Skip to main content
NCBI home page
Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring logoLink to Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring
. 2025 Oct 15;17(4):e70192. doi: 10.1002/dad2.70192

The impact of amyloid beta burden on white matter dysfunction and associated transcriptomic signatures in cognitively normal elderly individuals

Ziyun Li 1,2, Yuxiao Sun 1,2, Ting Li 1,2, Jiawen Liu 1,2, Biying Peng 1,2, Zhong Li 1,2, Lei Liu 3, Yaojing Chen 1,2,✉, Zhanjun Zhang 1,2,✉; for the Alzheimer's Disease Neuroimaging Initiative
PMCID: PMC12522070  PMID: 41104121

Abstract

INTRODUCTION

Amyloid beta (Aβ), a hallmark of early Alzheimer's disease (AD), disrupts white matter (WM) microstructure, but its spatial patterns and transcriptomic links in cognitively normal individuals remain underexplored.

METHODS

We compared the WM microstructure between Aβ‐positive (Aβ+) and Aβ‐negative (Aβ−) individuals at the cognitively normal stage. We investigated the relationship between the fibers and the cortical and subcortical regions to which they are connected, as well as the underlying gene expression.

RESULTS

WM damage observed in Aβ+ individuals was characterized across eight fiber tracts, even prior to the evidence of atrophy and during the cognitive normal stage. This damage is primarily associated with cortical Aβ accumulation and may be linked to genes that regulate oligodendrocyte function and myelination.

DISCUSSION

Cortical Aβ‐related WM changes precede gray matter atrophy in preclinical AD, highlighting their potential as early biomarkers. Oligodendrocyte dysfunction and myelination pathways may underlie Aβ‐driven WM vulnerability, offering targets for intervention.

Highlights

  • WM microstructural changes precede gray matter atrophy in preclinical AD.

  • Aβ‐driven WM damage persists even after adjusting for age.

  • WM microstructural damage is primarily linked to cortical Aβ burden in cognitively normal individuals.

  • Oligodendrocytes and myelin underlie the vulnerability of WM‐related to Aβ.

Keywords: Alzheimer's disease, amyloid beta, transcriptomics, white matter microstructure

1. BACKGROUND

Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder among the elderly and the leading cause of dementia, accounting for approximately 60% to 80% of all dementia cases. 1 The aggregation of amyloid beta (Aβ) is recognized as an early pathological hallmark of AD, with this process commencing several decades before the clinical onset of dementia. 2 , 3 Recently, the National Institute on Aging and the Alzheimer's Association (NIA‐AA) further emphasized the significance of Aβ proteinopathy, designating it as a primary core 1 biomarker for detecting early pathological changes in AD. 4 Positron emission tomography (PET) facilitates in vivo imaging of Aβ aggregation, thereby enhancing the ability to detect and monitor Aβ‐associated changes in the brain. 5

AD has long been considered primarily a disorder of gray matter. However, accumulating evidence suggests that abnormalities in white matter (WM) also play a significant role in its pathology. 6 Notably, damage to WM integrity, as assessed by diffusion weighted imaging (DWI), has emerged as a prominent finding in AD. 7 , 8 , 9 Specifically, widespread disruptions in WM integrity are observed in patients exhibiting cognitive impairment. 10 While most prior studies focused on the later stages of AD, where cognitive impairment is evident, some explored the relationship between Aβ deposition and WM integrity during the cognitively normal (CN) stage. In particular, higher global Aβ deposition has been linked to reduced WM integrity in CN individuals. 11 Furthermore, another study identified a similar association in CN individuals, especially within tracts that are susceptible to Aβ accumulation, including the anterior cingulum, posterior cingulum, and uncinate fasciculus. 12 Transcriptomic studies enable the integration of macroscale WM abnormalities with microscale gene expression, providing valuable insights into the molecular mechanisms of disease. However, most current transcriptomic research has focused on the AD stage, 13 with limited investigation into earlier pathogenic processes. So far, evidence regarding Aβ‐related changes in WM integrity during the preclinical stage remains limited. Furthermore, existing studies have largely overlooked the spatial patterns of these changes, with insufficient attention paid to understanding the underlying transcriptomic mechanisms, which remains a significant gap in the field.

Based on this, we aim to investigate the impact of Aβ deposition on WM integrity and its spatial associations, as well as to explore the transcriptomic signatures underlying these effects in CN individuals. To achieve this, we utilized multimodal neuroimaging data from CN individuals in the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, categorizing them into Aβ‐positive (Aβ+, n = 83) and Aβ‐negative (Aβ−, n = 174) groups. We calculated WM metrics for whole‐brain fiber tracts to compare differences in Aβ deposition and WM integrity between the groups. Additionally, we evaluated the association between WM integrity metrics and regional Aβ deposition in connected brain regions. Finally, we employed gene expression profiles to elucidate the effects of Aβ deposition on WM integrity. We hypothesized that the impact of Aβ deposition on WM integrity would exhibit a spatial association, such that Aβ accumulation would affect the WM fiber tracts connected to the corresponding cortical regions. Furthermore, transcriptomic analysis may reveal the molecular mechanisms underlying this process.

