Contrastive learning for neural fingerprinting from limited neuroimaging data

IntroductionNeural fingerprinting is a technique used to identify individuals based on their unique brain activity patterns. While deep learning techniques have been demonstrated to outperform traditional correlation-based methods, they often require retraining to accommodate new subjects. Furthermo...

Full description

Saved in:
Bibliographic Details
Main Authors: Nikolas Kampel, Farah Abdellatif, N. Jon Shah, Irene Neuner, Jürgen Dammers
Format: Article
Language:English
Published: Frontiers Media S.A. 2024-11-01
Series:Frontiers in Nuclear Medicine
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fnume.2024.1332747/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846169288351678464
author Nikolas Kampel
Nikolas Kampel
Nikolas Kampel
Farah Abdellatif
Farah Abdellatif
N. Jon Shah
N. Jon Shah
N. Jon Shah
N. Jon Shah
Irene Neuner
Irene Neuner
Irene Neuner
Irene Neuner
Jürgen Dammers
Jürgen Dammers
Jürgen Dammers
author_facet Nikolas Kampel
Nikolas Kampel
Nikolas Kampel
Farah Abdellatif
Farah Abdellatif
N. Jon Shah
N. Jon Shah
N. Jon Shah
N. Jon Shah
Irene Neuner
Irene Neuner
Irene Neuner
Irene Neuner
Jürgen Dammers
Jürgen Dammers
Jürgen Dammers
author_sort Nikolas Kampel
collection DOAJ
description IntroductionNeural fingerprinting is a technique used to identify individuals based on their unique brain activity patterns. While deep learning techniques have been demonstrated to outperform traditional correlation-based methods, they often require retraining to accommodate new subjects. Furthermore, the limited availability of samples in neuroscience research can impede the quick adoption of deep learning methods, presenting a challenge for their broader application in neural fingerprinting.MethodsThis study addresses these challenges by using contrastive learning to eliminate the need for retraining with new subjects and developing a data augmentation methodology to enhance model robustness in limited sample size conditions. We utilized the LEMON dataset, comprising 3 Tesla MRI and resting-state fMRI scans from 138 subjects, to compute functional connectivity as a baseline for fingerprinting performance based on correlation metrics. We adapted a recent deep learning model by incorporating data augmentation with short random temporal segments for training and reformulated the fingerprinting task as a contrastive problem, comparing the efficacy of contrastive triplet loss against conventional cross-entropy loss.ResultsThe results of this study confirm that deep learning methods can significantly improve fingerprinting performance over correlation-based methods, achieving an accuracy of about 98% in identifying a single subject out of 138 subjects utilizing 39 different functional connectivity profiles.DiscussionThe contrastive method showed added value in the “leave subject out” scenario, demonstrating flexibility comparable to correlation-based methods and robustness across different data sizes. These findings suggest that contrastive learning and data augmentation offer a scalable solution for neural fingerprinting, even with limited sample sizes.
format Article
id doaj-art-a51a340b90294f46a41e6baf42fc3bd3
institution Kabale University
issn 2673-8880
language English
publishDate 2024-11-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Nuclear Medicine
spelling doaj-art-a51a340b90294f46a41e6baf42fc3bd32024-11-13T05:15:30ZengFrontiers Media S.A.Frontiers in Nuclear Medicine2673-88802024-11-01410.3389/fnume.2024.13327471332747Contrastive learning for neural fingerprinting from limited neuroimaging dataNikolas Kampel0Nikolas Kampel1Nikolas Kampel2Farah Abdellatif3Farah Abdellatif4N. Jon Shah5N. Jon Shah6N. Jon Shah7N. Jon Shah8Irene Neuner9Irene Neuner10Irene Neuner11Irene Neuner12Jürgen Dammers13Jürgen Dammers14Jürgen Dammers15Institute of Neuroscience and Medicine (INM-4), Forschungszentrum Jülich GmbH, Jülich, GermanyFaculty of Medicine, RWTH Aachen University, Aachen, GermanyJülich Aachen Research Alliance (JARA) – CSD – Center for Simulation and Data Science, Aachen, GermanyInstitute of Neuroscience and Medicine (INM-4), Forschungszentrum Jülich GmbH, Jülich, GermanyFaculty of Mathematics, Computer Science and Natural Sciences, RWTH Aachen University, Aachen, GermanyInstitute of Neuroscience and Medicine (INM-4), Forschungszentrum Jülich GmbH, Jülich, GermanyJülich Aachen Research Alliance (JARA) – BRAIN – Translational Medicine, Aachen, GermanyInstitute of Neuroscience and Medicine (INM-11), Jülich Aachen Research Alliance (JARA), Forschungszentrum Jülich GmbH, Jülich, GermanyDepartment of Neurology, University Hospital RWTH Aachen, Aachen, GermanyInstitute of Neuroscience and Medicine (INM-4), Forschungszentrum Jülich GmbH, Jülich, GermanyJülich Aachen Research Alliance (JARA) – CSD – Center for