Comparative analysis of machine learning algorithms for multi-syndrome classification of neurodegenerative syndromes

Abstract

Importance The entry of artificial intelligence into medicine is pending. Several methods have been used for the predictions of structured neuroimaging data, yet nobody compared them in this context. Objective Multi-class prediction is key for building computational aid systems for differential diagnosis. We compared support vector machine, random forest, gradient boosting, and deep feed-forward neural networks for the classification of different neurodegenerative syndromes based on structural magnetic resonance imaging. Design, setting, and participants Atlas-based volumetry was performed on multi-centric T1-weighted MRI data from 940 subjects, i.e., 124 healthy controls and 816 patients with ten different neurodegenerative diseases, leading to a multi-diagnostic multi-class classification task with eleven different classes. Interventions N.A. Main outcomes and measures Cohen’s kappa, accuracy, and F1-score to assess model performance. Results Overall, the neural network produced both the best performance measures and the most robust results. The smaller classes however were better classified by either the ensemble learning methods or the support vector machine, while performance measures for small classes were comparatively low, as expected. Diseases with regionally specific and pronounced atrophy patterns were generally better classified than diseases with widespread and rather weak atrophy. Conclusions and relevance Our study furthermore underlines the necessity of larger data sets but also calls for a careful consideration of different machine learning methods that can handle the type of data and the classification task best.

Artificial neural networks for non-linear age correction of diffusion metrics in the brain

“…Data augmentation, which can also be considered a form of regularization ( Kukaèka et al, 2017 ), is common in advanced ML techniques such as deep neural networks ( Abadi et al, 2016 ). In various neurodegenerative diseases, ML models based on neuroimaging data can strengthen diagnostic accuracy ( Lampe et al, 2022 ). However, collecting a large amount of neuroimaging data in rare brain diseases is often challenging, as it is to train complex yet accurate ML models ( Castiglioni et al, 2021 ).…”

A review on the new age methodologies for early detection of Alzheimer's and Parkinson's disease

“…36 While neural networks performed best overall, smaller classes were better classified by ensemble methods and support vector machines. 36 Diseases with specific atrophy patterns fared better.…”

“…ML and novel biomarkers such as neuronal injury and neuroinflammation for AD diagnosis, in addition to traditional biomarkers like Aβ and tau, show potential for improved accuracy. Lampe et al compared various ML methods for classifying NDDs based on structural MRI data from 940 subjects 36 . While neural networks performed best overall, smaller classes were better classified by ensemble methods and support vector machines 36 .…”

“…Lampe et al compared various ML methods for classifying NDDs based on structural MRI data from 940 subjects 36 . While neural networks performed best overall, smaller classes were better classified by ensemble methods and support vector machines 36 . Diseases with specific atrophy patterns fared better.…”

Opaque Ontology: Neuroimaging Classification of ICD-10 Diagnostic Groups in the UK Biobank

“…Machine learning tools are increasingly adopted in clinical care settings, yet their disease-specificity cannot be assessed in single-disease studies. Although some transdiagnostic comparisons have been performed across related disorders (e.g., bipolar-schizophrenia [6]; mild cognitive impairment-Alzheimer’s disease [7]), broader comparisons are lacking. Second, our findings provide a comprehensive benchmark for future research on diagnostic classification models in the UKB and beyond.…”