Automated MRI measures identify individuals with mild cognitive impairment and Alzheimer's disease

Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA 02129, USA.
Brain (Impact Factor: 9.2). 06/2009; 132(Pt 8):2048-57. DOI: 10.1093/brain/awp123
Source: PubMed


Mild cognitive impairment can represent a transitional state between normal ageing and Alzheimer's disease. Non-invasive diagnostic methods are needed to identify mild cognitive impairment individuals for early therapeutic interventions. Our objective was to determine whether automated magnetic resonance imaging-based measures could identify mild cognitive impairment individuals with a high degree of accuracy. Baseline volumetric T1-weighted magnetic resonance imaging scans of 313 individuals from two independent cohorts were examined using automated software tools to identify the volume and mean thickness of 34 neuroanatomic regions. The first cohort included 49 older controls and 48 individuals with mild cognitive impairment, while the second cohort included 94 older controls and 57 mild cognitive impairment individuals. Sixty-five patients with probable Alzheimer's disease were also included for comparison. For the discrimination of mild cognitive impairment, entorhinal cortex thickness, hippocampal volume and supramarginal gyrus thickness demonstrated an area under the curve of 0.91 (specificity 94%, sensitivity 74%, positive likelihood ratio 12.12, negative likelihood ratio 0.29) for the first cohort and an area under the curve of 0.95 (specificity 91%, sensitivity 90%, positive likelihood ratio 10.0, negative likelihood ratio 0.11) for the second cohort. For the discrimination of Alzheimer's disease, these three measures demonstrated an area under the curve of 1.0. The three magnetic resonance imaging measures demonstrated significant correlations with clinical and neuropsychological assessments as well as with cerebrospinal fluid levels of tau, hyperphosphorylated tau and abeta 42 proteins. These results demonstrate that automated magnetic resonance imaging measures can serve as an in vivo surrogate for disease severity, underlying neuropathology and as a non-invasive diagnostic method for mild cognitive impairment and Alzheimer's disease.

Download full-text


Available from: Nick Schmansky, Oct 02, 2015
67 Reads
  • Source
    • "We choose an area of language where we have firm evidence on first language acquisition, where properties of developmental ordering have been argued to have a universal component, and where a uniform methodology is possible . We compare these findings to those from a population with mild cognitive impairment (MCI), a known precursor to AD in which recent neuroscientific study of MCI suggests that specific hypotheses can now be generated regarding a course of cortical degeneration in the pathway from MCI to AD (e.g., Desikan et al., 2009; Greene & Killiany, 2010; Hanggi, Streffer, Jancke, & Hock, 2011; Harasty, Halliday, Kril, & Code, 1999; Spreng & Turner, 2013). "
    [Show abstract] [Hide abstract]
    ABSTRACT: We consider the regression or retrogenesis hypothesis, which argues that order of acquisition in development is reversed in neurodegeneration or pathology. Originally proposed as a regression hypothesis for the study of memory disorders, specifically retrograde amnesia, by Ribot (1881), it has been extended to the study of brain aging and pathology and to language. We investigate this hypothesis in a new study of language development, aging, and pathology. Through interuniversity collaboration using a matched experimental design and task, we compare production of complex sentences containing relative clauses by normal monolingual children during normal development, healthy young adults, healthy aging adults, and aging adults diagnosed with mild cognitive impairment, a recognized potential harbinger of Alzheimer's disease. Our results refute the regression hypothesis in this area and lead to potential syntactic markers for prodromal Alzheimer's disease and predictions for future brain imaging analyses. Copyright © 2015 Elsevier Inc. All rights reserved.
    Brain and Language 02/2015; 143C:1-10. DOI:10.1016/j.bandl.2015.01.013 · 3.22 Impact Factor
  • Source
    • "The identified anatomical regions mainly correspond to cerebral cortex, part of temporal lobe, parietal lobe, and frontal lobe (Braak and Braak, 1991; Desikan et al., 2009; Yao et al., 2012). With AD, patients experience significant widespread damage over the brain, causing shrinkage of brain volume (Yao et al., 2012; Harasty et al., 1999) and thinning of cortical thickness (Desikan et al., 2009; Yao et al., 2012). The affected brain regions include those involved in controlling language (Broca's area) (Harasty et al., 1999), reasoning (superior and inferior frontal gyri) (Harasty et al., 1999), part of sensory area (primary auditory cortex, olfactory cortex , insula, and operculum) (Braak and Braak, 1991; Lee et al., 2013), somatosensory association area (Yao et al., 2012; Tales et al., 2005; Mapstone et al., 2003), memory loss (hippocampus) (den Heijer et al., 2010), and motor function (Buchman and Bennett , 2011). "
    [Show abstract] [Hide abstract]
    ABSTRACT: In an increasing number of neuroimaging studies, brain images, which are in the form of multidimensional arrays (tensors), have been collected on multiple subjects at multiple time points. Of scientific interest is to analyze such massive and complex longitudinal images to diagnose neurodegenerative disorders and to identify disease relevant brain regions. In this article, we treat those problems in a unifying regression framework with image predictors, and propose tensor generalized estimating equations (GEE) for longitudinal imaging analysis. The GEE approach takes into account intra-subject correlation of responses, whereas a low rank tensor decomposition of the coefficient array enables effective estimation and prediction with limited sample size. We propose an efficient estimation algorithm, study the asymptotics in both fixed $p$ and diverging $p$ regimes, and also investigate tensor GEE with regularization that is particularly useful for region selection. The efficacy of the proposed tensor GEE is demonstrated on both simulated data and a real data set from the Alzheimer's Disease Neuroimaging Initiative (ADNI).
  • Source
    • "In recent literature a vast variety of features has been extracted from MR images for the identification of AD. These features are voxel based [14, 18, 40], vertex based [20, 41], or ROI based features [21, 29]. Features that we have used in proposed approach are volume of GM, WM, and CSF and size of hippocampus. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Structural brain imaging is playing a vital role in identification of changes that occur in brain associated with Alzheimer's disease. This paper proposes an automated image processing based approach for the identification of AD from MRI of the brain. The proposed approach is novel in a sense that it has higher specificity/accuracy values despite the use of smaller feature set as compared to existing approaches. Moreover, the proposed approach is capable of identifying AD patients in early stages. The dataset selected consists of 85 age and gender matched individuals from OASIS database. The features selected are volume of GM, WM, and CSF and size of hippocampus. Three different classification models (SVM, MLP, and J48) are used for identification of patients and controls. In addition, an ensemble of classifiers, based on majority voting, is adopted to overcome the error caused by an independent base classifier. Ten-fold cross validation strategy is applied for the evaluation of our scheme. Moreover, to evaluate the performance of proposed approach, individual features and combination of features are fed to individual classifiers and ensemble based classifier. Using size of left hippocampus as feature, the accuracy achieved with ensemble of classifiers is 93.75%, with 100% specificity and 87.5% sensitivity.
    Computational and Mathematical Methods in Medicine 09/2014; 2014:862307. DOI:10.1155/2014/862307 · 0.77 Impact Factor
Show more