Elsevier

NeuroImage

Volume 18, Issue 4, April 2003, Pages 895-907
NeuroImage

Regular article
A comprehensive study of gray matter loss in patients with Alzheimer’s disease using optimized voxel-based morphometry

https://doi.org/10.1016/S1053-8119(03)00041-7Get rights and content

Abstract

Voxel-based morphometry (VBM) has already been applied to MRI scans of patients with Alzheimer’s disease (AD). The results of these studies demonstrated atrophy of the hippocampus, temporal pole, and insula, but did not describe any global brain changes or atrophy of deep cerebral structures. We propose an optimized VBM method, which accounts for these shortcomings. Additional processing steps are incorporated in the method, to ensure that the whole spectrum of brain atrophy is visualized. A local group template was created to avoid registration bias, morphological opening was performed to eliminate cerebrospinal fluid voxel misclassifications, and volume preserving modulation was used to correct for local volume changes. Group differences were assessed and thresholded at P < 0.05 (corrected). Our results confirm earlier findings, but additionally we demonstrate global cortical atrophy with sparing of the sensorimotor cortex, occipital poles, and cerebellum. Moreover, we show atrophy of the caudate head nuclei and medial thalami. Our findings are in full agreement with the established neuropathological descriptions, offering a comprehensive view of atrophy patterns in AD.

Introduction

Alzheimer’s disease (AD) will become one of the major public health problems for the Western World in the coming decades. Aging of the population will lead to more and more people developing AD. In the long run, caring for AD patients will be a major financial backlash for the health care system Gutterman et al 1999, Trabucchi 1999. In order to provide care for these patients though, it is important to diagnose this disease and differentiate it from other dementia subtypes.

Medial temporal lobe (MTL) atrophy, as seen on MRI scans of AD patients, is a sensitive marker of AD, even in mild forms of the disease de Leon et al 1993, Fox et al 1996, Scheltens et al 1992. There are various ways to estimate brain atrophy on MRI. Visual inspection for the presence of atrophy (Scheltens et al., 1997) may be adequate in a routine clinical setting, but is not enough when quantitative measures are needed, for example, to estimate rate of tissue loss during a clinical trial. To assess such changes in a single brain structure, the region-of-interest (ROI) analysis technique may be employed (Pruessner et al., 2000), in which an experienced operator outlines the structure in question in a series of contiguous sections on a computer screen. ROI analysis constitutes, up to now, the gold standard in brain atrophy measurements (Jack et al., 1992), but there are major shortcomings such as observer/operator dependency and bias in brain structure and anatomical region boundary selection (Pruessner et al., 2000). In order to overcome these shortcomings, automated techniques have been developed Freeborough and Fox 1997, Freeborough and Fox 1998, Frisoni et al 1996, Smith et al 2001, Thompson et al 2000, the goal of which is to automatically analyze whole-brain structural MRI scans, avoiding a priori selection of regions and eliminating observer variability.

Voxel-based morphometry (VBM) is one such method developed for automated unbiased analysis of structural MRI scans. It has been used in a variety of settings and diseases (Ashburner and Friston, 2000), including VBM analysis of brain tissue loss in AD Baron et al 2001, Rombouts et al 2000. However, the initial implementation of VBM has led to debate about the limitations of affine and low-order registration algorithms Ashburner and Friston 2001, Bookstein 2001. It has been argued that imperfect registration of MRI scans to a common template can lead to false estimates of atrophy (Bookstein, 2001). Moreover, tissue classification errors during automated segmentation of brain tissue classes will produce an artificial thin rim of periventricular gray matter. In the current analysis, we tackled these two issues and applied dementia-specific VBM to study patterns of atrophy in AD compared to elderly controls, attempting to visualize similar atrophy patterns on the MRI analysis to the ones described neuropathologically Braak et al 1999, Braak and Braak 1991.

Section snippets

Subjects

Twenty-five patients with AD and twenty-five healthy controls of comparable age, recruited among spouses and friends of the patients, were included in the study. Written informed consent was obtained for all subjects. Approval of performance of MRI scans by local ethics committee was granted. All dementia subjects underwent a neuropsychological test battery. Cognitive function was measured using the Cambridge Mental Disorders of the Elderly Examination (CAMCOG) (Roth et al., 1986), which

Display of results

For qualitative analysis, mean gray matter images for each group were created and visually inspected (Fig. 5). These images did not demonstrate any collateral “spreading” of gray matter structures such as the caudate nucleus and medial thalamus. For the visualization of quantitative results areas of significant change were displayed in the coronal plane and overlaid on the group specific template (Fig. 3). We avoid the “glass brain” display, because we do not find it informative enough when

Discussion

It is well known that Alzheimer’s disease is characterized by cerebral atrophy. A recent review describes the current hypotheses about atrophy patterns in AD (Smith, 2002). In the very early stages only the entorhinal cortex is affected. Later, atrophy spreads to the hippocampus proper and the MTL. Projection neurons from the MTL die, disrupting their neurotrophic support to pyramidal cells in the association areas of the cortex. Based on projections from the MTL, the cortex in the

Conclusion

The optimized VBM we describe here allows a comprehensive in vivo analysis of patterns of atrophy in AD, and the results are in perfect agreement with histopathological findings. In addition to hippocampal and medial temporal lobe involvement, diffuse cortical atrophy with sparing of sensorimotor cortex, occipital poles, and cerebellum was noted. Furthermore, insula, caudate head, medial thalamus, and cingulum also appeared to be involved.

Acknowledgements

G. Karas is the recipient of Grant 2001-014 from Stichting Alzheimer Nederland and S.A.R.B. Rombouts is the recipient of Grant H00.17 from Hersenstichting Nederland. Additional funds were received from the Stichting Alzheimer and Neuropsychiatry Foundation, Amsterdam.

References (44)

  • C.D. Good et al.

    A voxel-based morphometric study of ageing in 465 normal adult human brains

    Neuroimage

    (2001)
  • T.L. Jernigan et al.

    Cerebral structure on MRI, Part II: specific changes in Alzheimer’s and Huntington’s diseases

    Biol. Psychiatry

    (1991)
  • S.M. Montgomery

    Depressive symptoms in acute schizophrenia

    Prog. Neuropsychopharmacol.

    (1979)
  • S.A. Rombouts et al.

    Unbiased whole-brain analysis of gray matter loss in Alzheimer’s disease

    Neurosci. Lett.

    (2000)
  • A.W. Toga et al.

    The role of image registration in brain mapping

    Image Vision Comput.

    (2001)
  • J. Ashburner et al.

    Nonlinear spatial normalization using basis functions

    Hum. Brain Mapp.

    (1999)
  • E. Braak et al.

    Neuropathology of Alzheimer’s diseasewhat is new since A. Alzheimer?

    Eur. Arch. Psychiatry Clin. Neurosci.

    (1999)
  • H. Braak et al.

    Neuropathological stageing of Alzheimer-related changes

    Acta Neuropathol.

    (1991)
  • A. Brun et al.

    Regional pattern of degeneration in Alzheimer’s disease: neuronal loss and histopathological grading

    Histopathology

    (1981)
  • M. de Leon et al.

    Usefulness of MRI measures of entorhinal cortex versus hippocampus in AD

    Neurology

    (2001)
  • M.J. de Leon et al.

    The radiologic prediction of Alzheimer disease: the atrophic hippocampal formation

    AJNR Am. J. Neuroradiol.

    (1993)
  • T. Erkinjuntti et al.

    Temporal lobe atrophy on magnetic resonance imaging in the diagnosis of early Alzheimer’s disease

    Arch. Neurol.

    (1993)
  • Cited by (378)

    • Smoking is associated with lower brain volume and cognitive differences: A large population analysis based on the UK Biobank

      2023, Progress in Neuro-Psychopharmacology and Biological Psychiatry
      Citation Excerpt :

      In this study, the combination of the statistical power of the large brain-imaging sample of the UK Biobank and the use of a highly robust metric (i.e., GMV) made it possible to discover widespread associations between smoking and brain GMV (Madan and Kensinger, 2017). Lower brain volume in chronic and regular smokers has also been associated with a higher risk for the neurocognitive disorder (Karas et al., 2003; Knight et al., 2016; Lee et al., 2013). Very interestingly, the current investigation revealed a high correlation between the lower GMV and the amount of smoking (“Current smoker” group, r = 0.94, p < 0.001, Fig. S7).

    View all citing articles on Scopus
    View full text