Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data

Debbrata K. Saha, Anees Abrol, Eswar Damaraju, Barnaly Rashid, Sergey M. Plis, Vince D. Calhoun

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

Abstract

Currently, there is extensive research ongoing to analyze the dynamics of brain connectivity in resting fMRI data. In such approaches, there is often a selection of the number of connectivity states which are present within the data. Typically, an order selection approach such as the elbow criteria is used for this when time-varying connectivity estimates are acquired using sliding-window approach and clustered using k-means algorithm. In this work we evaluate the benefits of using classification (e.g. of patients versus controls) as a criterion for evaluating the optimal number of states (or clusters). We compare different classification strategies to perform the classification while optimizing for the number of clusters (selected via a k-means approach). In our approach, we compute cross-validated classification accuracy and variability at different numbers of states for healthy control (HC) versus schizophrenia (SZ) patients. Consistent with our earlier reports, we find improvement in classification performance when dynamic connectivity measures are combined with static connectivity measures. We also show that the model order at which maximal classification accuracy is obtained (four dynamic states for this data) can be different from the order obtained using standard k-means model order selection methods (that result in five states for the data at hand) across different combinations of features trained. We also investigate if additional information from hierarchical clustering of first level states can contribute to the performance of classification accuracy and observe no evidence for additional sub-clusters in short 5-minute resting scans. In sum, the results suggest the use of classification accuracy as a promising metric for selecting the number of states in a dynamic connectivity analysis.

Original languageEnglish (US)
Title of host publicationISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging
PublisherIEEE Computer Society
Pages1602-1605
Number of pages4
ISBN (Electronic)9781538636411
DOIs
StatePublished - Apr 2019
Externally publishedYes
Event16th IEEE International Symposium on Biomedical Imaging, ISBI 2019 - Venice, Italy
Duration: Apr 8 2019Apr 11 2019

Publication series

NameProceedings - International Symposium on Biomedical Imaging
Volume2019-April
ISSN (Print)1945-7928
ISSN (Electronic)1945-8452

Conference

Conference16th IEEE International Symposium on Biomedical Imaging, ISBI 2019
CountryItaly
CityVenice
Period4/8/194/11/19

Fingerprint

Magnetic Resonance Imaging
Elbow
Dynamic analysis
Cluster Analysis
Brain
Schizophrenia
Hand
Research

Keywords

  • Classification
  • DFNC
  • Hierarchical clustering
  • Model order

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Saha, D. K., Abrol, A., Damaraju, E., Rashid, B., Plis, S. M., & Calhoun, V. D. (2019). Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data. In ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging (pp. 1602-1605). [8759146] (Proceedings - International Symposium on Biomedical Imaging; Vol. 2019-April). IEEE Computer Society. https://doi.org/10.1109/ISBI.2019.8759146

Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data. / Saha, Debbrata K.; Abrol, Anees; Damaraju, Eswar; Rashid, Barnaly; Plis, Sergey M.; Calhoun, Vince D.

ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE Computer Society, 2019. p. 1602-1605 8759146 (Proceedings - International Symposium on Biomedical Imaging; Vol. 2019-April).

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

Saha, DK, Abrol, A, Damaraju, E, Rashid, B, Plis, SM & Calhoun, VD 2019, Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data. in ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging., 8759146, Proceedings - International Symposium on Biomedical Imaging, vol. 2019-April, IEEE Computer Society, pp. 1602-1605, 16th IEEE International Symposium on Biomedical Imaging, ISBI 2019, Venice, Italy, 4/8/19. https://doi.org/10.1109/ISBI.2019.8759146
Saha DK, Abrol A, Damaraju E, Rashid B, Plis SM, Calhoun VD. Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data. In ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE Computer Society. 2019. p. 1602-1605. 8759146. (Proceedings - International Symposium on Biomedical Imaging). https://doi.org/10.1109/ISBI.2019.8759146
Saha, Debbrata K. ; Abrol, Anees ; Damaraju, Eswar ; Rashid, Barnaly ; Plis, Sergey M. ; Calhoun, Vince D. / Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data. ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging. IEEE Computer Society, 2019. pp. 1602-1605 (Proceedings - International Symposium on Biomedical Imaging).
@inproceedings{6a837f8b9cde4ff0b53411fc3e2c018c,
title = "Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data",
abstract = "Currently, there is extensive research ongoing to analyze the dynamics of brain connectivity in resting fMRI data. In such approaches, there is often a selection of the number of connectivity states which are present within the data. Typically, an order selection approach such as the elbow criteria is used for this when time-varying connectivity estimates are acquired using sliding-window approach and clustered using k-means algorithm. In this work we evaluate the benefits of using classification (e.g. of patients versus controls) as a criterion for evaluating the optimal number of states (or clusters). We compare different classification strategies to perform the classification while optimizing for the number of clusters (selected via a k-means approach). In our approach, we compute cross-validated classification accuracy and variability at different numbers of states for healthy control (HC) versus schizophrenia (SZ) patients. Consistent with our earlier reports, we find improvement in classification performance when dynamic connectivity measures are combined with static connectivity measures. We also show that the model order at which maximal classification accuracy is obtained (four dynamic states for this data) can be different from the order obtained using standard k-means model order selection methods (that result in five states for the data at hand) across different combinations of features trained. We also investigate if additional information from hierarchical clustering of first level states can contribute to the performance of classification accuracy and observe no evidence for additional sub-clusters in short 5-minute resting scans. In sum, the results suggest the use of classification accuracy as a promising metric for selecting the number of states in a dynamic connectivity analysis.",
keywords = "Classification, DFNC, Hierarchical clustering, Model order",
author = "Saha, {Debbrata K.} and Anees Abrol and Eswar Damaraju and Barnaly Rashid and Plis, {Sergey M.} and Calhoun, {Vince D.}",
year = "2019",
month = "4",
doi = "10.1109/ISBI.2019.8759146",
language = "English (US)",
series = "Proceedings - International Symposium on Biomedical Imaging",
publisher = "IEEE Computer Society",
pages = "1602--1605",
booktitle = "ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging",