2. METHODS

2.1. Participants

Data used in this study were obtained from the ADNI database. The study included subjects who had at least one T1‐weighted scan, one DWI scan, and one amyloid PET scan using 18F‐florbetapir (FBP), resulting in a total of 257 participants. All participants included in the current study were CN, as indicated by a Mini‐Mental State Examination (MMSE) score of 24 or higher, a Clinical Dementia Rating (CDR) of 0, and no reported cognitive concerns at the time of neuroimaging. On average, all subjects underwent magnetic resonance imaging (MRI) scans 26.5 ± 40.6 days prior to the amyloid PET scans. Further details are available in Supplementary Text 1.

2.2. MRI acquisition and analysis

All MRI acquisition protocols were standardized across different vendor platforms to ensure compatibility with ADNI. The preprocessing pipeline was executed using QSIPrep 14 version 0.19.1. TractSeg 15 version 2.9, a deep learning‐based method for WM segmentation, was employed to reconstruct fiber tracts. Volumetric T1‐weighted data were processed using FreeSurfer segmentation for all cortical and subcortical regions, based on the Desikan–Killiany atlas. 16 Additionally, the estimated total intracranial volume (TIV) was calculated to account for individual variability in head size. For each fiber tract (Figure S1), the following diffusion metrics were generated: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity, and radial diffusivity (RD). To assess the regional association between Aβ burden and fiber tract metrics, we identified the cortical and subcortical brain regions connected by the fiber tracts. To achieve this, we intersected the start and end masks of the fiber tracts obtained from TractSeg with the Desikan–Killiany atlas, thereby determining the brain regions traversed by the fiber tracts. Finally, to harmonize scans collected from different scanners, the measures were adjusted using the ComBat batch‐effect correction tool, 17 which has been shown to effectively harmonize multisite DWI data. 18 Further details are available in Supplementary Text 2.

2.3. PET acquisition and analysis

In this study, amyloid standardized uptake value ratio (SUVR) data were obtained from the University of California at Berkeley 18F‐florbetapir datasets within the ADNI. The computation of the regional amyloid SUVR for each region of interest (ROI) used the whole cerebellum as the reference region. The global SUVR was calculated by averaging the SUVR values from the frontal, anterior/posterior cingulate, lateral parietal, and lateral temporal regions. A cut‐off of 1.11 SUVR for the global SUVR level was established to determine a positive amyloid scan (Aβ+), with values below this threshold indicating a negative amyloid scan (Aβ−). 19 Additional information can be found in Supplementary Text 3.

2.4. WM integrity and transcriptomic profiles

As previously described, 20 , 21 brain‐wide gene expression maps in WM were constructed using the Allen Human Brain Atlas (AHBA), 22 which includes 3702 spatially distinct samples from six post mortem brains (mean age = 42.50 ± 13.38 years; male/female ratio = 5:1). The AHBA dataset bridges the gap between macroscale WM integrity and microscale gene expression. In brief, gene expression profiles in WM fiber tracts were computed using the Abagen toolbox version 0.1.4. Subsequently, partial least squares (PLS) regression 23 was employed to investigate the relationship between WM integrity and transcriptional activity across all genes. Metascape 24 was then utilized to identify enriched ontological terms within the PLS1 gene set. Finally, we investigated the overlap between the PLS1 genes and gene sets enriched in seven typical cell types: astrocytes (Astro), endothelial cells (Endo), microglia (Micro), excitatory neurons (NeuroEx), inhibitory neurons (NeuroIn), oligodendrocytes (Oligo), and oligodendrocyte precursors (OPC). Additional information can be found in Supplementary Text 4.

2.5. Statistical analyses

All statistical analyses were performed using R version 4.2.3. Multiple comparisons were controlled for using the false discovery rate (FDR) according to the Benjamini–Hochberg method. In the absence of a priori hypotheses regarding laterality and to minimize the number of comparisons, all measurements for the left and right hemispheres were averaged (the Aβ SUVR was volume‐weighted and averaged). Participant characteristics were compared between the Aβ+ and Aβ− groups using the chi‐squared test for categorical variables and two‐tailed t tests for continuous variables.

RESEARCH IN CONTEXT

  1. Systematic review: The authors conducted a systematic review of the literature using PubMed and cited relevant articles. The literature highlights Aβ as an early biomarker for AD; however, prior studies primarily focused on WM damage during symptomatic stages or on the global burden of Aβ. The spatial patterns of Aβ‐related WM, as well as the transcriptomic mechanisms in cognitively normal individuals, remain underexplored.

  2. Interpretation: Our findings indicate that cortical Aβ deposition selectively disrupts the integrity of WM in fiber tracts through oligodendrocyte dysfunction and myelination pathways, potentially preceding gray matter atrophy. This observation is consistent with emerging evidence of glial vulnerability in the early pathology of AD.

  3. Future directions: Investigating the neurobiological basis underlying the pathological processes of AD may yield insights into the relationship between Aβ deposition and the impairment of WM integrity. This research could also open new avenues for exploring disease pathogenesis and identifying potential therapeutic targets.

