Machine learning methods for predicting progression from mild cognitive impairment to Alzheimer’s disease dementia: a systematic review

Abstract

Background An increase in lifespan in our society is a double-edged sword that entails a growing number of patients with neurocognitive disorders, Alzheimer’s disease being the most prevalent. Advances in medical imaging and computational power enable new methods for the early detection of neurocognitive disorders with the goal of preventing or reducing cognitive decline. Computer-aided image analysis and early detection of changes in cognition is a promising approach for patients with mild cognitive impairment, sometimes a prodromal stage of Alzheimer’s disease dementia. Methods We conducted a systematic review following PRISMA guidelines of studies where machine learning was applied to neuroimaging data in order to predict whether patients with mild cognitive impairment might develop Alzheimer’s disease dementia or remain stable. After removing duplicates, we screened 452 studies and selected 116 for qualitative analysis. Results Most studies used magnetic resonance image (MRI) and positron emission tomography (PET) data but also magnetoencephalography. The datasets were mainly extracted from the Alzheimer’s disease neuroimaging initiative (ADNI) database with some exceptions. Regarding the algorithms used, the most common was support vector machine with a mean accuracy of 75.4%, but convolutional neural networks achieved a higher mean accuracy of 78.5%. Studies combining MRI and PET achieved overall better classification accuracy than studies that only used one neuroimaging technique. In general, the more complex models such as those based on deep learning, combined with multimodal and multidimensional data (neuroimaging, clinical, cognitive, genetic, and behavioral) achieved the best performance. Conclusions Although the performance of the different methods still has room for improvement, the results are promising and this methodology has a great potential as a support tool for clinicians and healthcare professionals.

Oscillatory Activity of the Hippocampus in Prodromal Alzheimer’s Disease: A Source-Space Magnetoencephalography Study

“…Besides augmenting our understanding of AD pathophysiology, new measures of altered hippocampal activity could aid in developing stronger prognostic and/or diagnostic markers, since current EEG/MEG-based markers lack specificity [ 3 ]. However, larger groups, longitudinal data, multimodal data, and machine-learning based approaches may be required [ 75 ]. So far, relatively straightforward measures such as spectral power perform best, so using them to look at specific regions or conditions may be an equally valid approach.…”

Machine Learning Decomposition of the Anatomy of Neuropsychological Deficit in Alzheimer’s Disease and Mild Cognitive Impairment

“…This was performed in order to identify specific anatomical regions associated with cognitive test performance, rather than develop a diagnostic tool. Nonetheless, many studies utilizing machine learning have reported excellent performance, with accuracy, sensitivity and specificity at times exceeding 99% ( Odusami et al, 2021 ), though a recent systematic review demonstrated a mean accuracy of 75.4% for support vector machines in predicting progression of MCI to AD ( Grueso and Viejo-Sobera, 2021 ). However, a primary reason preventing the implementation of these models in clinical practice is their replicability, where the same models which have often overfitted to the study sample will underperform with independent data sets ( Beam et al, 2020 ; Crowley et al, 2020 ).…”

“…Due to the advancement of machine learning algorithms and new data, we have the opportunity to analyze large data sets and build prediction models to inform clinical practice ( Panch et al, 2018 ; Davenport and Kalakota, 2019 ). Various studies have utilized machine learning to predict the conversion of MCI to AD based on neuropsychological measures and clinical markers, with several studies focusing on neuroimaging models ( Huang et al, 2020 ; Stamate et al, 2020 ; Syaifullah et al, 2020 ; Grueso and Viejo-Sobera, 2021 ). Wee et al (2012) employed support vector machines to demonstrate that a multimodal approach combining functional and structural connectivity data improved the accuracy of MCI classification.…”

Ten years of image analysis and machine learning competitions in dementia

“…While the challenges using ADNI for testing took measures to limit the advantage of researchers being more familiar to the data, the challenge results may still be biased towards its specific data characteristics. Data from the ADNI plays a major role in the research field of machine learning in Alzheimer's disease and is used for optimization and validation in 60 − 90% of the published articles (Rathore et al, 2017;Grueso and Viejo-Sobera, 2021). Therefore, state-of-theart performance in ADNI is likely to be an overestimation of performance in clinical practice.…”