulu, United States
Duration: Jul 18 2018Jul 21 2018

Other

Other40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018
CountryUnited States
CityHonolulu
Period7/18/187/21/18

Fingerprint

Learning
Brain-Computer Interfaces
Learning algorithms
Space Simulation
Unsupervised learning
Action Potentials
Brain
Trajectories
Decoding

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Cite this

Abbaspourazad, H., Wong, Y., Pesaran, B., & Shanechi, M. M. (2018). Identifying multiscale hidden states to decode behavior. In 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018 (Vol. 2018-July, pp. 3778-3781). [8513242] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EMBC.2018.8513242

Identifying multiscale hidden states to decode behavior. / Abbaspourazad, Hamidreza; Wong, Yan; Pesaran, Bijan; Shanechi, Maryam M.

40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018. Vol. 2018-July Institute of Electrical and Electronics Engineers Inc., 2018. p. 3778-3781 8513242.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abbaspourazad, H, Wong, Y, Pesaran, B & Shanechi, MM 2018, Identifying multiscale hidden states to decode behavior. in 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018. vol. 2018-July, 8513242, Institute of Electrical and Electronics Engineers Inc., pp. 3778-3781, 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018, Honolulu, United States, 7/18/18. https://doi.org/10.1109/EMBC.2018.8513242
Abbaspourazad H, Wong Y, Pesaran B, Shanechi MM. Identifying multiscale hidden states to decode behavior. In 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018. Vol. 2018-July. Institute of Electrical and Electronics Engineers Inc. 2018. p. 3778-3781. 8513242 https://doi.org/10.1109/EMBC.2018.8513242
Abbaspourazad, Hamidreza ; Wong, Yan ; Pesaran, Bijan ; Shanechi, Maryam M. / Identifying multiscale hidden states to decode behavior. 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018. Vol. 2018-July Institute of Electrical and Electronics Engineers Inc., 2018. pp. 3778-3781
@inproceedings{0445d90c7b3c49bfafe51a24720ccdda,
title = "Identifying multiscale hidden states to decode behavior",
abstract = "A key element needed in a brain-machine interface (BMI) decoder is the encoding model, which relates the neural activity to intended movement. The vast majority of work have used a representational encoding model, which assumes movement parameters are directly encoded in neural activity. Recent work have in turn suggested the existence of neural dynamics that represent behavior. This recent evidence motivates developing dynamical encoding models with hidden states that encode movement. Regardless of their type, encoding models have vastly characterized a single scale of activity, e.g., either spikes or local field potentials (LFP). In our recent work we developed a multiscale representational encoding model to simultaneously characterize and decode discrete spikes and continuous field activity. However, learning a multiscale dynamical model from simultaneous spike-field recordings in the presence of hidden states is challenging. Here we present an unsupervised learning algorithm for estimating a multiscale state-space model with hidden states and validate it using spike-LFP activity during a reaching movement. We use the learned multiscale statespace model and a corresponding decoder to identify hidden states from spike-LFP activity. We then decode the movement trajectories using these hidden states. We find that the identified states can accurately decode the trajectories. Moreover, we demonstrate that adding LFP to spikes improves the decoding accuracy, suggesting that our unsupervised learning algorithm incorporates information across scales. This learning algorithm could serve as a new tool to study encoding across scales and to enhance future BMI systems.",
author = "Hamidreza Abbaspourazad and Yan Wong and Bijan Pesaran and Shanechi, {Maryam M.}",
year = "2018",
month = "10",
day = "26",
doi = "10.1109/EMBC.2018.8513242",
language = "English (US)",
volume = "2018-July",
pages = "3778--3781",
booktitle = "40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Identifying multiscale hidden states to decode behavior

AU - Abbaspourazad, Hamidreza

AU - Wong, Yan

AU - Pesaran, Bijan

AU - Shanechi, Maryam M.

PY - 2018/10/26

Y1 - 2018/10/26

N2 - A key element needed in a brain-machine interface (BMI) decoder is the encoding model, which relates the neural activity to intended movement. The vast majority of work have used a representational encoding model, which assumes movement parameters are directly encoded in neural activity. Recent work have in turn suggested the existence of neural dynamics that represent behavior. This recent evidence motivates developing dynamical encoding models with hidden states that encode movement. Regardless of their type, encoding models have vastly characterized a single scale of activity, e.g., either spikes or local field potentials (LFP). In our recent work we developed a multiscale representational encoding model to simultaneously characterize and decode discrete spikes and continuous field activity. However, learning a multiscale dynamical model from simultaneous spike-field recordings in the presence of hidden states is challenging. Here we present an unsupervised learning algorithm for estimating a multiscale state-space model with hidden states and validate it using spike-LFP activity during a reaching movement. We use the learned multiscale statespace model and a corresponding decoder to identify hidden states from spike-LFP activity. We then decode the movement trajectories using these hidden states. We find that the identified states can accurately decode the trajectories. Moreover, we demonstrate that adding LFP to spikes improves the decoding accuracy, suggesting that our unsupervised learning algorithm incorporates information across scales. This learning algorithm could serve as a new tool to study encoding across scales and to enhance future BMI systems.

AB - A key element needed in a brain-machine interface (BMI) decoder is the encoding model, which relates the neural activity to intended movement. The vast majority of work have used a representational encoding model, which assumes movement parameters are directly encoded in neural activity. Recent work have in turn suggested the existence of neural dynamics that represent behavior. This recent evidence motivates developing dynamical encoding models with hidden states that encode movement. Regardless of their type, encoding models have vastly characterized a single scale of activity, e.g., either spikes or local field potentials (LFP). In our recent work we developed a multiscale representational encoding model to simultaneously characterize and decode discrete spikes and continuous field activity. However, learning a multiscale dynamical model from simultaneous spike-field recordings in the presence of hidden states is challenging. Here we present an unsupervised learning algorithm for estimating a multiscale state-space model with hidden states and validate it using spike-LFP activity during a reaching movement. We use the learned multiscale statespace model and a corresponding decoder to identify hidden states from spike-LFP activity. We then decode the movement trajectories using these hidden states. We find that the identified states can accurately decode the trajectories. Moreover, we demonstrate that adding LFP to spikes improves the decoding accuracy, suggesting that our unsupervised learning algorithm incorporates information across scales. This learning algorithm could serve as a new tool to study encoding across scales and to enhance future BMI systems.

UR - http://www.scopus.com/inward/record.url?scp=85056620375&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056620375&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2018.8513242

DO - 10.1109/EMBC.2018.8513242

M3 - Conference contribution

C2 - 30441189

AN - SCOPUS:85056620375

VL - 2018-July

SP - 3778

EP - 3781

BT - 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -