Attention mechanism‐based prediction of early tau accumulation using MRI
Background Assessing tau accumulation in early affected areas like the lateral entorhinal cortex (EC) and inferior temporal gyrus (ITG) enables early prediction of disease progression and cognitive decline. However, positron emission tomography (PET) imaging poses radiation exposure and cost concern...
Saved in:
| Published in: | Alzheimer's & dementia Vol. 20; no. S1 |
|---|---|
| Main Authors: | , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Hoboken
John Wiley and Sons Inc
01.12.2024
|
| Subjects: | |
| ISSN: | 1552-5260, 1552-5279 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Background
Assessing tau accumulation in early affected areas like the lateral entorhinal cortex (EC) and inferior temporal gyrus (ITG) enables early prediction of disease progression and cognitive decline. However, positron emission tomography (PET) imaging poses radiation exposure and cost concerns. This research aims to develop a deep learning model predicting tau positivity in these regions using MRI.
Method
In this study, we used the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort, of which dataset was partitioned into train, validation, and test sets (8:1:1 ratio), encompassing a total of 1010 scans, all of whom underwent T1‐weighted magnetic resonance imaging (MRI) and [18F] flortaucipir‐PET imaging. For the T1‐weighted MRI images, FreeSurfer v7.2 was employed to perform pre‐processing and extract cortical thickness measurements. Simultaneously, [18F] flortaucipir‐PET imaging was processed to compute voxel‐wise regions of interest (ROIs) for 66 specific brain regions. Regional tau positivity was established using a cutoff at a z‐score of 1.25, with a focus on cognitive normal (CN) subjects within the train set. To predict early tau accumulation regions, we developed an attention mechanism‐based encoder‐decoder model by adopting a Transformer model into our problem setting, performing sequential predictions for each of the 66 regions. Notably, the model’s predictive performance in initial regions significantly influences subsequent predictions. Consequently, we implemented a prioritization strategy, emphasizing predictions from areas where the model demonstrated high accuracy. This approach was designed to enhance the overall predictive accuracy of the model.
Result
Predicting five early tau accumulation regions per hemisphere, our model achieved an average AUC of 0.84 and accuracy of 84% for the test dataset (112 participants). Notably, in critical early disease progression regions (fusiform gyrus and ITG), AUC values of 0.84, 0.85, and accuracies of 84.4%, 84% were observed. Furthermore, the proposed prioritization strategy improved performance compared to predictions using vanilla attention‐based model.
Conclusion
We developed an attention mechanism‐based architecture with an encoder‐decoder structure. By predicting outcomes not only based on cortical thickness values but also their cross‐attention‐based contexture information, we could achieve highly accurate tau prediction in early and challenging regions |
|---|---|
| AbstractList | Background
Assessing tau accumulation in early affected areas like the lateral entorhinal cortex (EC) and inferior temporal gyrus (ITG) enables early prediction of disease progression and cognitive decline. However, positron emission tomography (PET) imaging poses radiation exposure and cost concerns. This research aims to develop a deep learning model predicting tau positivity in these regions using MRI.
Method
In this study, we used the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort, of which dataset was partitioned into train, validation, and test sets (8:1:1 ratio), encompassing a total of 1010 scans, all of whom underwent T1‐weighted magnetic resonance imaging (MRI) and [18F] flortaucipir‐PET imaging. For the T1‐weighted MRI images, FreeSurfer v7.2 was employed to perform pre‐processing and extract cortical thickness measurements. Simultaneously, [18F] flortaucipir‐PET imaging was processed to compute voxel‐wise regions of interest (ROIs) for 66 specific brain regions. Regional tau positivity was established using a cutoff at a z‐score of 1.25, with a focus on cognitive normal (CN) subjects within the train set. To predict early tau accumulation regions, we developed an attention mechanism‐based encoder‐decoder model by adopting a Transformer model into our problem setting, performing sequential predictions for each of the 66 regions. Notably, the model’s predictive performance in initial regions significantly influences subsequent predictions. Consequently, we implemented a prioritization strategy, emphasizing predictions from areas where the model demonstrated high accuracy. This approach was designed to enhance the overall predictive accuracy of the model.
Result
Predicting five early tau accumulation regions per hemisphere, our model achieved an average AUC of 0.84 and accuracy of 84% for the test dataset (112 participants). Notably, in critical early disease progression regions (fusiform gyrus and ITG), AUC values of 0.84, 0.85, and accuracies of 84.4%, 84% were observed. Furthermore, the proposed prioritization strategy improved performance compared to predictions using vanilla attention‐based model.