First, a general linear model (GLM) was used to examine the differences in (1) regional Aβ SUVR, (2) fiber tract metrics, and (3) regional volume between the Aβ− and Aβ+ groups in CN subjects. In this analysis, SUVR, fiber tract metrics, and regional volume each served as the dependent variables, while group status was the independent variable. Age, sex, apolipoprotein E (APOE) ɛ4 status, education, and TIV were included as covariates for the analysis. Fiber tracts exhibiting widespread significant metric differences were classified as damaged (p FDR < 0.05). Additionally, Cohen's d was calculated to assess the effect size of group differences in each brain region. Second, Pearson partial correlations were calculated using the ppcor package (version 1.0) to evaluate associations between damaged fiber tract metrics and (1) age (with covariates: sex, APOE ɛ4 status, and education); and (2) global SUVR (with covariates: age, sex, APOE ɛ4 status, and education). Third, for each damaged fiber tract, the associations between metrics and the regional SUVR of the connected cortical and subcortical regions were assessed, controlling for covariates such as age, sex, APOE ɛ4 status, and education. Finally, the Shapiro–Wilk test was employed to assess the normality of the absolute correlation coefficients. Depending on whether the data followed a normal distribution, either a t test or a Wilcoxon rank‐sum test was conducted to compare the association strengths between connected cortical and subcortical regions, as well as between connected and non‐connected cortical regions. Brain and tract visualizations were generated using DSI Studio 25 and the R packages ggplot2 26 and ggseg. 27

To verify the robustness of our findings, we conducted additional analyses controlling for cardiovascular risk factors, including body mass index (BMI), systolic blood pressure, and diastolic blood pressure.

3. RESULTS

3.1. Participant characteristics

The participant characteristics are presented in Table 1. The data used in this study were derived from 174 Aβ− (mean age: 72.6 ± 7.3 years) and 83 Aβ+ (mean age: 75.5 ± 7.0 years) CN subjects, each of whom had undergone at least one scan across all three imaging modalities. A significant difference was observed between the two groups, with the Aβ+ group being older (p < 0.001) and having a higher prevalence of APOE ɛ4 carriers (p < 0.001). No significant differences were found between the two groups regarding sex, education, cognitive assessments (MMSE, CDR, Alzheimer's Disease Assessment Scale [ADAS]), Functional Activities Questionnaire (FAQ) score, Geriatric Depression Scale (GDS) score, or psychiatric symptoms (Neuropsychiatric Inventory [NPI]).

TABLE 1.

Participant characteristics.

CN Aβ− CN Aβ+ Test statistic p value
N 174 83 – –
Age (years) 72.6 (7.3) 75.5 (7.0) 3.046 0.003
Sex (M/F) 95/79 56/27 3.329 0.068
Education (years) 16.6 (2.4) 16.6 (2.6) 0.010 0.992
APOE ɛ4 carriers (%) 23.6% 51.8% 19.111 <0.001
MMSE score 29.05 (1.23) 28.88 (1.15) 1.070 0.286
CDR‐SB score 0.04 (0.13) 0.04 (0.14) 0.268 0.789
ADAS‐Q4 score 2.57 (1.84) 2.80 (1.87) 0.893 0.373
ADAS‐11 score 5.09 (2.52) 5.57 (3.12) 1.306 0.193
ADAS‐13 score 7.95 (3.87) 8.77 (4.53) 1.508 0.133
GDS score 1.05 (1.40) 0.94 (1.36) 0.575 0.566
NPI score 1.52 (3.29) 1.65 (3.61) 0.282 0.779
FAQ score 0.23 (0.82) 0.29 (0.97) 0.510 0.611
Global Aβ SUVR 1.00 (0.52) 1.30 (0.17) 20.650 <0.001
Open in a new tab

Note: Values denoted as mean (standard) or frequency. Participants with at least one ɛ4 allele were considered APOE ɛ4 carriers. P values were derived from two‐tailed t‐tests for continuous measures and from chi‐squared tests for categorical measures. Bolded values indicate p < 0.05.

Abbreviations: ADAS, Alzheimer's Disease Assessment Scale; APOE ɛ4, apolipoprotein E ε4; Aβ, amyloid beta; CDR‐SB, Clinical Dementia Rating Scale Sum of Boxes; CN, cognitively normal; FAQ, Functional Activities Questionnaire; GDS, Geriatric Depression Scale; MMSE, Mini‐Mental State Examination; NPI, Neuropsychiatric Inventory; SUVR, standardized uptake value ratio.

3.2. Regional Aβ SUVR differences

To examine regional variation in Aβ deposition within the Aβ+ group, we compared the Aβ SUVR across all brain regions between the Aβ+ and Aβ− groups. The mean Aβ SUVR for cortical and subcortical regions in the Aβ+ and Aβ− groups is illustrated in Figures 1A and 1B. Significant regional differences in Aβ burden were identified between the Aβ+ and Aβ− groups, with individuals in the Aβ+ group demonstrating predominant deposition in cortical regions (Figure 1C).

FIGURE 1.

FIGURE 1

Open in a new tab

Regional differences in Aβ burden between CN Aβ+ and CN Aβ− group. (A) Average SUVR values in cortical and subcortical regions for the CN Aβ+ group, represented by color. (B) Average SUVR values in cortical and subcortical regions for CN Aβ− group, represented by color. (C) Difference in regional Aβ burden between CN Aβ+ and CN Aβ− group quantified with t value represented by color. Aβ burden is primarily in cortical regions. Only the left hemisphere is shown. Aβ, amyloid beta; CN, cognitively normal; SUVR, standardized uptake value ratio.

