Combining anatomical, diffusion, and resting state functional magnetic resonance imaging for individual classification of mild and moderate Alzheimer's disease

Abstract

Magnetic resonance imaging (MRI) is sensitive to structural and functional changes in the brain caused by Alzheimer's disease (AD), and can therefore be used to help in diagnosing the disease. Improving classification of AD patients based on MRI scans might help to identify AD earlier in the disease's progress, which may be key in developing treatments for AD. In this study we used an elastic net classifier based on several measures derived from the MRI scans of mild to moderate AD patients (N = 77) from the prospective registry on dementia study and controls (N = 173) from the Austrian Stroke Prevention Family Study. We based our classification on measures from anatomical MRI, diffusion weighted MRI and resting state functional MRI. Our unimodal classification performance ranged from an area under the curve (AUC) of 0.760 (full correlations between functional networks) to 0.909 (grey matter density). When combining measures from multiple modalities in a stepwise manner, the classification performance improved to an AUC of 0.952. This optimal combination consisted of grey matter density, white matter density, fractional anisotropy, mean diffusivity, and sparse partial correlations between functional networks. Classification performance for mild AD as well as moderate AD also improved when using this multimodal combination. We conclude that different MRI modalities provide complementary information for classifying AD. Moreover, combining multiple modalities can substantially improve classification performance over unimodal classification.

Single Subject Classification of Alzheimer’s Disease and Behavioral Variant Frontotemporal Dementia Using Anatomical, Diffusion Tensor, and Resting-State Functional Magnetic Resonance Imaging

“…Furthermore, in line with previous work, we observed that concatenating all measures did not result in the most accurate classification model and as more measures were added the level of improvement became less apparent [29,31,61]. The marginal improvements observed for some measures or combinations may indicate a negligible contribution to the model.…”

“…Nevertheless, the role of rsfMRI as a separate measure for disease-type classifications has not yet been established. It has found limited application for single subject disease-type classifications [27,29,30,32]. In AD versus control classifications its performance rates are consistently among the lowest compared with other MRI measures [29,30,32] and we also found that rs-fMRI as a single measure was neither effective in differentiating dementia types nor between patients and controls subjects.…”

“…Substantial improvement was for instance not observed for AD versus control classifications while others did [29,32] or did not show benefit [30] from GM, DTI, and rs-fMRI measures. Furthermore, in line with previous work, we observed that concatenating all measures did not result in the most accurate classification model and as more measures were added the level of improvement became less apparent [29,31,61].…”

“…Multiparametric MRI-based statistical algorithms integrate various MRI measures to compute a single, quantitative probabilistic index and have shown to be more accurate in differentiating cognitively healthy from demented patients than single MRI measure-based algorithms alone [18,19,29,31,32]. We speculate that MRI-based classification algorithms that differentiate dementia-types benefit from integrating structural and functional connectivity measures.…”

“…Differences in functional connectivity strengths as measured with restingstate functional MRI (rs-fMRI) have been heralded as early and useful markers to differentiate AD and bvFTD patients [27,28]. However, the contribution of rs-fMRI measures as a (complementary) measure to differentiate dementia-types on an individual basis is still largely unclear [27,29,30].…”