Conclusion
We developed an attention mechanism‐based architecture with an encoder‐decoder structure. By predicting outcomes not only based on cortical thickness values but also their cross‐attention‐based contexture information, we could achieve highly accurate tau prediction in early and challenging regions |
| Author | Lee, Wha Jin Seong, Joon‐Kyung Kim, Jin‐Yang Song, Yeong‐Hun |
| AuthorAffiliation | 3 Alzheimer’s Disease Neuroimaging Initiative, http://adni.loni.usc.edu/, CA USA 1 Korea University, Seoul Korea, Republic of (South) 2 NeuroXT, Seoul Korea, Republic of (South) |
| AuthorAffiliation_xml | – name: 1 Korea University, Seoul Korea, Republic of (South) – name: 2 NeuroXT, Seoul Korea, Republic of (South) – name: 3 Alzheimer’s Disease Neuroimaging Initiative, http://adni.loni.usc.edu/, CA USA |
| Author_xml | – sequence: 1 givenname: Jin‐Yang surname: Kim fullname: Kim, Jin‐Yang email: a9613789@naver.com organization: Korea University, Seoul – sequence: 2 givenname: Yeong‐Hun surname: Song fullname: Song, Yeong‐Hun organization: Korea University, Seoul – sequence: 3 givenname: Wha Jin surname: Lee fullname: Lee, Wha Jin organization: NeuroXT, Seoul – sequence: 4 givenname: Joon‐Kyung surname: Seong fullname: Seong, Joon‐Kyung organization: NeuroXT, Seoul |
| BookMark | eNp9kM1Kw0AUhQepYK1ufIKshdR7p5NJspJS_KlUBNGNm2EyuWlH8lMyiVJXPoLP6JMYm1Jw4-oeON85XM4xG5RVSYydIYwRgF_o_GMMMUcBB2yIQcD9gIfxYK8lHLFj514BBEQYDNndtGmobGxVegWZlS6tK74_vxLtKPXWNaXWbM0q80jX-cZrdOtpY9qizfXWaZ0tl9794_yEHWY6d3S6uyP2fH31NLv1Fw8389l04RuUMfg6mQiDJoIECDlkKIJECpPImBtIKcoET0OJqOMoCyOudQBJBCQFZmQojiYjdtn3rtukoNR079c6V-vaFrreqEpb9dcp7UotqzeFGCIIKbuG877B1JVzNWX7MIL63VF1O6p-xw7GHn63OW3-IdV08bLL_ADzQXmn |
| ContentType | Journal Article |
| Copyright | 2024 The Alzheimer's Association. published by Wiley Periodicals LLC on behalf of Alzheimer's Association. |
| Copyright_xml | – notice: 2024 The Alzheimer's Association. published by Wiley Periodicals LLC on behalf of Alzheimer's Association. |
| CorporateAuthor | ADNI |
| CorporateAuthor_xml | – name: ADNI |
| DBID | 24P AAYXX CITATION 5PM |
| DOI | 10.1002/alz.092140 |
| DatabaseName | Wiley Online Library Open Access CrossRef PubMed Central (Full Participant titles) |
| DatabaseTitle | CrossRef |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: 24P name: Wiley Online Library Open Access url: https://authorservices.wiley.com/open-science/open-access/browse-journals.html sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| DocumentTitleAlternate | BASIC SCIENCE AND PATHOGENESIS |
| EISSN | 1552-5279 |
| EndPage | n/a |
| ExternalDocumentID | PMC11710466 10_1002_alz_092140 ALZ092140 |
| Genre | abstract |
| GroupedDBID | --- --K --M .~1 0R~ 1B1 1OC 1~. 1~5 24P 33P 4.4 457 4G. 53G 5VS 7-5 71M 7RV 7X7 8FI 8FJ 8P~ AAEDT AAIKJ AAKOC AALRI AAMMB AANLZ AAOAW AAXLA AAXUO AAYCA AAYWO ABBQC ABCQJ ABCUV ABIVO ABJNI ABMAC ABMZM ABUWG ABWVN ACCMX ACCZN ACGFS ACGOF ACPOU ACRPL ACVFH ACXQS ADBBV ADBTR ADCNI ADEZE ADHUB ADKYN ADMUD ADNMO ADPDF ADVLN ADZMN AEFGJ AEIGN AEKER AENEX AEUPX AEUYR AEVXI AFKRA AFPUW AFTJW AFWVQ AGHFR AGHNM AGUBO AGWIK AGXDD AGYEJ AIDQK AIDYY AIGII AITUG AIURR AJRQY AKBMS AKRWK AKYEP ALMA_UNASSIGNED_HOLDINGS ALUQN AMRAJ AMYDB ANZVX AZQEC BENPR BFHJK BLXMC C45 CCPQU DCZOG EBS EJD EMOBN EO8 EO9 EP2 EP3 F5P FDB FEDTE FIRID FNPLU FYUFA G-Q GBLVA HMCUK HVGLF HX~ HZ~ IHE J1W K9- LATKE LEEKS M0R M41 MO0 MOBAO N9A NAPCQ O-L O9- OAUVE OVD OVEED OZT P-8 P-9 P2P PC. PGMZT PHGZM PHGZT PIMPY PSYQQ Q38 QTD RIG ROL RPM RPZ SDF SDG SEL SES SSZ SUPJJ TEORI UKHRP ~G- 9DU AAYXX AFFHD CITATION EFLBG PJZUB PPXIY ~HD 5PM |