3.3. Regional volume and WM metric differences

To examine volumetric changes in the Aβ+ group, we compared the volumes of all brain regions between the Aβ+ and Aβ− groups. No significant volumetric differences were found between Aβ+ and Aβ− individuals across all brain regions after applying multiple corrections (Table S1), indicating that there was no significant atrophy in CN Aβ+ individuals.

To examine changes in WM metric within the Aβ+ group, we compared FA, MD, axial diffusivity, and RD of fiber tracts between the Aβ+ and Aβ− groups (Tables S2–S5). Individuals in the Aβ+ group exhibited significantly increased MD across eight fiber tracts (t = 2.52 to 3.50, Cohen's d = 0.38–0.57, p FDR < 0.05, Figure 2). These tracts included anterior thalamic radiation (ATR), fronto‐pontine tract (FPT), superior thalamic radiation (STR), thalamo‐premotor tract (T‐PREM), thalamo‐parietal tract (T‐PAR), striato‐premotor tract (ST‐PREM), corpus callosum rostral body (CC‐III), and corpus callosum anterior midbody (CC‐IV). Additionally, RD was significantly elevated in these fiber tracts, as well as in the corticospinal tract (CST), inferior occipito‐frontal fasciculus (IFO), optic radiation (OR), parieto‐occipital pontine tract (POPT), superior longitudinal fasciculus I (SLF‐I), and thalamo‐occipital tract (T‐OCC)(t = 2.30 to 4.22, Cohen's d = 0.28 to 0.57, p FDR < 0.05, Figure S2). These results were adjusted for age, sex, APOE ɛ4 status, and education. No significant differences in FA and axial diffusivity were observed among all fiber tracts. However, significant increases in MD and RD were noted in the ATR, FPT, STR, T‐PREM, T‐PAR, ST‐PREM, CC‐III, and CC‐IV in the CN Aβ+ group. To avoid collinearity and prioritize MD due to its stronger association with AD pathology, these eight tracts were identified as impaired fiber tracts. Subsequent analyses focused exclusively on MD for these tracts.

FIGURE 2.

FIGURE 2

Open in a new tab

Comparative analysis of WM integrity between groups. (A) The WM fiber tracts analyzed are color‐coded: ATR in blue, FPT in green, STR in light blue, T‐PREM in pink, T‐PAR in purple, ST‐PREM in orange, CC‐III in yellow, and CC‐IV in dark blue. (B) Combined box and violin plots depict MD values in Aβ− (orange) and Aβ+ (purple) groups across each tract. MD were significantly higher in Aβ+ group compared to Aβ− group. Only MD and significant results are shown. *p FDR < 0.05, **p FDR < 0.01, ***p FDR < 0.001. Values are residuals, adjusted for age, sex, APOE ɛ4 status, and education. Aβ, amyloid beta; APOE, apolipoprotein E; ATR, anterior thalamic radiation; CC‐III, corpus callosum rostral body; CC‐IV, corpus callosum anterior midbody; FDR, false discovery rate; FPT, fronto‐pontine tract; STR, superior thalamic radiation; T‐PREM, thalamo‐premotor tract; T‐PAR, thalamo‐parietal tract; ST‐PREM, striato‐premotor tract; WM, white matter.

3.4. WM metric association with age and global Aβ SUVR

Higher MD in the impaired fiber tracts, including ATR, FPT, STR, T‐PREM, T‐PAR, and ST‐PREM, was significantly associated with elevated global Aβ SUVR, while controlling for age, sex, education, and APOE ɛ4 status as covariates (r = 0.13 to 0.23, p FDR < 0.05) (Figure 3A). Furthermore, higher MD in all impaired fiber tracts was significantly correlated with increased age, after adjusting for sex, education, and APOE ɛ4 status as covariates (r = 0.27 to 0.46, p FDR < 0.001) (Figure 3B). In summary, these results suggest that Aβ burden is related to WM damage, which is associated with Aβ levels and significantly influenced by age.

FIGURE 3.

FIGURE 3

Open in a new tab

Association of MD with age and global SUVR in impaired fiber tracts. (A) Increasing global SUVR is associated with increased MD in impaired fiber tracts, including age, sex, education, and APOE ɛ4 status as covariates. (B) Increasing age is associated with increased MD in impaired fiber tracts, including sex, education, and APOE ɛ4 status as covariates. Partial correlation coefficient (r) and p value adjusted by FDR (p FDR) are provided for each tract, with Aβ− represented by orange dots and Aβ+ by purple dots. Solid lines reflect fitted regressions. Shading reflects 95% confidence interval. Only impaired fiber tracts are shown. Values are residuals, adjusted for age (only for [A]), sex, APOE ɛ4 status, and education. Aβ, amyloid beta; APOE, apolipoprotein E; FDR, false discovery rate; MD, mean diffusivity; SUVR, standardized uptake value ratio.

