Background: Recent advances in data-driven computational approaches have been helpful in devising tools to objectively diagnose psychiatric disorders. However, current machine learning studies... Show moreBackground: Recent advances in data-driven computational approaches have been helpful in devising tools to objectively diagnose psychiatric disorders. However, current machine learning studies limited to small homogeneous samples, different methodologies, and different imaging collection protocols, limit the ability to directly compare and generalize their results. Here we aimed to classify individuals with PTSD versus controls and assess the generalizability using a large heterogeneous brain datasets from the ENIGMA-PGC PTSD Working group. Methods: We analyzed brain MRI data from 3,477 structural-MRI; 2,495 resting state-fMRI; and 1,952 diffusion-MRI. First, we identified the brain features that best distinguish individuals with PTSD from controls using traditional machine learning methods. Second, we assessed the utility of the denoising variational autoencoder (DVAE) and evaluated its classification performance. Third, we assessed the generalizability and reproducibility of both models using leave-one-site-out cross-validation procedure for each modality. Results: We found lower performance in classifying PTSD vs. controls with data from over 20 sites (60 % test AUC for s-MRI, 59 % for rs-fMRI and 56 % for D-MRI), as compared to other studies run on single-site data. The performance increased when classifying PTSD from HC without trauma history in each modality (75 % AUC). The classification performance remained intact when applying the DVAE framework, which reduced the number of features. Finally, we found that the DVAE framework achieved better generalization to unseen datasets compared with the traditional machine learning frameworks, albeit performance was slightly above chance. Conclusion: These results have the potential to provide a baseline classification performance for PTSD when using large scale neuroimaging datasets. Our findings show that the control group used can heavily affect classification performance. The DVAE framework provided better generalizability for the multi-site data. This may be more significant in clinical practice since the neuroimaging-based diagnostic DVAE classification models are much less site-specific, rendering them more generalizable. Show less
Alzheimer’s disease (AD) is a neurodegenerative disease characterized by memory loss and declined cognitive functioning. Brain changes in AD involve grey matter atrophy and changes in brain... Show moreAlzheimer’s disease (AD) is a neurodegenerative disease characterized by memory loss and declined cognitive functioning. Brain changes in AD involve grey matter atrophy and changes in brain function. These different brain characteristics can respectively be visualized with structural and functional MRI scans. These MRI modalities have been used for AD classification, but studies typically only include a limited number of features. In this thesis we derived multiple types of features from each MRI modality, and combined those to discriminate AD patients and elderly controls. First, we showed that AD classification accuracy increases when combining multiple types of measures from a single MRI modality. This was shown for structural MRI scans in chapter 2, and for resting state fMRI scans in chapter 3. In chapter 4 we evaluated whether MRI based AD classification models can discriminate AD in a diverse clinical population as well. This worked to some extent, and it worked best using structural MRI scans. In chapter 5 we used baseline multimodal MRI scans from the same diverse clinical population to predict two-year follow-up cognitive decline. Decline was predicted above chance level for the MMSE, but not for six other neuropsychological tests. Show less