Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression
Abstract Introduction Genome‐wide association studies (GWAS) for late onset Alzheimer's disease (AD) may miss genetic variants relevant for delineating disease stages when using clinically defined case/control as a phenotype due to its loose definition and heterogeneity. Methods We use a transf...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2021-01-01
|
| Series: | Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/dad2.12140 |
| _version_ | 1797975502444560384 |
|---|---|
| author | Yi Li Annat Haber Christoph Preuss Cai John Asli Uyar Hongtian Stanley Yang Benjamin A. Logsdon Vivek Philip R. Krishna Murthy Karuturi Gregory W. Carter The Alzheimer's Disease Neuroimaging Initiative |
| author_facet | Yi Li Annat Haber Christoph Preuss Cai John Asli Uyar Hongtian Stanley Yang Benjamin A. Logsdon Vivek Philip R. Krishna Murthy Karuturi Gregory W. Carter The Alzheimer's Disease Neuroimaging Initiative |
| author_sort | Yi Li |
| collection | DOAJ |
| description | Abstract Introduction Genome‐wide association studies (GWAS) for late onset Alzheimer's disease (AD) may miss genetic variants relevant for delineating disease stages when using clinically defined case/control as a phenotype due to its loose definition and heterogeneity. Methods We use a transfer learning technique to train three‐dimensional convolutional neural network (CNN) models based on structural magnetic resonance imaging (MRI) from the screening stage in the Alzheimer's Disease Neuroimaging Initiative consortium to derive image features that reflect AD progression. Results CNN‐derived image phenotypes are significantly associated with fasting metabolites related to early lipid metabolic changes as well as insulin resistance and with genetic variants mapped to candidate genes enriched for amyloid beta degradation, tau phosphorylation, calcium ion binding‐dependent synaptic loss, APP‐regulated inflammation response, and insulin resistance. Discussion This is the first attempt to show that non‐invasive MRI biomarkers are linked to AD progression characteristics, reinforcing their use in early AD diagnosis and monitoring. |
| first_indexed | 2024-04-11T04:36:34Z |
| format | Article |
| id | doaj.art-203af5651b5e44f899b9f095b9b2508e |
| institution | Directory Open Access Journal |
| issn | 2352-8729 |
| language | English |
| last_indexed | 2024-04-11T04:36:34Z |
| publishDate | 2021-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring |
| spelling | doaj.art-203af5651b5e44f899b9f095b9b2508e2022-12-28T09:12:13ZengWileyAlzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring2352-87292021-01-01131n/an/a10.1002/dad2.12140Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progressionYi Li0Annat Haber1Christoph Preuss2Cai John3Asli Uyar4Hongtian Stanley Yang5Benjamin A. Logsdon6Vivek Philip7R. Krishna Murthy Karuturi8Gregory W. Carter9The Alzheimer's Disease Neuroimaging InitiativeThe Jackson Laboratory Farmington Connecticut USAThe Jackson Laboratory Farmington Connecticut USAThe Jackson Laboratory Bar Harbor Maine USAThe Jackson Laboratory Farmington Connecticut USAThe Jackson Laboratory Farmington Connecticut USAThe Jackson Laboratory Bar Harbor Maine USASage Bionetworks Seattle Washington USAThe Jackson Laboratory Bar Harbor Maine USAThe Jackson Laboratory Farmington Connecticut USAThe Jackson Laboratory Farmington Connecticut USAAbstract Introduction Genome‐wide association studies (GWAS) for late onset Alzheimer's disease (AD) may miss genetic variants relevant for delineating disease stages when using clinically defined case/control as a phenotype due to its loose definition and heterogeneity. Methods We use a transfer learning technique to train three‐dimensional convolutional neural network (CNN) models based on structural magnetic resonance imaging (MRI) from the screening stage in the Alzheimer's Disease Neuroimaging Initiative consortium to derive image features that reflect AD progression. Results CNN‐derived image phenotypes are significantly associated with fasting metabolites related to early lipid metabolic changes as well as insulin resistance and with genetic variants mapped to candidate genes enriched for amyloid beta degradation, tau phosphorylation, calcium ion binding‐dependent synaptic loss, APP‐regulated inflammation response, and insulin resistance. Discussion This is the first attempt to show that non‐invasive MRI biomarkers are linked to AD progression characteristics, reinforcing their use in early AD diagnosis and monitoring.https://doi.org/10.1002/dad2.12140Alzheimer's diseaseconvolutional neural networksdeep learningdisease progressionimaging phenotypesmachine learning |
| spellingShingle | Yi Li Annat Haber Christoph Preuss Cai John Asli Uyar Hongtian Stanley Yang Benjamin A. Logsdon Vivek Philip R. Krishna Murthy Karuturi Gregory W. Carter The Alzheimer's Disease Neuroimaging Initiative Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring Alzheimer's disease convolutional neural networks deep learning disease progression imaging phenotypes machine learning |
| title | Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression |
| title_full | Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression |
| title_fullStr | Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression |
| title_full_unstemmed | Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression |
| title_short | Transfer learning‐trained convolutional neural networks identify novel MRI biomarkers of Alzheimer's disease progression |
| title_sort | transfer learning trained convolutional neural networks identify novel mri biomarkers of alzheimer s disease progression |
| topic | Alzheimer's disease convolutional neural networks deep learning disease progression imaging phenotypes machine learning |
| url | https://doi.org/10.1002/dad2.12140 |
| work_keys_str_mv | AT yili transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT annathaber transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT christophpreuss transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT caijohn transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT asliuyar transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT hongtianstanleyyang transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT benjaminalogsdon transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT vivekphilip transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT rkrishnamurthykaruturi transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT gregorywcarter transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression AT thealzheimersdiseaseneuroimaginginitiative transferlearningtrainedconvolutionalneuralnetworksidentifynovelmribiomarkersofalzheimersdiseaseprogression |