3.5. Association between WM metrics and Aβ SUVR in connected regions

In the damaged fiber tracts, we conducted partial correlation analyses to investigate the relationship between regional Aβ SUVR in the connected cortical and subcortical regions and fiber tract damage (Table S6 and Figure 4). Higher MD in the damaged fiber tracts was significantly associated with increased Aβ SUVR in the connected cortical regions, after controlling for sex, education, and APOE ɛ4 status as covariates (r = 0.137 to 0.276, p FDR < 0.05). However, no significant association was observed between MD in the damaged fiber tracts of the ATR, FPT, STR, T‐PREM, T‐PAR, ST‐PREM, and Aβ SUVR in the connected subcortical regions, with the exception of the CC‐III and CC‐IV fiber tracts. Furthermore, MD in CC‐III (r = −0.248, p FDR < 0.001) and CC‐IV (r = −0.220, p FDR < 0.001) exhibited a negative association with Aβ SUVR in the connected subcortical regions. By comparing the absolute partial correlation coefficients between MD and regional Aβ SUVR in connected cortical and subcortical regions across various fiber tracts, we found that the correlation between MD and connected cortical Aβ SUVR was significantly higher than the correlation between MD and connected subcortical Aβ SUVR (Figure S3, t = 4.801, p < 0.001). Moreover, the associations between WM tract and its anatomically connected cortical regions were significantly stronger than the associations between the tract and non‐connecting cortical regions (Figure S4, W Wilcoxon = 2010, p = 0.025). The primary results remained largely consistent after adjustment for cardiovascular risk factors (Figure S5).

FIGURE 4.

FIGURE 4

Open in a new tab

Association of MD of damaged fiber tracts and Aβ SUVR in connected cortical and subcortical regions. Increased SUVR in connected cortical regions is associated with increased MD in impaired fiber tracts, including age, sex, education, and APOE ɛ4 status as covariates. No significant association is found between MD in damaged fiber tracts and Aβ SUVR in connected subcortical regions, except for CC‐III and CC‐IV fiber tracts. Partial correlation coefficient (r) are provided for each tract. Solid lines reflect fitted regressions. Shading reflects 95% confidence interval. Only impaired fiber tracts are shown. Values are residuals, adjusted for age, sex, APOE ɛ4 status, and education. *p FDR < 0.05, **pP FDR < 0.01, ***p FDR < 0.001. All partial correlation coefficients (r) and p‐value adjusted by the FDR (p FDR) for each tract can be found in Table S6. Aβ, amyloid beta; APOE, apolipoprotein E; CC‐IV, corpus callosum anterior midbody; MD, mean diffusivity; SUVR, standardized uptake value ratio.

3.6. WM integrity and transcriptomic signatures

We further identified the transcriptomic signatures associated with fiber tract damage resulting from Aβ burden, utilizing data from the Allen human dataset. In the first component of PLS (PLS1), we identified 509 genes that were closely linked to fiber tract damage (Figure 5A, p FDR < 0.001). To investigate whether molecular biological signatures influence fiber tract damage, we used Metascape to compare Gene Ontology (GO) biological processes and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways with the PLS1 gene set. The PLS1 gene set was significantly enriched in multiple pathways, including sensory organ development, extraembryonic membrane development, myelination, regulation of cellular ketone metabolism, neuron projection development, regulation of establishment of protein localization, monocarboxylic acid metabolism, glial cell fate commitment, phospholipid biosynthesis, cellular export, and brain development (Figure 5B, p FDR < 0.0001).

FIGURE 5.

FIGURE 5

Open in a new tab

Transcriptomic signatures and cognitive traits of WM integrity. (A) The distribution of PLS1 genes in the brain is depicted as a word cloud, highlighting 509 genes with significant expression (p FDR < 0.0001). (B) Functional enrichment analysis of PLS1 genes (two‐tailed hypergeometric test; p FDR  <  0.0001). Only significant results are shown. (C) Cell type specificity of WM integrity (two‐tailed permutation test; ***p FDR  <  0.0001). (D) Neurosynth‐based decoding of WM integrity. FDR, false discovery rate; WM, white matter.

Finally, we investigated the overlap between the PLS1 gene set and gene sets from seven typical cell types. Genes associated with oligodendrocytes were significantly enriched (p FDR < 0.0001), whereas no significant enrichment was observed in the other six cell types (Figure 5C, Table S7).

4. DISCUSSION

We investigated Aβ‐mediated WM disruption in CN individuals and its transcriptomic correlates. The key findings are as follows: (1) CN Aβ+ individuals showed cortical Aβ deposition with WM damage but preserved gray matter; (2) the effect of Aβ on WM remained significant after age adjustment, though age‐dependent; (3) disrupted integrity specifically correlated with connected cortical (not subcortical) Aβ; and (4) genes displaying spatial expression patterns paralleling WM alterations were enriched in relevant biological processes and predominantly expressed in oligodendrocytes. These findings provide novel insights into Aβ‐induced WM disruption, potentially informing Aβ accumulation and damage mechanisms.

Consistent with Thal amyloid phases from autopsy studies, 28 our Aβ+ cohort exhibited neocortical (frontal/parietal/temporal) deposition pattern characteristic of phase 1, without subcortical involvement typically emerging in phase 2. Structural analyses revealed preserved gray matter volumes in Aβ+ individuals, aligning with AD biomarker trajectories where amyloid pathology precedes neurodegeneration. 4

Although our study did not find significant gray matter atrophy in CN Aβ+ individuals, we observed widespread damage to WM fibers. Aβ+ individuals exhibited increased MD and RD in multiple fiber tracts, which extends and confirms previous DWI studies that consistently reported abnormal integrity changes, such as widespread in MD in increases with MCI and AD. 8 , 9 Additionally, another study found that no significant changes in cortical MD when comparing CN Aβ+ individuals to Aβ− individuals. 29 This suggests that WM fiber integrity may be more sensitive to damage than gray matter and could serve as an earlier predictive indicator of AD.

