Adopting transfer learning for neuroimaging: a comparative analysis with a custom 3D convolution neural network model.
JSON
Soliman A, Chang JR, Etminani K, Byttner S, Davidsson A, Martínez-Sanchis B, Camacho V, Bauckneht M, Stegeran R, Ressner M, Agudelo-Cifuentes M, Chincarini A, Brendel M, Rominger A, Bruffaerts R, Vandenberghe R, Kramberger MG, Trost M, Nicastro N, Frisoni GB, Lemstra AW, Berckel BNMV, Pilotto A, Padovani A, Morbelli S, Aarsland D, Nobili F, Garibotto V, Alzheimer’s Disease Neuroimaging Initiative , Ochoa-Figueroa M
BMC Med Inform Decis Mak 22 (Suppl 6) 318 [2022-12-07; online 2022-12-07]
In recent years, neuroimaging with deep learning (DL) algorithms have made remarkable advances in the diagnosis of neurodegenerative disorders. However, applying DL in different medical domains is usually challenged by lack of labeled data. To address this challenge, transfer learning (TL) has been applied to use state-of-the-art convolution neural networks pre-trained on natural images. Yet, there are differences in characteristics between medical and natural images, also image classification and targeted medical diagnosis tasks. The purpose of this study is to investigate the performance of specialized and TL in the classification of neurodegenerative disorders using 3D volumes of 18F-FDG-PET brain scans. Results show that TL models are suboptimal for classification of neurodegenerative disorders, especially when the objective is to separate more than two disorders. Additionally, specialized CNN model provides better interpretations of predicted diagnosis. TL can indeed lead to superior performance on binary classification in timely and data efficient manner, yet for detecting more than a single disorder, TL models do not perform well. Additionally, custom 3D model performs comparably to TL models for binary classification, and interestingly perform better for diagnosis of multiple disorders. The results confirm the superiority of the custom 3D-CNN in providing better explainable model compared to TL adopted ones.
PubMed 36476613
DOI 10.1186/s12911-022-02054-7
Crossref 10.1186/s12911-022-02054-7
pmc: PMC9727842
pii: 10.1186/s12911-022-02054-7