Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks

Yue Qiu, Qiu Hua Lin, Li Dan Kuang, Wen Da Zhao, Xiao Feng Gong, Fengyu Cong, Vince D. Calhoun

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

Abstract

Deep learning has contributed greatly to functional magnetic resonance imaging (fMRI) analysis, however, spatial maps derived from fMRI data by independent component analysis (ICA), as promising biomarkers, have rarely been directly used to perform individualized diagnosis. As such, this study proposes a novel framework combining ICA and convolutional neural network (CNN) for classifying schizophrenia patients (SZs) and healthy controls (HCs). ICA is first used to obtain components of interest which have been previously implicated in schizophrenia. Functionally informative slices of these components are then selected and labelled. CNN is finally employed to learn hierarchical diagnostic features from the slices and classify SZs and HCs. We use complex-valued fMRI data instead of magnitude fMRI data, in order to obtain more contiguous spatial activations. Spatial maps estimated by ICA with multiple model orders are employed for data argumentation to enhance the training process. Evaluations are performed using 82 resting-state complex-valued fMRI datasets including 42 SZs and 40 HCs. The proposed method shows an average accuracy of 72.65% in the default mode network and 78.34% in the auditory cortex for slice-level classification. When performing subject-level classification based on majority voting, the result shows 91.32% and 98.75% average accuracy, highlighting the potential of the proposed method for diagnosis of schizophrenia and other neurological diseases.

Original languageEnglish (US)
Title of host publicationAdvances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings
EditorsHuajin Tang, Zhanshan Wang, Huchuan Lu
PublisherSpringer Verlag
Pages540-547
Number of pages8
ISBN (Print)9783030228071
DOIs
StatePublished - Jan 1 2019
Externally publishedYes
Event16th International Symposium on Neural Networks, ISNN 2019 - Moscow, Russian Federation
Duration: Jul 10 2019Jul 12 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11555 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Symposium on Neural Networks, ISNN 2019
CountryRussian Federation
CityMoscow
Period7/10/197/12/19

Fingerprint

Functional Magnetic Resonance Imaging
Independent component analysis
Independent Component Analysis
Neural Networks
Neural networks
Slice
Majority Voting
Spatial Analysis
Argumentation
Biomarkers
Multiple Models
Cortex
Activation
Diagnostics
Chemical activation
Classify
Magnetic Resonance Imaging
Evaluation

Keywords

  • Data argumentation
  • Deep learning
  • fMRI
  • ICA
  • Model order
  • Schizophrenia

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Qiu, Y., Lin, Q. H., Kuang, L. D., Zhao, W. D., Gong, X. F., Cong, F., & Calhoun, V. D. (2019). Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks. In H. Tang, Z. Wang, & H. Lu (Eds.), Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings (pp. 540-547). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11555 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-22808-8_53

Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks. / Qiu, Yue; Lin, Qiu Hua; Kuang, Li Dan; Zhao, Wen Da; Gong, Xiao Feng; Cong, Fengyu; Calhoun, Vince D.

Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings. ed. / Huajin Tang; Zhanshan Wang; Huchuan Lu. Springer Verlag, 2019. p. 540-547 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11555 LNCS).

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

Qiu, Y, Lin, QH, Kuang, LD, Zhao, WD, Gong, XF, Cong, F & Calhoun, VD 2019, Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks. in H Tang, Z Wang & H Lu (eds), Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11555 LNCS, Springer Verlag, pp. 540-547, 16th International Symposium on Neural Networks, ISNN 2019, Moscow, Russian Federation, 7/10/19. https://doi.org/10.1007/978-3-030-22808-8_53
Qiu Y, Lin QH, Kuang LD, Zhao WD, Gong XF, Cong F et al. Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks. In Tang H, Wang Z, Lu H, editors, Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings. Springer Verlag. 2019. p. 540-547. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-22808-8_53
Qiu, Yue ; Lin, Qiu Hua ; Kuang, Li Dan ; Zhao, Wen Da ; Gong, Xiao Feng ; Cong, Fengyu ; Calhoun, Vince D. / Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks. Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings. editor / Huajin Tang ; Zhanshan Wang ; Huchuan Lu. Springer Verlag, 2019. pp. 540-547 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2480b588a2404945b23b2c0b5bf2c87e,
title = "Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks",
abstract = "Deep learning has contributed greatly to functional magnetic resonance imaging (fMRI) analysis, however, spatial maps derived from fMRI data by independent component analysis (ICA), as promising biomarkers, have rarely been directly used to perform individualized diagnosis. As such, this study proposes a novel framework combining ICA and convolutional neural network (CNN) for classifying schizophrenia patients (SZs) and healthy controls (HCs). ICA is first used to obtain components of interest which have been previously implicated in schizophrenia. Functionally informative slices of these components are then selected and labelled. CNN is finally employed to learn hierarchical diagnostic features from the slices and classify SZs and HCs. We use complex-valued fMRI data instead of magnitude fMRI data, in order to obtain more contiguous spatial activations. Spatial maps estimated by ICA with multiple model orders are employed for data argumentation to enhance the training process. Evaluations are performed using 82 resting-state complex-valued fMRI datasets including 42 SZs and 40 HCs. The proposed method shows an average accuracy of 72.65{\%} in the default mode network and 78.34{\%} in the auditory cortex for slice-level classification. When performing subject-level classification based on majority voting, the result shows 91.32{\%} and 98.75{\%} average accuracy, highlighting the potential of the proposed method for diagnosis of schizophrenia and other neurological diseases.",
keywords = "Data argumentation, Deep learning, fMRI, ICA, Model order, Schizophrenia",
author = "Yue Qiu and Lin, {Qiu Hua} and Kuang, {Li Dan} and Zhao, {Wen Da} and Gong, {Xiao Feng} and Fengyu Cong and Calhoun, {Vince D.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-22808-8_53",
language = "English (US)",
isbn = "9783030228071",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "540--547",
editor = "Huajin Tang and Zhanshan Wang and Huchuan Lu",
booktitle = "Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings",