Simulation and Data Science, Aachen, GermanyJülich Aachen Research Alliance (JARA) – BRAIN – Translational Medicine, Aachen, GermanyDepartment of Psychiatry, Psychotherapy and Psychosomatics, RWTH Aachen University, Aachen, GermanyInstitute of Neuroscience and Medicine (INM-4), Forschungszentrum Jülich GmbH, Jülich, GermanyFaculty of Medicine, RWTH Aachen University, Aachen, GermanyJülich Aachen Research Alliance (JARA) – CSD – Center for Simulation and Data Science, Aachen, GermanyIntroductionNeural fingerprinting is a technique used to identify individuals based on their unique brain activity patterns. While deep learning techniques have been demonstrated to outperform traditional correlation-based methods, they often require retraining to accommodate new subjects. Furthermore, the limited availability of samples in neuroscience research can impede the quick adoption of deep learning methods, presenting a challenge for their broader application in neural fingerprinting.MethodsThis study addresses these challenges by using contrastive learning to eliminate the need for retraining with new subjects and developing a data augmentation methodology to enhance model robustness in limited sample size conditions. We utilized the LEMON dataset, comprising 3 Tesla MRI and resting-state fMRI scans from 138 subjects, to compute functional connectivity as a baseline for fingerprinting performance based on correlation metrics. We adapted a recent deep learning model by incorporating data augmentation with short random temporal segments for training and reformulated the fingerprinting task as a contrastive problem, comparing the efficacy of contrastive triplet loss against conventional cross-entropy loss.ResultsThe results of this study confirm that deep learning methods can significantly improve fingerprinting performance over correlation-based methods, achieving an accuracy of about 98% in identifying a single subject out of 138 subjects utilizing 39 different functional connectivity profiles.DiscussionThe contrastive method showed added value in the “leave subject out” scenario, demonstrating flexibility comparable to correlation-based methods and robustness across different data sizes. These findings suggest that contrastive learning and data augmentation offer a scalable solution for neural fingerprinting, even with limited sample sizes.https://www.frontiersin.org/articles/10.3389/fnume.2024.1332747/fulldeep learningcontrastive learningneural fingerprintingfMRIresting-statefunctional connectivity
spellingShingle Nikolas Kampel
Nikolas Kampel
Nikolas Kampel
Farah Abdellatif
Farah Abdellatif
N. Jon Shah
N. Jon Shah
N. Jon Shah
N. Jon Shah
Irene Neuner
Irene Neuner
Irene Neuner
Irene Neuner
Jürgen Dammers
Jürgen Dammers
Jürgen Dammers
Contrastive learning for neural fingerprinting from limited neuroimaging data
Frontiers in Nuclear Medicine
deep learning
contrastive learning
neural fingerprinting
fMRI
resting-state
functional connectivity
title Contrastive learning for neural fingerprinting from limited neuroimaging data
title_full Contrastive learning for neural fingerprinting from limited neuroimaging data
title_fullStr Contrastive learning for neural fingerprinting from limited neuroimaging data
title_full_unstemmed Contrastive learning for neural fingerprinting from limited neuroimaging data
title_short Contrastive learning for neural fingerprinting from limited neuroimaging data
title_sort contrastive learning for neural fingerprinting from limited neuroimaging data
topic deep learning
contrastive learning
neural fingerprinting
fMRI
resting-state
functional connectivity
url https://www.frontiersin.org/articles/10.3389/fnume.2024.1332747/full
work_keys_str_mv AT nikolaskampel contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT nikolaskampel contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT nikolaskampel contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT farahabdellatif contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT farahabdellatif contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT njonshah contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT njonshah contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT njonshah contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT njonshah contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT ireneneuner contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT ireneneuner contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT ireneneuner contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT ireneneuner contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT jurgendammers contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT jurgendammers contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata
AT jurgendammers contrastivelearningforneuralfingerprintingfromlimitedneuroimagingdata