| ID | FETCH-LOGICAL-c1690-ab34c1c80b0e120f145b64cb692c0de8f42d7611a98f782aa50b80e641fece983 |
| IEDL.DBID | 24P |
| ISSN | 1552-5260 |
| IngestDate | Tue Nov 04 02:04:47 EST 2025 Sat Nov 29 07:19:14 EST 2025 Sun Jul 06 04:45:08 EDT 2025 |
| IsDoiOpenAccess | true |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | S1 |
| Language | English |
| License | Attribution This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c1690-ab34c1c80b0e120f145b64cb692c0de8f42d7611a98f782aa50b80e641fece983 |
| OpenAccessLink | https://onlinelibrary.wiley.com/doi/abs/10.1002%2Falz.092140 |
| PageCount | 2 |
| ParticipantIDs | pubmedcentral_primary_oai_pubmedcentral_nih_gov_11710466 crossref_primary_10_1002_alz_092140 wiley_primary_10_1002_alz_092140_ALZ092140 |
| PublicationCentury | 2000 |
| PublicationDate | December 2024 |
| PublicationDateYYYYMMDD | 2024-12-01 |
| PublicationDate_xml | – month: 12 year: 2024 text: December 2024 |
| PublicationDecade | 2020 |
| PublicationPlace | Hoboken |
| PublicationPlace_xml | – name: Hoboken |
| PublicationTitle | Alzheimer's & dementia |
| PublicationYear | 2024 |
| Publisher | John Wiley and Sons Inc |
| Publisher_xml | – name: John Wiley and Sons Inc |
| SSID | ssj0040815 |
| Score | 2.4104605 |
| Snippet | Background
Assessing tau accumulation in early affected areas like the lateral entorhinal cortex (EC) and inferior temporal gyrus (ITG) enables early... |
| SourceID | pubmedcentral crossref wiley |
| SourceType | Open Access Repository Index Database Publisher |
| SubjectTerms | Basic Science and Pathogenesis |
| Title | Attention mechanism‐based prediction of early tau accumulation using MRI |
| URI | https://onlinelibrary.wiley.com/doi/abs/10.1002%2Falz.092140 https://pubmed.ncbi.nlm.nih.gov/PMC11710466 |
| Volume | 20 |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVPQU databaseName: Family Health Database (Proquest) customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: M0R dateStart: 20240101 isFulltext: true titleUrlDefault: https://search.proquest.com/familyhealth providerName: ProQuest – providerCode: PRVPQU databaseName: Health & Medical Collection customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: 7X7 dateStart: 20240101 isFulltext: true titleUrlDefault: https://search.proquest.com/healthcomplete providerName: ProQuest – providerCode: PRVPQU databaseName: Nursing & Allied Health Database customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: 7RV dateStart: 20240101 isFulltext: true titleUrlDefault: https://search.proquest.com/nahs providerName: ProQuest – providerCode: PRVPQU databaseName: ProQuest Central customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: BENPR dateStart: 20240101 isFulltext: true titleUrlDefault: https://www.proquest.com/central providerName: ProQuest – providerCode: