Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification
Abstract Missing Alzheimer's disease (AD) data is prevalent and poses significant challenges for AD diagnosis. Previous studies have explored various data imputation approaches on AD data, but the systematic evaluation of deep learning algorithms for imputing heterogeneous and comprehensive AD...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-12-01
|
| Series: | Healthcare Technology Letters |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/htl2.12091 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850132598265741312 |
|---|---|
| author | Namitha Thalekkara Haridas Jose M. Sanchez‐Bornot Paula L. McClean KongFatt Wong‐Lin Alzheimer's Disease Neuroimaging Initiative (ADNI) |
| author_facet | Namitha Thalekkara Haridas Jose M. Sanchez‐Bornot Paula L. McClean KongFatt Wong‐Lin Alzheimer's Disease Neuroimaging Initiative (ADNI) |
| author_sort | Namitha Thalekkara Haridas |
| collection | DOAJ |
| description | Abstract Missing Alzheimer's disease (AD) data is prevalent and poses significant challenges for AD diagnosis. Previous studies have explored various data imputation approaches on AD data, but the systematic evaluation of deep learning algorithms for imputing heterogeneous and comprehensive AD data is limited. This study investigates the efficacy of denoising autoencoder‐based imputation of missing key features of heterogeneous data that comprised tau‐PET, MRI, cognitive and functional assessments, genotype, sociodemographic, and medical history. The authors focused on extreme (≥40%) missing at random of key features which depend on AD progression; identified as the history of a mother having AD, APoE ε4 alleles, and clinical dementia rating. Along with features selected using traditional feature selection methods, latent features extracted from the denoising autoencoder are incorporated for subsequent classification. Using random forest classification with 10‐fold cross‐validation, robust AD predictive performance of imputed datasets (accuracy: 79%–85%; precision: 71%–85%) across missingness levels, and high recall values with 40% missingness are found. Further, the feature‐selected dataset using feature selection methods, including autoencoder, demonstrated higher classification score than that of the original complete dataset. These results highlight the effectiveness and robustness of autoencoder in imputing crucial information for reliable AD prediction in AI‐based clinical decision support systems. |
| format | Article |
| id | doaj-art-5e4e2b74107d489cb0ecbfa1117bdec3 |
| institution | OA Journals |
| issn | 2053-3713 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Wiley |
| record_format | Article |
| series | Healthcare Technology Letters |
| spelling | doaj-art-5e4e2b74107d489cb0ecbfa1117bdec32025-08-20T02:32:11ZengWileyHealthcare Technology Letters2053-37132024-12-0111645246010.1049/htl2.12091Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classificationNamitha Thalekkara Haridas0Jose M. Sanchez‐Bornot1Paula L. McClean2KongFatt Wong‐Lin3Alzheimer's Disease Neuroimaging Initiative (ADNI)Intelligent Systems Research Centre, School of Computing, Engineering and Intelligent Systems Ulster University, Magee campus Derry∼Londonderry Northern Ireland UKIntelligent Systems Research Centre, School of Computing, Engineering and Intelligent Systems Ulster University, Magee campus Derry∼Londonderry Northern Ireland UKPersonalised Medicine Centre, School of Medicine Ulster University, Magee campus Derry∼Londonderry Northern Ireland UKIntelligent Systems Research Centre, School of Computing, Engineering and Intelligent Systems Ulster University, Magee campus Derry∼Londonderry Northern Ireland UKAbstract Missing Alzheimer's disease (AD) data is prevalent and poses significant challenges for AD diagnosis. Previous studies have explored various data imputation approaches on AD data, but the systematic evaluation of deep learning algorithms for imputing heterogeneous and comprehensive AD data is limited. This study investigates the efficacy of denoising autoencoder‐based imputation of missing key features of heterogeneous data that comprised tau‐PET, MRI, cognitive and functional assessments, genotype, sociodemographic, and medical history. The authors focused on extreme (≥40%) missing at random of key features which depend on AD progression; identified as the history of a mother having AD, APoE ε4 alleles, and clinical dementia rating. Along with features selected using traditional feature selection methods, latent features extracted from the denoising autoencoder are incorporated for subsequent classification. Using random forest classification with 10‐fold cross‐validation, robust AD predictive performance of imputed datasets (accuracy: 79%–85%; precision: 71%–85%) across missingness levels, and high recall values with 40% missingness are found. Further, the feature‐selected dataset using feature selection methods, including autoencoder, demonstrated higher classification score than that of the original complete dataset. These results highlight the effectiveness and robustness of autoencoder in imputing crucial information for reliable AD prediction in AI‐based clinical decision support systems.https://doi.org/10.1049/htl2.12091data miningdata reductiondecision support systemsfeature extractionfeature selectionlearning (artificial intelligence) |
| spellingShingle | Namitha Thalekkara Haridas Jose M. Sanchez‐Bornot Paula L. McClean KongFatt Wong‐Lin Alzheimer's Disease Neuroimaging Initiative (ADNI) Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification Healthcare Technology Letters data mining data reduction decision support systems feature extraction feature selection learning (artificial intelligence) |
| title | Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification |
| title_full | Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification |
| title_fullStr | Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification |
| title_full_unstemmed | Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification |
| title_short | Autoencoder imputation of missing heterogeneous data for Alzheimer's disease classification |
| title_sort | autoencoder imputation of missing heterogeneous data for alzheimer s disease classification |
| topic | data mining data reduction decision support systems feature extraction feature selection learning (artificial intelligence) |
| url | https://doi.org/10.1049/htl2.12091 |
| work_keys_str_mv | AT namithathalekkaraharidas autoencoderimputationofmissingheterogeneousdataforalzheimersdiseaseclassification AT josemsanchezbornot autoencoderimputationofmissingheterogeneousdataforalzheimersdiseaseclassification AT paulalmcclean autoencoderimputationofmissingheterogeneousdataforalzheimersdiseaseclassification AT kongfattwonglin autoencoderimputationofmissingheterogeneousdataforalzheimersdiseaseclassification AT alzheimersdiseaseneuroimaginginitiativeadni autoencoderimputationofmissingheterogeneousdataforalzheimersdiseaseclassification |