Optimal control-based bayesian detection of clinical and behavioral state transitions

Sabato Santaniello, David L. Sherman, Nitish V. Thakor, Emad N. Eskandar, Sridevi V. Sarma

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

Accurately detecting hidden clinical or behavioral states from sequential measurements is an emerging topic in neuroscience and medicine, which may dramatically impact neural prosthetics, brain-computer interface and drug delivery. For example, early detection of an epileptic seizure from sequential electroencephalographic (EEG) measurements would allow timely administration of anticonvulsant drugs or neurostimulation, thus reducing physical impairment and risks of overtreatment. We develop a Bayesian paradigm for state transition detection that combines optimal control and Markov processes. We define a hidden Markov model of the state evolution and develop a detection policy that minimizes a loss function of both probability of false positives and accuracy (i.e., lag between estimated and actual transition time). Our strategy automatically adapts to each newly acquired measurement based on the state evolution model and the relative loss for false positives and accuracy, thus resulting in a time varying threshold policy. The paradigm was used in two applications: 1) detection of movement onset (behavioral state) from subthalamic single unit recordings in Parkinson's disease patients performing a motor task; 2) early detection of an approaching seizure (clinical state) from multichannel intracranial EEG recordings in rodents treated with pentylenetetrazol chemoconvulsant. Our paradigm performs significantly better than chance and improves over widely used detection algorithms.

Original languageEnglish (US)
Article number6263308
Pages (from-to)708-719
Number of pages12
JournalIEEE Transactions on Neural Systems and Rehabilitation Engineering
Volume20
Issue number5
DOIs
StatePublished - Sep 17 2012
Externally publishedYes

Fingerprint

Brain-Computer Interfaces
Markov Chains
Pentylenetetrazole
Neurosciences
Anticonvulsants
Parkinson Disease
Rodentia
Epilepsy
Seizures
Brain computer interface
Medicine
Hidden Markov models
Prosthetics
Drug delivery
Markov processes
Pharmaceutical Preparations
Medical Overuse

Keywords

  • Bayesian estimation
  • neural systems
  • optimal control
  • quickest detection (QD)

ASJC Scopus subject areas

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

Cite this

Optimal control-based bayesian detection of clinical and behavioral state transitions. / Santaniello, Sabato; Sherman, David L.; Thakor, Nitish V.; Eskandar, Emad N.; Sarma, Sridevi V.

In: IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol. 20, No. 5, 6263308, 17.09.2012, p. 708-719.

Research output: Contribution to journalArticle

Santaniello, Sabato ; Sherman, David L. ; Thakor, Nitish V. ; Eskandar, Emad N. ; Sarma, Sridevi V. / Optimal control-based bayesian detection of clinical and behavioral state transitions. In: IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2012 ; Vol. 20, No. 5. pp. 708-719.
@article{a55f94b97bd0446ba1f725c45145cac6,
title = "Optimal control-based bayesian detection of clinical and behavioral state transitions",
abstract = "Accurately detecting hidden clinical or behavioral states from sequential measurements is an emerging topic in neuroscience and medicine, which may dramatically impact neural prosthetics, brain-computer interface and drug delivery. For example, early detection of an epileptic seizure from sequential electroencephalographic (EEG) measurements would allow timely administration of anticonvulsant drugs or neurostimulation, thus reducing physical impairment and risks of overtreatment. We develop a Bayesian paradigm for state transition detection that combines optimal control and Markov processes. We define a hidden Markov model of the state evolution and develop a detection policy that minimizes a loss function of both probability of false positives and accuracy (i.e., lag between estimated and actual transition time). Our strategy automatically adapts to each newly acquired measurement based on the state evolution model and the relative loss for false positives and accuracy, thus resulting in a time varying threshold policy. The paradigm was used in two applications: 1) detection of movement onset (behavioral state) from subthalamic single unit recordings in Parkinson's disease patients performing a motor task; 2) early detection of an approaching seizure (clinical state) from multichannel intracranial EEG recordings in rodents treated with pentylenetetrazol chemoconvulsant. Our paradigm performs significantly better than chance and improves over widely used detection algorithms.",
keywords = "Bayesian estimation, neural systems, optimal control, quickest detection (QD)",
author = "Sabato Santaniello and Sherman, {David L.} and Thakor, {Nitish V.} and Eskandar, {Emad N.} and Sarma, {Sridevi V.}",
year = "2012",
month = "9",
day = "17",
doi = "10.1109/TNSRE.2012.2210246",
language = "English (US)",
volume = "20",
pages = "708--719",
journal = "IEEE Transactions on Neural Systems and Rehabilitation Engineering",
issn = "1534-4320",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Optimal control-based bayesian detection of clinical and behavioral state transitions