PRVPQU databaseName: ProQuest Publicly Available Content Database customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: PIMPY dateStart: 20240101 isFulltext: true titleUrlDefault: http://search.proquest.com/publiccontent providerName: ProQuest – providerCode: PRVWIB databaseName: Wiley Online Library Open Access customDbUrl: eissn: 1552-5279 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0040815 issn: 1552-5260 databaseCode: 24P dateStart: 20240101 isFulltext: true titleUrlDefault: https://authorservices.wiley.com/open-science/open-access/browse-journals.html providerName: Wiley-Blackwell |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1JSwMxFA5aPXhxQcW6EdCTMDbJJJkMeCliUbGlFIXiZchkMlqwC-3Ugyd_gr_RX2KWLtaDIF6GgWQWXt6a9_I9AE4lCjOUYxbQKA8Dqo0oitzwMpYx5YJJRrRvNhE1GqLdjptL4GJ6FsbjQ8w23KxkOH1tBVymo8ocNFS-vJ2jmJgAYRmsYBwK27iB0OZUD1Nj7JhDS2U23OJoBk5KKvNnF8zRz7LI7-6qsze1jf_96SZYn_iZsOoZYwss6d42uK0WhS9vhF1tj_x2Rt3P9w9ryTI4GNqcjRvs51Bb4GNYyDGUSo27kyZf0JbJP8F662YHPNSu7i-vg0kzhUDZTFgg05AqrARKkcbELA9lKacq5TFRKNMipySLODZrJHLjNUjJUCqQ5hTnWulYhLug1Ov39B6AWnGjJwRVhJtghDCZuQSdlBHJsijnZXAypWky8JgZiUdHJomhReJpUQZigdyzqRbwenGk13l2wNcYRzYlbT5w5uj9y9uT6t2jv9v_y-QDsEaM1-LrVQ5BqRiO9RFYVa9FZzQ8dvxlrnXU-gIVt9Uk |
| linkProvider | Wiley-Blackwell |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV3NS8MwFA86Bb34gYrzM6AnoS5J0zQ9DnFsuo0hE4aXkqapDlw3ts6DJ_8E_0b_EpOmbs6DIN4KSdry8j7zXn4PgHOB3Bgl2HOon7gOVVoUeaJ5GYuAMu4JjyjbbMJvt3mvF3SK2hxzF8biQ8wO3Ixk5PraCLg5kK7MUUPF8-slCoiOEJbBCtVmxrA5oZ0vRUy1tfNyuFTPxFsMzdBJSWW-dsEe_ayL_O6v5gantvnPX90CG4WnCauWNbbBkkp3wE01y2yBIxwoc-m3Pxl8vL0bWxbD0dhkbfLBYQKVgT6GmZhCIeV0ULT5gqZQ_hG27hq74L523b2qO0U7BUeaXJgjIpdKLDmKkMJEbxD1IkZlxAIiUax4QknsM6x3iSfabxDCQxFHilGcKKkC7u6BUjpM1T6ASjKtKTiVhOlwhHgizlN0Qvgkjv2ElcHZF1HDkUXNCC0-Mgk1LUJLizLgC_SeTTWQ14sjaf8ph77G2DdJaf2Bi5zgv7w9rDYf7NPBXyafgrV6t9UMm4327SFYJ9qHsdUrR6CUjafqGKzKl6w_GZ_kzPYJyPnYHw |
| linkToPdf | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1JSwMxFA5aRby4oGJdA3oSRpM0yWSORS1WaymiULwMmUyiBbvQTj148if4G_0lJpMu1oMg3gaSSYaXt857-R4AxxKVUmQwC2hoSgHVVhSFsbyMZUS5YJIR7ZtNhPW6aDajxqg2x92F8fgQkx9uTjJyfe0EXPdSczZFDZUvb6coIjZCmAcLlIXYMTWhjbEiptbasRwulbl4i6MJOik5m747Y49-1kV-91dzg1NZ_eenroGVkacJy5411sGc7myA63KW-QJH2Nbu0m9r0P58_3C2LIW9vsva5INdA7WDPoaZHEKp1LA9avMFXaH8E7y9q26Ch8rl_flVMGqnECiXCwtkUqIKK4ESpDGxB0RZwqlKeEQUSrUwlKQhx_aUhLF-g5QMJQJpTrHRSkeitAUKnW5HbwOoFbeaQlBFuA1HCJNpnqKTMiRpGhpeBEdjosY9j5oRe3xkEltaxJ4WRSBm6D2Z6iCvZ0c6recc-hrj0CWl7QYnOcF_WT0u1x79085fJh-CpcZFJa5V6ze7YJlYF8YXr-yBQtYf6n2wqF6z1qB_kPPaF6W116M |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Attention+mechanism%E2%80%90based+prediction+of+early+tau+accumulation+using+MRI&rft.jtitle=Alzheimer%27s+%26+dementia&rft.au=Kim%2C+Jin%E2%80%90Yang&rft.au=Song%2C+Yeong%E2%80%90Hun&rft.au=Lee%2C+Wha+Jin&rft.au=Seong%2C+Joon%E2%80%90Kyung&rft.date=2024-12-01&rft.pub=John+Wiley+and+Sons+Inc&rft.issn=1552-5260&rft.eissn=1552-5279&rft.volume=20&rft.issue=Suppl+1&rft_id=info:doi/10.1002%2Falz.092140&rft.externalDocID=PMC11710466 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1552-5260&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1552-5260&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1552-5260&client=summon |