Cognitive signature of brain FDG PET based on deep learning: domain transfer from Alzheimer’s disease to Parkinson’s disease

for the Alzheimer’s Disease Neuroimaging Initiative

Research output: Contribution to journalArticle

Abstract

Purpose: Although functional brain imaging has been used for the early and objective assessment of cognitive dysfunction, there is a lack of generalized image-based biomarker which can evaluate individual’s cognitive dysfunction in various disorders. To this end, we developed a deep learning-based cognitive signature of FDG brain PET adaptable for Parkinson’s disease (PD) as well as Alzheimer’s disease (AD). Methods: A deep learning model for discriminating AD from normal controls (NCs) was built by a training set consisting of 636 FDG PET obtained from Alzheimer’s Disease Neuroimaging Initiative database. The model was directly transferred to images of mild cognitive impairment (MCI) patients (n = 666) for identifying who would rapidly convert to AD and another independent cohort consisting of 62 PD patients to differentiate PD patients with dementia. The model accuracy was measured by area under curve (AUC) of receiver operating characteristic (ROC) analysis. The relationship between all images was visualized by two-dimensional projection of the deep learning-based features. The model was also designed to predict cognitive score of the subjects and validated in PD patients. Cognitive dysfunction-related regions were visualized by feature maps of the deep CNN model. Results: AUC of ROC for differentiating AD from NC was 0.94 (95% CI 0.89–0.98). The transfer of the model could differentiate MCI patients who would convert to AD (AUC = 0.82) and PD with dementia (AUC = 0.81). The two-dimensional projection mapping visualized the degree of cognitive dysfunction compared with normal brains regardless of different disease cohorts. Predicted cognitive score, an output of the model, was highly correlated with the mini-mental status exam scores. Individual cognitive dysfunction-related regions included cingulate and high frontoparietal cortices, while they showed individual variability. Conclusion: The deep learning-based cognitive function evaluation model could be successfully transferred to multiple disease domains. We suggest that this approach might be extended to an objective cognitive signature that provides quantitative biomarker for cognitive dysfunction across various neurodegenerative disorders.

Original languageEnglish
Pages (from-to)403-412
Number of pages10
JournalEuropean Journal of Nuclear Medicine and Molecular Imaging
Volume47
Issue number2
DOIs
StatePublished - 1 Feb 2020

Fingerprint

Parkinson Disease
Alzheimer Disease
Brain
Area Under Curve
Learning
ROC Curve
Dementia
Biomarkers
Functional Neuroimaging
Transfer (Psychology)
Cognitive Dysfunction
Gyrus Cinguli
Neuroimaging
Neurodegenerative Diseases
Cognition
Databases

Keywords

  • Deep learning
  • Dementia
  • FDG PET
  • Parkinson disease
  • Transfer learning

Cite this

@article{649aba3adf6c4cd6928e94b26cca0eb5,
title = "Cognitive signature of brain FDG PET based on deep learning: domain transfer from Alzheimer’s disease to Parkinson’s disease",
abstract = "Purpose: Although functional brain imaging has been used for the early and objective assessment of cognitive dysfunction, there is a lack of generalized image-based biomarker which can evaluate individual’s cognitive dysfunction in various disorders. To this end, we developed a deep learning-based cognitive signature of FDG brain PET adaptable for Parkinson’s disease (PD) as well as Alzheimer’s disease (AD). Methods: A deep learning model for discriminating AD from normal controls (NCs) was built by a training set consisting of 636 FDG PET obtained from Alzheimer’s Disease Neuroimaging Initiative database. The model was directly transferred to images of mild cognitive impairment (MCI) patients (n = 666) for identifying who would rapidly convert to AD and another independent cohort consisting of 62 PD patients to differentiate PD patients with dementia. The model accuracy was measured by area under curve (AUC) of receiver operating characteristic (ROC) analysis. The relationship between all images was visualized by two-dimensional projection of the deep learning-based features. The model was also designed to predict cognitive score of the subjects and validated in PD patients. Cognitive dysfunction-related regions were visualized by feature maps of the deep CNN model. Results: AUC of ROC for differentiating AD from NC was 0.94 (95{\%} CI 0.89–0.98). The transfer of the model could differentiate MCI patients who would convert to AD (AUC = 0.82) and PD with dementia (AUC = 0.81). The two-dimensional projection mapping visualized the degree of cognitive dysfunction compared with normal brains regardless of different disease cohorts. Predicted cognitive score, an output of the model, was highly correlated with the mini-mental status exam scores. Individual cognitive dysfunction-related regions included cingulate and high frontoparietal cortices, while they showed individual variability. Conclusion: The deep learning-based cognitive function evaluation model could be successfully transferred to multiple disease domains. We suggest that this approach might be extended to an objective cognitive signature that provides quantitative biomarker for cognitive dysfunction across various neurodegenerative disorders.",
keywords = "Deep learning, Dementia, FDG PET, Parkinson disease, Transfer learning",
author = "{for the Alzheimer’s Disease Neuroimaging Initiative} and Hongyoon Choi and Kim, {Yu Kyeong} and Yoon, {Eun Jin} and Lee, {Jee Young} and Lee, {Dong Soo}",
year = "2020",
month = "2",
day = "1",
doi = "10.1007/s00259-019-04538-7",
language = "English",
volume = "47",
pages = "403--412",
journal = "European Journal of Nuclear Medicine and Molecular Imaging",
issn = "1619-7070",
publisher = "Springer Verlag",
number = "2",

}

Cognitive signature of brain FDG PET based on deep learning : domain transfer from Alzheimer’s disease to Parkinson’s disease. / for the Alzheimer’s Disease Neuroimaging Initiative.