}

TY - GEN

T1 - Classification as a criterion to select model order for dynamic functional connectivity states in rest-fMRI data

AU - Saha, Debbrata K.

AU - Abrol, Anees

AU - Damaraju, Eswar

AU - Rashid, Barnaly

AU - Plis, Sergey M.

AU - Calhoun, Vince D.

PY - 2019/4

Y1 - 2019/4

N2 - Currently, there is extensive research ongoing to analyze the dynamics of brain connectivity in resting fMRI data. In such approaches, there is often a selection of the number of connectivity states which are present within the data. Typically, an order selection approach such as the elbow criteria is used for this when time-varying connectivity estimates are acquired using sliding-window approach and clustered using k-means algorithm. In this work we evaluate the benefits of using classification (e.g. of patients versus controls) as a criterion for evaluating the optimal number of states (or clusters). We compare different classification strategies to perform the classification while optimizing for the number of clusters (selected via a k-means approach). In our approach, we compute cross-validated classification accuracy and variability at different numbers of states for healthy control (HC) versus schizophrenia (SZ) patients. Consistent with our earlier reports, we find improvement in classification performance when dynamic connectivity measures are combined with static connectivity measures. We also show that the model order at which maximal classification accuracy is obtained (four dynamic states for this data) can be different from the order obtained using standard k-means model order selection methods (that result in five states for the data at hand) across different combinations of features trained. We also investigate if additional information from hierarchical clustering of first level states can contribute to the performance of classification accuracy and observe no evidence for additional sub-clusters in short 5-minute resting scans. In sum, the results suggest the use of classification accuracy as a promising metric for selecting the number of states in a dynamic connectivity analysis.

AB - Currently, there is extensive research ongoing to analyze the dynamics of brain connectivity in resting fMRI data. In such approaches, there is often a selection of the number of connectivity states which are present within the data. Typically, an order selection approach such as the elbow criteria is used for this when time-varying connectivity estimates are acquired using sliding-window approach and clustered using k-means algorithm. In this work we evaluate the benefits of using classification (e.g. of patients versus controls) as a criterion for evaluating the optimal number of states (or clusters). We compare different classification strategies to perform the classification while optimizing for the number of clusters (selected via a k-means approach). In our approach, we compute cross-validated classification accuracy and variability at different numbers of states for healthy control (HC) versus schizophrenia (SZ) patients. Consistent with our earlier reports, we find improvement in classification performance when dynamic connectivity measures are combined with static connectivity measures. We also show that the model order at which maximal classification accuracy is obtained (four dynamic states for this data) can be different from the order obtained using standard k-means model order selection methods (that result in five states for the data at hand) across different combinations of features trained. We also investigate if additional information from hierarchical clustering of first level states can contribute to the performance of classification accuracy and observe no evidence for additional sub-clusters in short 5-minute resting scans. In sum, the results suggest the use of classification accuracy as a promising metric for selecting the number of states in a dynamic connectivity analysis.

KW - Classification

KW - DFNC

KW - Hierarchical clustering

KW - Model order

UR - http://www.scopus.com/inward/record.url?scp=85073891053&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073891053&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2019.8759146

DO - 10.1109/ISBI.2019.8759146

M3 - Conference contribution

AN - SCOPUS:85073891053

T3 - Proceedings - International Symposium on Biomedical Imaging

SP - 1602

EP - 1605

BT - ISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging

PB - IEEE Computer Society

ER -