}

TY - GEN

T1 - Classification of Schizophrenia Patients and Healthy Controls Using ICA of Complex-Valued fMRI Data and Convolutional Neural Networks

AU - Qiu, Yue

AU - Lin, Qiu Hua

AU - Kuang, Li Dan

AU - Zhao, Wen Da

AU - Gong, Xiao Feng

AU - Cong, Fengyu

AU - Calhoun, Vince D.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Deep learning has contributed greatly to functional magnetic resonance imaging (fMRI) analysis, however, spatial maps derived from fMRI data by independent component analysis (ICA), as promising biomarkers, have rarely been directly used to perform individualized diagnosis. As such, this study proposes a novel framework combining ICA and convolutional neural network (CNN) for classifying schizophrenia patients (SZs) and healthy controls (HCs). ICA is first used to obtain components of interest which have been previously implicated in schizophrenia. Functionally informative slices of these components are then selected and labelled. CNN is finally employed to learn hierarchical diagnostic features from the slices and classify SZs and HCs. We use complex-valued fMRI data instead of magnitude fMRI data, in order to obtain more contiguous spatial activations. Spatial maps estimated by ICA with multiple model orders are employed for data argumentation to enhance the training process. Evaluations are performed using 82 resting-state complex-valued fMRI datasets including 42 SZs and 40 HCs. The proposed method shows an average accuracy of 72.65% in the default mode network and 78.34% in the auditory cortex for slice-level classification. When performing subject-level classification based on majority voting, the result shows 91.32% and 98.75% average accuracy, highlighting the potential of the proposed method for diagnosis of schizophrenia and other neurological diseases.

AB - Deep learning has contributed greatly to functional magnetic resonance imaging (fMRI) analysis, however, spatial maps derived from fMRI data by independent component analysis (ICA), as promising biomarkers, have rarely been directly used to perform individualized diagnosis. As such, this study proposes a novel framework combining ICA and convolutional neural network (CNN) for classifying schizophrenia patients (SZs) and healthy controls (HCs). ICA is first used to obtain components of interest which have been previously implicated in schizophrenia. Functionally informative slices of these components are then selected and labelled. CNN is finally employed to learn hierarchical diagnostic features from the slices and classify SZs and HCs. We use complex-valued fMRI data instead of magnitude fMRI data, in order to obtain more contiguous spatial activations. Spatial maps estimated by ICA with multiple model orders are employed for data argumentation to enhance the training process. Evaluations are performed using 82 resting-state complex-valued fMRI datasets including 42 SZs and 40 HCs. The proposed method shows an average accuracy of 72.65% in the default mode network and 78.34% in the auditory cortex for slice-level classification. When performing subject-level classification based on majority voting, the result shows 91.32% and 98.75% average accuracy, highlighting the potential of the proposed method for diagnosis of schizophrenia and other neurological diseases.

KW - Data argumentation

KW - Deep learning

KW - fMRI

KW - ICA

KW - Model order

KW - Schizophrenia

UR - http://www.scopus.com/inward/record.url?scp=85068617538&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068617538&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-22808-8_53

DO - 10.1007/978-3-030-22808-8_53

M3 - Conference contribution

AN - SCOPUS:85068617538

SN - 9783030228071

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 540

EP - 547

BT - Advances in Neural Networks – ISNN 2019 - 16th International Symposium on Neural Networks, ISNN 2019, Proceedings

A2 - Tang, Huajin

A2 - Wang, Zhanshan

A2 - Lu, Huchuan

PB - Springer Verlag

ER -