Fast and stable signal deconvolution via compressible state-space models

Abbas Kazemipour, Ji Liu, Krystyna Solarana, Daniel A. Nagode, Patrick Kanold, Min Wu, Behtash Babadi

Research output: Contribution to journalArticle

Abstract

Objective: Common biological measurements are in the form of noisy convolutions of signals of interest with possibly unknown and transient blurring kernels. Examples include EEG and calcium imaging data. Thus, signal deconvolution of these measurements is crucial in understanding the underlying biological processes. The objective of this paper is to develop fast and stable solutions for signal deconvolution from noisy, blurred, and undersampled data, where the signals are in the form of discrete events distributed in time and space. Methods: We introduce compressible state-space models as a framework to model and estimate such discrete events. These state-space models admit abrupt changes in the states and have a convergent transition matrix, and are coupled with compressive linear measurements. We consider a dynamic compressive sensing optimization problem and develop a fast solution, using two nested expectation maximization algorithms, to jointly estimate the states as well as their transition matrices. Under suitable sparsity assumptions on the dynamics, we prove optimal stability guarantees for the recovery of the states and present a method for the identification of the underlying discrete events with precise confidence bounds. Results: We present simulation studies as well as application to calcium deconvolution and sleep spindle detection, which verify our theoretical results and show significant improvement over existing techniques. Conclusion: Our results show that by explicitly modeling the dynamics of the underlying signals, it is possible to construct signal deconvolution solutions that are scalable, statistically robust, and achieve high temporal resolution. Significance: Our proposed methodology provides a framework for modeling and deconvolution of noisy, blurred, and undersampled measurements in a fast and stable fashion, with potential application to a wide range of biological data.

Original languageEnglish (US)
Pages (from-to)74-86
Number of pages13
JournalIEEE Transactions on Biomedical Engineering
Volume65
Issue number1
DOIs
StatePublished - Jan 1 2018
Externally publishedYes

Fingerprint

Deconvolution
Calcium
Electroencephalography
Convolution
Imaging techniques
Recovery

Keywords

  • Calcium imaging
  • Compressive sensing
  • Signal deconvolution
  • Sleep spindles
  • State-space models

ASJC Scopus subject areas

  • Biomedical Engineering

Cite this

Fast and stable signal deconvolution via compressible state-space models. / Kazemipour, Abbas; Liu, Ji; Solarana, Krystyna; Nagode, Daniel A.; Kanold, Patrick; Wu, Min; Babadi, Behtash.

In: IEEE Transactions on Biomedical Engineering, Vol. 65, No. 1, 01.01.2018, p. 74-86.

Research output: Contribution to journalArticle

Kazemipour, Abbas ; Liu, Ji ; Solarana, Krystyna ; Nagode, Daniel A. ; Kanold, Patrick ; Wu, Min ; Babadi, Behtash. / Fast and stable signal deconvolution via compressible state-space models. In: IEEE Transactions on Biomedical Engineering. 2018 ; Vol. 65, No. 1. pp. 74-86.
@article{b95223896aa94cfeb61df186f3f56f89,
title = "Fast and stable signal deconvolution via compressible state-space models",
abstract = "Objective: Common biological measurements are in the form of noisy convolutions of signals of interest with possibly unknown and transient blurring kernels. Examples include EEG and calcium imaging data. Thus, signal deconvolution of these measurements is crucial in understanding the underlying biological processes. The objective of this paper is to develop fast and stable solutions for signal deconvolution from noisy, blurred, and undersampled data, where the signals are in the form of discrete events distributed in time and space. Methods: We introduce compressible state-space models as a framework to model and estimate such discrete events. These state-space models admit abrupt changes in the states and have a convergent transition matrix, and are coupled with compressive linear measurements. We consider a dynamic compressive sensing optimization problem and develop a fast solution, using two nested expectation maximization algorithms, to jointly estimate the states as well as their transition matrices. Under suitable sparsity assumptions on the dynamics, we prove optimal stability guarantees for the recovery of the states and present a method for the identification of the underlying discrete events with precise confidence bounds. Results: We present simulation studies as well as application to calcium deconvolution and sleep spindle detection, which verify our theoretical results and show significant improvement over existing techniques. Conclusion: Our results show that by explicitly modeling the dynamics of the underlying signals, it is possible to construct signal deconvolution solutions that are scalable, statistically robust, and achieve high temporal resolution. Significance: Our proposed methodology provides a framework for modeling and deconvolution of noisy, blurred, and undersampled measurements in a fast and stable fashion, with potential application to a wide range of biological data.",
keywords = "Calcium imaging, Compressive sensing, Signal deconvolution, Sleep spindles, State-space models",
author = "Abbas Kazemipour and Ji Liu and Krystyna Solarana and Nagode, {Daniel A.} and Patrick Kanold and Min Wu and Behtash Babadi",
year = "2018",
month = "1",
day = "1",
doi = "10.1109/TBME.2017.2694339",
language = "English (US)",
volume = "65",
pages = "74--86",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",
number = "1",