Over the past few decades, patients with AD have shown both spatial and temporal changes in WM integrity, which manifest at both microstructural and microstructural levels. 30 Our findings indicate that even at the CN stage, individuals who are Aβ+ demonstrate damage in multiple fiber tracts. This damage may play a crucial role in the subsequent progression of AD and further accumulation of Aβ. Neuropathological studies have shown that these fiber tracts are associated with various functions and diseases. For instance, the integrity of the ATR and the STR has been compromised in conditions such as attention‐deficit/hyperactivity disorder 31 and post‐traumatic headache 32 and is linked to cognitive function. 33 Conversely, the FTP, 34 T‐PREM, 35 , 36 T‐PAR 37 , and ST‐PREM 35 have been reported to be potentially associated with motor function. Notably, CC‐III and CC‐IV connect the premotor and primary motor cortices, respectively. 38 This information may provide insights into the mechanisms underlying cognitive impairments in AD and offer an explanation for the cognitive improvements observed with physical exercise. 39 , 40

In contrast to findings from later stages of AD, our results indicate preserved axonal integrity but early alterations in myelination related diffusion metrics within the cognitively normal cohort. These microstructural changes appear closely linked to cortical Aβ deposition, with WM disruptions localized to connected cortical but not subcortical regions, suggesting cortical Aβ plays a primary role in initiating early neurodegeneration. Additionally, we observed an inverse pattern in the corpus callosum related to Aβ in subcortical areas, potentially reflecting non‐linear diffusion changes, compensatory mechanisms, or the influence of adjacent WM hyperintensities. Further methodological details and interpretations are provided in Supplementary Text 5.

This study explored the transcriptomic signatures that underlie the effects of Aβ deposition on WM integrity. The identification of PLS1 genes provides valuable insights into the molecular mechanisms associated with Aβ‐related WM damage. The enrichment of PLS1 genes in biological processes such as myelination, glial cell fate commitment, and neuron projection development suggests that Aβ deposition may disrupt critical pathways involved in maintaining WM integrity. Previous studies also confirmed the association of these biological processes with AD and cognitive function, 41 , 42 particularly myelination, 43 which aligns with our earlier hypothesis. Furthermore, the significant enrichment of PLS1 genes in oligodendrocytes underscores their potential role in mediating the effects of Aβ on WM. Oligodendrocytes are essential for the formation and maintenance of myelin, which is crucial for efficient axonal conduction and neural communication. A recent study demonstrated an increase in oligodendrocyte proliferation and axonal remyelination during the early stages of AD. 44 Consequently, oligodendrocyte dysfunction may hinder myelin repair and maintenance, exacerbating WM damage and promoting disease progression. Similarly, several studies 45 , 46 have highlighted the role of oligodendrocytes in the progression of AD, further validating our hypothesis that early Aβ deposition primarily causes myelin damage, with oligodendrocytes playing a critical role in this process. Further details are available in Supplementary Text 6.

Several limitations of this study should be acknowledged. First, the sample of Aβ+ individuals exhibited a notable sex imbalance, with over two‐thirds being male. This is inconsistent with epidemiological patterns showing that women are more frequently affected by AD. The overrepresentation of males in our sample may be due to recruitment bias, sex differences in disease progression, or other underlying biological factors. Second, as a cross‐sectional investigation, it cannot ascertain whether Aβ deposition is a consequence or a contributing factor to WM damage. Addressing this question is crucial for identifying early diagnostic biomarkers and elucidating the pathological mechanisms of AD. Third, our study currently encompasses older participants with an average age of 73.5 years. Future research should involve larger samples across diverse age groups to enhance the generalizability of these findings and employ longitudinal designs to monitor changes in WM integrity in relation to increasing Aβ burden. Furthermore, due to the influence of various biological processes such as inflammation, axonal swelling, and demyelination, the interpretation of some results may be challenging. Multicomponent relaxometry techniques, which are more sensitive to these processes, could be used to better assess them. These techniques have already been applied in cognitively healthy individuals and those with subjective cognitive decline. 47 , 48 Finally, integrating additional imaging modalities to assess functional connectivity and metabolic activity could facilitate exploration of the causal relationship between Aβ burden and WM integrity, particularly focusing on the underlying cellular mechanisms.

In summary, our study demonstrates that Aβ deposition is associated with early WM damage in CN individuals, characterized by specific spatial patterns and transcriptomic signatures that underlie these effects. These findings advance our understanding of the initial impact of Aβ on WM and open new avenues for investigating the pathogenesis and potential therapeutic targets for AD.

CONFLICT OF INTEREST STATEMENT

The authors have no relevant financial or non‐financial interests to disclose.

CONSENT STATEMENT

All participants provided written informed consent under protocols approved by ADNI‐affiliated Institutional Review Boards. De‐identified data were obtained from the ADNI database in accordance with the Declaration of Helsinki and ADNI ethical guidelines.

