Age Correction in Dementia – Matching to a Healthy Brain

Abstract

In recent research, many univariate and multivariate approaches have been proposed to improve automatic classification of various dementia syndromes using imaging data. Some of these methods do not provide the possibility to integrate possible confounding variables like age into the statistical evaluation. A similar problem sometimes exists in clinical studies, as it is not always possible to match different clinical groups to each other in all confounding variables, like for example, early-onset (age<65 years) and late-onset (age≥65) patients with Alzheimer's disease (AD). Here, we propose a simple method to control for possible effects of confounding variables such as age prior to statistical evaluation of magnetic resonance imaging (MRI) data using support vector machine classification (SVM) or voxel-based morphometry (VBM). We compare SVM results for the classification of 80 AD patients and 79 healthy control subjects based on MRI data with and without prior age correction. Additionally, we compare VBM results for the comparison of three different groups of AD patients differing in age with the same group of control subjects obtained without including age as covariate, with age as covariate or with prior age correction using the proposed method. SVM classification using the proposed method resulted in higher between-group classification accuracy compared to uncorrected data. Further, applying the proposed age correction substantially improved univariate detection of disease-related grey matter atrophy using VBM in AD patients differing in age from control subjects. The results suggest that the approach proposed in this work is generally suited to control for confounding variables such as age in SVM or VBM analyses. Accordingly, the approach might improve and extend the application of these methods in clinical neurosciences.

A Novel Grading Biomarker for the Prediction of Conversion From Mild Cognitive Impairment to Alzheimer's Disease

“…(2) Since normal aging has similar atrophy effects on certain regions as AD [20], [25], [48], it would cause a confounding effect on using the diseasespecific changes for classification. Thus, a linear regression model [25], [49] was used to remove the confounding effect of normal aging. Although the improvement by adding age correction is not significant, it is consistently helpful for all the classification experiments as shown in Figures 4 and 5.…”

“…The effects of normal aging were then estimated on the voxel-wise intensities of NC subjects by fitting a linear regression model [25] at each location. Given that there are N healthy subjects and each image contains M brain voxels, the normalized intensity values of the preprocessed MR images of all NC subjects can be represented by a matrix…”

“…For example, in a recent challenge [24], eight out of fifteen teams incorporated aging effects in their algorithms for the classification of AD, where the normal aging effect was eliminated using age-dependent normalization [24] or regression [20], [24]. The removal of effects associated with normal aging has been reported to improve the classification performance significantly [20], [25] or slightly [22] depending on the adopted features and classification methods.…”

Ensemble Learning Incorporating Uncertain Registration

“…Subject age is regressed out for each voxel based on the empirical expectation of P (d|w) for the healthy controls in the training set using ordinary least squares linear regression [36], and each voxel is given 0 mean and unit standard deviation.…”

Anatomically Constrained Weak Classifier Fusion for Early Detection of Alzheimer’s Disease

“…For each intermediate classifier, we estimated the age-related effect on the CN population using linear regression. Intermediate classifiers in the MCI populations were then corrected using the estimated linear regression coefficients (see [14]). …”