AU - Santaniello, Sabato

AU - Sherman, David L.

AU - Thakor, Nitish V.

AU - Eskandar, Emad N.

AU - Sarma, Sridevi V.

PY - 2012/9/17

Y1 - 2012/9/17

N2 - Accurately detecting hidden clinical or behavioral states from sequential measurements is an emerging topic in neuroscience and medicine, which may dramatically impact neural prosthetics, brain-computer interface and drug delivery. For example, early detection of an epileptic seizure from sequential electroencephalographic (EEG) measurements would allow timely administration of anticonvulsant drugs or neurostimulation, thus reducing physical impairment and risks of overtreatment. We develop a Bayesian paradigm for state transition detection that combines optimal control and Markov processes. We define a hidden Markov model of the state evolution and develop a detection policy that minimizes a loss function of both probability of false positives and accuracy (i.e., lag between estimated and actual transition time). Our strategy automatically adapts to each newly acquired measurement based on the state evolution model and the relative loss for false positives and accuracy, thus resulting in a time varying threshold policy. The paradigm was used in two applications: 1) detection of movement onset (behavioral state) from subthalamic single unit recordings in Parkinson's disease patients performing a motor task; 2) early detection of an approaching seizure (clinical state) from multichannel intracranial EEG recordings in rodents treated with pentylenetetrazol chemoconvulsant. Our paradigm performs significantly better than chance and improves over widely used detection algorithms.

AB - Accurately detecting hidden clinical or behavioral states from sequential measurements is an emerging topic in neuroscience and medicine, which may dramatically impact neural prosthetics, brain-computer interface and drug delivery. For example, early detection of an epileptic seizure from sequential electroencephalographic (EEG) measurements would allow timely administration of anticonvulsant drugs or neurostimulation, thus reducing physical impairment and risks of overtreatment. We develop a Bayesian paradigm for state transition detection that combines optimal control and Markov processes. We define a hidden Markov model of the state evolution and develop a detection policy that minimizes a loss function of both probability of false positives and accuracy (i.e., lag between estimated and actual transition time). Our strategy automatically adapts to each newly acquired measurement based on the state evolution model and the relative loss for false positives and accuracy, thus resulting in a time varying threshold policy. The paradigm was used in two applications: 1) detection of movement onset (behavioral state) from subthalamic single unit recordings in Parkinson's disease patients performing a motor task; 2) early detection of an approaching seizure (clinical state) from multichannel intracranial EEG recordings in rodents treated with pentylenetetrazol chemoconvulsant. Our paradigm performs significantly better than chance and improves over widely used detection algorithms.

KW - Bayesian estimation

KW - neural systems

KW - optimal control

KW - quickest detection (QD)

UR - http://www.scopus.com/inward/record.url?scp=84866079024&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84866079024&partnerID=8YFLogxK

U2 - 10.1109/TNSRE.2012.2210246

DO - 10.1109/TNSRE.2012.2210246

M3 - Article

C2 - 22893447

AN - SCOPUS:84866079024

VL - 20

SP - 708

EP - 719

JO - IEEE Transactions on Neural Systems and Rehabilitation Engineering

JF - IEEE Transactions on Neural Systems and Rehabilitation Engineering

SN - 1534-4320

IS - 5

M1 - 6263308

ER -