Supporting information

Supporting Information

DAD2-17-e70192-s001.docx (14.1MB, docx)

Supporting Information

DAD2-17-e70192-s002.pdf (1.8MB, pdf)

ACKNOWLEDGMENTS

The authors would like to thank all the participants of the ADNI, as well as all the individuals who contributed to this study. Special thanks go to Kewei Chen for the critical review of the manuscript. Data collection and sharing for this project are based on the ADNI (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH‐12‐2‐0012). ADNI is funded by the National Institute on Aging and the National Institute of Biomedical Imaging and Bioengineering and through generous contributions from the following: AbbVie, Alzheimer's Association; Alzheimer's Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol‐Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann‐La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research provide funds to support ADNI clinical sites in Canada. Private‐sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer's Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California. This study was funded by National Key Research and Development Program of China (Grant No. 2023YFC3605400), Science and Technology Innovation 2030 Major Projects (Grant No. 2022ZD0211600), and Sponsored by Beijing Nova Program.

Li Z, Sun Y, Li T. et al; The impact of amyloid beta burden on white matter dysfunction and associated transcriptomic signatures in cognitively normal elderly individuals. Alzheimer's Dement. 2025;17:e70192. 10.1002/dad2.70192

Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (https://adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: https://adni.loni.usc.edu/wp‐content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf

Contributor Information

Yaojing Chen, Email: luckychen1989@gmail.com.

Zhanjun Zhang, Email: zhang_rzs@bnu.edu.cn.

DATA AVAILABILITY STATEMENT

The data used in this study are publicly available from the ADNI database (https://adni.loni.usc.edu) upon registration and adherence to the data use agreement. The whole‐brain gene expression maps supporting the findings of this study are available in the Allen Human Brain Atlas at http://human.brain‐map.org. The list of cell‐specific gene sets compiled from all available large‐scale single‐cell studies of the adult human cortex can be found at https://static‐content.springer.com/esm/art%3A10.1038%2Fs41467‐020‐17051‐5/MediaObjects/41467_2020_17051_MOESM8_ESM.xlsx. The source data underlying figures are provided as Supplementary Data. The code and data supporting the findings of this study can be obtained from the corresponding author upon reasonable request.