In: European Journal of Nuclear Medicine and Molecular Imaging, Vol. 47, No. 2, 01.02.2020, p. 403-412.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Cognitive signature of brain FDG PET based on deep learning

T2 - domain transfer from Alzheimer’s disease to Parkinson’s disease

AU - for the Alzheimer’s Disease Neuroimaging Initiative

AU - Choi, Hongyoon

AU - Kim, Yu Kyeong

AU - Yoon, Eun Jin

AU - Lee, Jee Young

AU - Lee, Dong Soo

PY - 2020/2/1

Y1 - 2020/2/1

N2 - Purpose: Although functional brain imaging has been used for the early and objective assessment of cognitive dysfunction, there is a lack of generalized image-based biomarker which can evaluate individual’s cognitive dysfunction in various disorders. To this end, we developed a deep learning-based cognitive signature of FDG brain PET adaptable for Parkinson’s disease (PD) as well as Alzheimer’s disease (AD). Methods: A deep learning model for discriminating AD from normal controls (NCs) was built by a training set consisting of 636 FDG PET obtained from Alzheimer’s Disease Neuroimaging Initiative database. The model was directly transferred to images of mild cognitive impairment (MCI) patients (n = 666) for identifying who would rapidly convert to AD and another independent cohort consisting of 62 PD patients to differentiate PD patients with dementia. The model accuracy was measured by area under curve (AUC) of receiver operating characteristic (ROC) analysis. The relationship between all images was visualized by two-dimensional projection of the deep learning-based features. The model was also designed to predict cognitive score of the subjects and validated in PD patients. Cognitive dysfunction-related regions were visualized by feature maps of the deep CNN model. Results: AUC of ROC for differentiating AD from NC was 0.94 (95% CI 0.89–0.98). The transfer of the model could differentiate MCI patients who would convert to AD (AUC = 0.82) and PD with dementia (AUC = 0.81). The two-dimensional projection mapping visualized the degree of cognitive dysfunction compared with normal brains regardless of different disease cohorts. Predicted cognitive score, an output of the model, was highly correlated with the mini-mental status exam scores. Individual cognitive dysfunction-related regions included cingulate and high frontoparietal cortices, while they showed individual variability. Conclusion: The deep learning-based cognitive function evaluation model could be successfully transferred to multiple disease domains. We suggest that this approach might be extended to an objective cognitive signature that provides quantitative biomarker for cognitive dysfunction across various neurodegenerative disorders.

AB - Purpose: Although functional brain imaging has been used for the early and objective assessment of cognitive dysfunction, there is a lack of generalized image-based biomarker which can evaluate individual’s cognitive dysfunction in various disorders. To this end, we developed a deep learning-based cognitive signature of FDG brain PET adaptable for Parkinson’s disease (PD) as well as Alzheimer’s disease (AD). Methods: A deep learning model for discriminating AD from normal controls (NCs) was built by a training set consisting of 636 FDG PET obtained from Alzheimer’s Disease Neuroimaging Initiative database. The model was directly transferred to images of mild cognitive impairment (MCI) patients (n = 666) for identifying who would rapidly convert to AD and another independent cohort consisting of 62 PD patients to differentiate PD patients with dementia. The model accuracy was measured by area under curve (AUC) of receiver operating characteristic (ROC) analysis. The relationship between all images was visualized by two-dimensional projection of the deep learning-based features. The model was also designed to predict cognitive score of the subjects and validated in PD patients. Cognitive dysfunction-related regions were visualized by feature maps of the deep CNN model. Results: AUC of ROC for differentiating AD from NC was 0.94 (95% CI 0.89–0.98). The transfer of the model could differentiate MCI patients who would convert to AD (AUC = 0.82) and PD with dementia (AUC = 0.81). The two-dimensional projection mapping visualized the degree of cognitive dysfunction compared with normal brains regardless of different disease cohorts. Predicted cognitive score, an output of the model, was highly correlated with the mini-mental status exam scores. Individual cognitive dysfunction-related regions included cingulate and high frontoparietal cortices, while they showed individual variability. Conclusion: The deep learning-based cognitive function evaluation model could be successfully transferred to multiple disease domains. We suggest that this approach might be extended to an objective cognitive signature that provides quantitative biomarker for cognitive dysfunction across various neurodegenerative disorders.

KW - Deep learning

KW - Dementia

KW - FDG PET

KW - Parkinson disease

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85076225755&partnerID=8YFLogxK

U2 - 10.1007/s00259-019-04538-7

DO - 10.1007/s00259-019-04538-7

M3 - Article

C2 - 31768599

AN - SCOPUS:85076225755

VL - 47

SP - 403

EP - 412

JO - European Journal of Nuclear Medicine and Molecular Imaging

JF - European Journal of Nuclear Medicine and Molecular Imaging

SN - 1619-7070

IS - 2

ER -