}

TY - JOUR

T1 - Fast and stable signal deconvolution via compressible state-space models

AU - Kazemipour, Abbas

AU - Liu, Ji

AU - Solarana, Krystyna

AU - Nagode, Daniel A.

AU - Kanold, Patrick

AU - Wu, Min

AU - Babadi, Behtash

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Objective: Common biological measurements are in the form of noisy convolutions of signals of interest with possibly unknown and transient blurring kernels. Examples include EEG and calcium imaging data. Thus, signal deconvolution of these measurements is crucial in understanding the underlying biological processes. The objective of this paper is to develop fast and stable solutions for signal deconvolution from noisy, blurred, and undersampled data, where the signals are in the form of discrete events distributed in time and space. Methods: We introduce compressible state-space models as a framework to model and estimate such discrete events. These state-space models admit abrupt changes in the states and have a convergent transition matrix, and are coupled with compressive linear measurements. We consider a dynamic compressive sensing optimization problem and develop a fast solution, using two nested expectation maximization algorithms, to jointly estimate the states as well as their transition matrices. Under suitable sparsity assumptions on the dynamics, we prove optimal stability guarantees for the recovery of the states and present a method for the identification of the underlying discrete events with precise confidence bounds. Results: We present simulation studies as well as application to calcium deconvolution and sleep spindle detection, which verify our theoretical results and show significant improvement over existing techniques. Conclusion: Our results show that by explicitly modeling the dynamics of the underlying signals, it is possible to construct signal deconvolution solutions that are scalable, statistically robust, and achieve high temporal resolution. Significance: Our proposed methodology provides a framework for modeling and deconvolution of noisy, blurred, and undersampled measurements in a fast and stable fashion, with potential application to a wide range of biological data.

AB - Objective: Common biological measurements are in the form of noisy convolutions of signals of interest with possibly unknown and transient blurring kernels. Examples include EEG and calcium imaging data. Thus, signal deconvolution of these measurements is crucial in understanding the underlying biological processes. The objective of this paper is to develop fast and stable solutions for signal deconvolution from noisy, blurred, and undersampled data, where the signals are in the form of discrete events distributed in time and space. Methods: We introduce compressible state-space models as a framework to model and estimate such discrete events. These state-space models admit abrupt changes in the states and have a convergent transition matrix, and are coupled with compressive linear measurements. We consider a dynamic compressive sensing optimization problem and develop a fast solution, using two nested expectation maximization algorithms, to jointly estimate the states as well as their transition matrices. Under suitable sparsity assumptions on the dynamics, we prove optimal stability guarantees for the recovery of the states and present a method for the identification of the underlying discrete events with precise confidence bounds. Results: We present simulation studies as well as application to calcium deconvolution and sleep spindle detection, which verify our theoretical results and show significant improvement over existing techniques. Conclusion: Our results show that by explicitly modeling the dynamics of the underlying signals, it is possible to construct signal deconvolution solutions that are scalable, statistically robust, and achieve high temporal resolution. Significance: Our proposed methodology provides a framework for modeling and deconvolution of noisy, blurred, and undersampled measurements in a fast and stable fashion, with potential application to a wide range of biological data.

KW - Calcium imaging

KW - Compressive sensing

KW - Signal deconvolution

KW - Sleep spindles

KW - State-space models

UR - http://www.scopus.com/inward/record.url?scp=85043518713&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85043518713&partnerID=8YFLogxK

U2 - 10.1109/TBME.2017.2694339

DO - 10.1109/TBME.2017.2694339

M3 - Article

VL - 65

SP - 74

EP - 86

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 1

ER -