Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy

Soren Fons Vind Nielsen, Yuri Levin-Schwartz, Diego Vidaurre, Tulay Adali, Vince Daniel Calhoun, Kristoffer H. Madsen, Lars Kai Hansen, Morten Morup

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

Abstract

Dynamic functional connectivity has become a prominent approach for tracking the changes of macroscale statistical dependencies between regions in the brain. Effective parametrization of these statistical dependencies, referred to as brain states, is however still an open problem. We investigate different emission models in the hidden Markov model framework, each representing certain assumptions about dynamic changes in the brain. We evaluate each model by how well they can discriminate between schizophrenic patients and healthy controls based on a group independent component analysis of resting-state functional magnetic resonance imaging data. We find that simple emission models without full covariance matrices can achieve similar classification results as the models with more parameters. This raises questions about the predictability of dynamic functional connectivity in comparison to simpler dynamic features when used as biomarkers. However, we must stress that there is a distinction between characterization and classification, which has to be investigated further.

Original languageEnglish (US)
Title of host publication2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2566-2570
Number of pages5
Volume2018-April
ISBN (Print)9781538646588
DOIs
StatePublished - Sep 10 2018
Externally publishedYes
Event2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Calgary, Canada
Duration: Apr 15 2018Apr 20 2018

Other

Other2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
CountryCanada
CityCalgary
Period4/15/184/20/18

Fingerprint

Brain
Independent component analysis
Biomarkers
Hidden Markov models
Covariance matrix
Magnetic Resonance Imaging

Keywords

  • Classification
  • Dynamic functional connectivity
  • Hidden Markov models
  • Schizophrenia

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Nielsen, S. F. V., Levin-Schwartz, Y., Vidaurre, D., Adali, T., Calhoun, V. D., Madsen, K. H., ... Morup, M. (2018). Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings (Vol. 2018-April, pp. 2566-2570). [8462310] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2018.8462310

Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy. / Nielsen, Soren Fons Vind; Levin-Schwartz, Yuri; Vidaurre, Diego; Adali, Tulay; Calhoun, Vince Daniel; Madsen, Kristoffer H.; Hansen, Lars Kai; Morup, Morten.

2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. p. 2566-2570 8462310.

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

Nielsen, SFV, Levin-Schwartz, Y, Vidaurre, D, Adali, T, Calhoun, VD, Madsen, KH, Hansen, LK & Morup, M 2018, Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy. in 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. vol. 2018-April, 8462310, Institute of Electrical and Electronics Engineers Inc., pp. 2566-2570, 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018, Calgary, Canada, 4/15/18. https://doi.org/10.1109/ICASSP.2018.8462310
Nielsen SFV, Levin-Schwartz Y, Vidaurre D, Adali T, Calhoun VD, Madsen KH et al. Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April. Institute of Electrical and Electronics Engineers Inc. 2018. p. 2566-2570. 8462310 https://doi.org/10.1109/ICASSP.2018.8462310
Nielsen, Soren Fons Vind ; Levin-Schwartz, Yuri ; Vidaurre, Diego ; Adali, Tulay ; Calhoun, Vince Daniel ; Madsen, Kristoffer H. ; Hansen, Lars Kai ; Morup, Morten. / Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy. 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. pp. 2566-2570
@inproceedings{8fdf3b17edf147f5acde4bd5e4c17614,
title = "Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy",
abstract = "Dynamic functional connectivity has become a prominent approach for tracking the changes of macroscale statistical dependencies between regions in the brain. Effective parametrization of these statistical dependencies, referred to as brain states, is however still an open problem. We investigate different emission models in the hidden Markov model framework, each representing certain assumptions about dynamic changes in the brain. We evaluate each model by how well they can discriminate between schizophrenic patients and healthy controls based on a group independent component analysis of resting-state functional magnetic resonance imaging data. We find that simple emission models without full covariance matrices can achieve similar classification results as the models with more parameters. This raises questions about the predictability of dynamic functional connectivity in comparison to simpler dynamic features when used as biomarkers. However, we must stress that there is a distinction between characterization and classification, which has to be investigated further.",
keywords = "Classification, Dynamic functional connectivity, Hidden Markov models, Schizophrenia",
author = "Nielsen, {Soren Fons Vind} and Yuri Levin-Schwartz and Diego Vidaurre and Tulay Adali and Calhoun, {Vince Daniel} and Madsen, {Kristoffer H.} and Hansen, {Lars Kai} and Morten Morup",
year = "2018",
month = "9",
day = "10",
doi = "10.1109/ICASSP.2018.8462310",
language = "English (US)",
isbn = "9781538646588",
volume = "2018-April",
pages = "2566--2570",
booktitle = "2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Evaluating Models of Dynamic Functional Connectivity Using Predictive Classification Accuracy

AU - Nielsen, Soren Fons Vind

AU - Levin-Schwartz, Yuri

AU - Vidaurre, Diego

AU - Adali, Tulay

AU - Calhoun, Vince Daniel

AU - Madsen, Kristoffer H.

AU - Hansen, Lars Kai

AU - Morup, Morten

PY - 2018/9/10

Y1 - 2018/9/10

N2 - Dynamic functional connectivity has become a prominent approach for tracking the changes of macroscale statistical dependencies between regions in the brain. Effective parametrization of these statistical dependencies, referred to as brain states, is however still an open problem. We investigate different emission models in the hidden Markov model framework, each representing certain assumptions about dynamic changes in the brain. We evaluate each model by how well they can discriminate between schizophrenic patients and healthy controls based on a group independent component analysis of resting-state functional magnetic resonance imaging data. We find that simple emission models without full covariance matrices can achieve similar classification results as the models with more parameters. This raises questions about the predictability of dynamic functional connectivity in comparison to simpler dynamic features when used as biomarkers. However, we must stress that there is a distinction between characterization and classification, which has to be investigated further.

AB - Dynamic functional connectivity has become a prominent approach for tracking the changes of macroscale statistical dependencies between regions in the brain. Effective parametrization of these statistical dependencies, referred to as brain states, is however still an open problem. We investigate different emission models in the hidden Markov model framework, each representing certain assumptions about dynamic changes in the brain. We evaluate each model by how well they can discriminate between schizophrenic patients and healthy controls based on a group independent component analysis of resting-state functional magnetic resonance imaging data. We find that simple emission models without full covariance matrices can achieve similar classification results as the models with more parameters. This raises questions about the predictability of dynamic functional connectivity in comparison to simpler dynamic features when used as biomarkers. However, we must stress that there is a distinction between characterization and classification, which has to be investigated further.

KW - Classification

KW - Dynamic functional connectivity

KW - Hidden Markov models

KW - Schizophrenia

UR - http://www.scopus.com/inward/record.url?scp=85054211560&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054211560&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2018.8462310

DO - 10.1109/ICASSP.2018.8462310

M3 - Conference contribution

AN - SCOPUS:85054211560

SN - 9781538646588

VL - 2018-April

SP - 2566

EP - 2570

BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -