Conductance-Based Structural Brain Connectivity in Aging and Dementia

Abstract

Background: Structural brain connectivity has been shown to be sensitive to the changes that the brain undergoes during Alzheimer's disease (AD) progression. Methods: In this work, we used our recently proposed structural connectivity quantification measure derived from diffusion magnetic resonance imaging, which accounts for both direct and indirect pathways, to quantify brain connectivity in dementia. We analyzed data from the second phase of Alzheimer's Disease Neuroimaging Initiative and third release in the Open Access Series of Imaging Studies data sets to derive relevant information for the study of the changes that the brain undergoes in AD. We also compared these data sets to the Human Connectome Project data set, as a reference, and eventually validated externally on two cohorts of the European DTI Study in Dementia database. Results: Our analysis shows expected trends of mean conductance with respect to age and cognitive scores, significant age prediction values in aging data, and regional effects centered among subcortical regions, and cingulate and temporal cortices. Discussion: Results indicate that the conductance measure has prediction potential, especially for age, that age and cognitive scores largely overlap, and that this measure could be used to study effects such as anticorrelation in structural connections. Impact statement This work presents a methodology and a set of analyses that open new possibilities in the study of healthy and pathological aging. The methodology used here is sensitive to direct and indirect pathways in deriving brain connectivity measures from diffusion-weighted magnetic resonance imaging, and therefore provides information that many state-of-the-art methods do not account for. As a result, this technique may provide the research community with ways to detect subtle effects of healthy aging and Alzheimer's disease.

Keywords: Alzheimer's disease; aging; brain connectivity; conductance; diffusion MRI.

FIG. 1.

Demographics of the AD data sets used here. (a) ADNI-2 cohort of 213 subjects (77 CN, 89 MCI, 47 AD), (b) OASIS-3 group of 652 subjects, consisting of 4 cohorts, each of which had more than 100 subjects with similar description in the “Scans” field of the data sheet: (c) 272-subject cohort (187 CN, 38 AD, 47 other dementias), (d) 139-subject cohort (86 CN, 33 AD, 20 other dementias), (e) 125-subject cohort (112 CN, 4 AD, 9 other dementias), and (f) 116-subject cohort (103 CN, 6 AD, 7 other dementias). Other dementias included vascular dementia, or AD dementia with depression or additional symptoms (refer to Fig. 2). (g) EDSD Freiburg cohort (16 CN, 11 MCI, 10 AD). (h) EDSD Rostock-3T cohort (20 CN, 22 MCI, 15 AD). AD, Alzheimer's disease; ADNI-2, second phase of Alzheimer's Disease Neuroimaging Initiative; CN, cognitively normal; EDSD, European DTI Study in Dementia; MCI, mild cognitive impairment; OASIS-3, third release in the Open Access Series of Imaging Studies.

FIG. 2.

Distribution of the MMSE score for each diagnostic category in each dementia database, color-coded by the CDR values of 0 (blue), 0.5 (green), 1 (orange), 2 (red), and 3 (brown) (refer to Fig. 3, right, for the color map). Diagnostic labels are quoted from the databases. CDR was not available for HCP and EDSD. CDR, Clinical Dementia Rating; HCP, Human Connectome Project; MMSE, Mini–Mental State Examination.

FIG. 3.

Correlation of mean conductance with age and cognitive scores of CDR and MMSE, color-coded with respect to the CDR: 0 (blue), 0.5 (green), 1 (orange), 2 (red), and 3 (brown). CDR was not available for HCP.

FIG. 4.

Sig-values for the correlation of conductance with age, CDR, and MMSE. We depict the negative logarithm of the Bonferroni-corrected p-value ($sig=-{log}_{10}\left(p_b\right)$), and consider significant values above 1.3 (i.e., $p_b<0.05$). In this figure, all OASIS-3 cohorts were used together.

FIG. 5.

Sig-values for the correlation of conductance with CDR and MMSE, once the age and sex effects have been removed. We show the negative logarithm of the Bonferroni-corrected p-value ($sig=-{log}_{10}\left(p_b\right)$), and consider significant values above 1.3 (i.e., $p_b<0.05$). In this figure, all OASIS-3 cohorts were used together.

FIG. 6.

Negative correlation between the insula/caudate and the precentral/entorhinal structural connections in the left and right hemispheres, across the ADNI-2 (top) and 272-subject cohort of OASIS-3 (bottom) populations. CDR values are encoded in the colors of the dots: 0 (blue), 0.5 (green), 1 (orange), 2 (red), and 3 (brown) (refer to Fig. 3).