REFERENCES

  • 1. 2021 Alzheimer's disease facts and figures. Alzheimers Dement. 2021;17(3):327‐406. [DOI] [PubMed] [Google Scholar]
  • 2. Jansen WJ, Ossenkoppele R, Knol DL, et al. Prevalence of cerebral amyloid pathology in persons without dementia: a meta‐analysis. JAMA. 2015;313(19):1924‐1938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Bateman RJ, Xiong C, Benzinger TL, et al. Clinical and biomarker changes in dominantly inherited Alzheimer's disease. N Engl J Med. 2012;367(9):795‐804. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Jack CR Jr, Andrews JS, Beach TG, et al. Revised criteria for diagnosis and staging of Alzheimer's disease: alzheimer's association workgroup. Alzheimers Dement. 2024;20(8):5143‐5169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Clark CM, Pontecorvo MJ, Beach TG, et al. Cerebral PET with florbetapir compared with neuropathology at autopsy for detection of neuritic amyloid‐β plaques: a prospective cohort study. Lancet Neurol. 2012;11(8):669‐678. [DOI] [PubMed] [Google Scholar]
  • 6. Sachdev PS, Zhuang L, Braidy N, Wen W. Is Alzheimer's a disease of the white matter?. Curr Opin Psychiatry. 2013;26(3):244‐251. [DOI] [PubMed] [Google Scholar]
  • 7. Lee SH, Coutu JP, Wilkens P, Yendiki A, Rosas HD, Salat DH. Tract‐based analysis of white matter degeneration in Alzheimer's disease. Neuroscience. 2015;301:79‐89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Zhang X, Sun Y, Li W, et al. Characterization of white matter changes along fibers by automated fiber quantification in the early stages of Alzheimer's disease. Neuroimage Clin. 2019;22:101723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Dou X, Yao H, Feng F, et al. Characterizing white matter connectivity in Alzheimer's disease and mild cognitive impairment: an automated fiber quantification analysis with two independent datasets. Cortex. 2020;129:390‐405. [DOI] [PubMed] [Google Scholar]
  • 10. Shirzadi Z, Schultz SA, Yau WW, et al. Etiology of white matter hyperintensities in autosomal dominant and sporadic Alzheimer disease. JAMA Neurol. 2023;80(12):1353‐1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Wang Y‐J, Hu H, Yang Y‐X, et al. Regional amyloid accumulation and white matter integrity in cognitively normal individuals. Journal of Alzheimer's Disease. 2020;74(4):1261‐1270. [DOI] [PubMed] [Google Scholar]
  • 12.Pichet Binette A, Theaud G, Rheault F, et al. Bundle‐specific associations between white matter microstructure and Aβ and tau pathology in preclinical Alzheimer's disease. eLife. 2021;10:e62929. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Li Y, Zhou G, Peng J, et al. White matter dysfunction in Alzheimer's disease is associated with disease‐related transcriptomic signatures. Commun Biol. 2025;8(1):820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Cieslak M, Cook PA, He X, et al. QSIPrep: an integrative platform for preprocessing and reconstructing diffusion MRI data. Nature Methods. 2021;18(7):775‐778. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Wasserthal J, Neher P, Maier‐Hein KH. TractSeg—Fast and accurate white matter tract segmentation. Neuroimage. 2018;183:239‐253. [DOI] [PubMed] [Google Scholar]
  • 16. Desikan RS, Ségonne F, Fischl B, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage. 2006;31(3):968‐980. [DOI] [PubMed] [Google Scholar]
  • 17. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics. 2007;8(1):118‐127. [DOI] [PubMed] [Google Scholar]
  • 18. Fortin JP, Parker D, Tunç B, et al. Harmonization of multi‐site diffusion tensor imaging data. Neuroimage. 2017;161:149‐170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Wang WE, Chen R, Mayrand RP, et al. Association of longitudinal cognitive decline with diffusion MRI in Gray Matter, Amyloid, and Tau deposition. Neurobiology of Aging. 2023;121:166‐178. [DOI] [PubMed] [Google Scholar]
  • 20. Ji G‐J, Sun J, Hua Q, et al. White matter dysfunction in psychiatric disorders is associated with neurotransmitter and genetic profiles. Nature Mental Health. 2023;1(9):655‐666. [Google Scholar]
  • 21. Li J, Wu GR, Li B, et al. Transcriptomic and macroscopic architectures of intersubject functional variability in human brain white‐matter. Commun Biol. 2021;4(1):1417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Hawrylycz MJ, Lein ES, Guillozet‐Bongaarts AL, et al. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature. 2012;489(7416):391‐399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Krishnan A, Williams LJ, McIntosh AR, Abdi H. Partial Least Squares (PLS) methods for neuroimaging: a tutorial and review. Neuroimage. 2011;56(2):455‐475. [DOI] [PubMed] [Google Scholar]
  • 24. Zhou Y, Zhou B, Pache L, et al. Metascape provides a biologist‐oriented resource for the analysis of systems‐level datasets. Nat Commun. 2019;10(1):1523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Yeh F‐C. DSI Studio: an integrated tractography platform and fiber data hub for accelerating brain research. Nature Methods. 2025;22(8):1617‐1619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer International Publishing; 2016. [Google Scholar]
  • 27. Mowinckel AM, Vidal‐Piñeiro D. Visualization of brain statistics with R packages ggseg and ggseg3d. Advances in Methods and Practices in Psychological Science. 2020;3(4):466‐483. [Google Scholar]
  • 28. Thal DR, Rüb U, Orantes M, Braak H. Phases of A beta‐deposition in the human brain and its relevance for the development of AD. Neurology. 2002;58(12):1791‐1800. [DOI] [PubMed] [Google Scholar]
  • 29. Sun P, He Z, Li A, et al. Spatial and temporal patterns of cortical mean diffusivity in Alzheimer's disease and suspected non‐Alzheimer's disease pathophysiology. Alzheimers Dement. 2024;20(10):7048‐7061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Mak E, Gabel S, Mirette H, et al. Structural neuroimaging in preclinical dementia: from microstructural deficits and grey matter atrophy to macroscale connectomic changes. Ageing Res Rev. 2017;35:250‐264. [DOI] [PubMed] [Google Scholar]
  • 31. Chen M, van der Pal Z, Poirot MG, et al. Prediction of methylphenidate treatment response for ADHD using conventional and radiomics T1 and DTI features: secondary analysis of a randomized clinical trial. Neuroimage Clin. 2024;45:103707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Yang HC, Nguyen T, Naugle KM, White FA, Wu YC>. White matter microstructural changes in post‐traumatic headache: a diffusion tensor imaging (DTI) study. medRxiv. 2024. [Google Scholar]
  • 33. Kokubun K, Nemoto K, Yamakawa Y. Smartphone app for lifestyle improvement improves brain health and boosts the vitality and cognitive function of healthy middle‐aged adults. Brain Behav. 2024;14(5):e3500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Hyde C, Fuelscher I, Rosch KS, et al. Subtle motor signs in children with ADHD and their white matter correlates. Hum Brain Mapp. 2024;45(14):e70002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Zhou W, He J, Zhang C, Pan Y, Sang T, Qiu X. Fiber‐specific white matter alterations in Parkinson's disease patients with freezing of gait. Brain Res. 2023;1815:148440. [DOI] [PubMed] [Google Scholar]
  • 36. Schnitzler A, Timmermann L, Gross J. Physiological and pathological oscillatory networks in the human motor system. J Physiol Paris. 2006;99(1):3‐7. [DOI] [PubMed] [Google Scholar]
  • 37. Simmons CM, Moseley SC, Ogg JD, et al. A thalamo‐parietal cortex circuit is critical for place‐action coordination. Hippocampus. 2023;33(12):1252‐1266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Hofer S, Frahm J. Topography of the human corpus callosum revisited–comprehensive fiber tractography using diffusion tensor magnetic resonance imaging. Neuroimage. 2006;32(3):989‐994. [DOI] [PubMed] [Google Scholar]
  • 39. Gaitán JM, Moon HY, Stremlau M, et al. Effects of aerobic exercise training on systemic biomarkers and cognition in late middle‐aged adults at risk for Alzheimer's disease. Front Endocrinol (Lausanne). 2021;12:660181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Lam LCW, Chan WC, Kwok TCY, et al. Combined physical exercise‐working memory training on slowing down cognitive decline in elders with mild clinical Alzheimer disease: a randomised controlled study (abridged secondary publication). Hong Kong Med J. 2022;28(3):28‐30. Suppl 3. [PubMed] [Google Scholar]
  • 41. Bathini P, Brai E, Balin BJ, et al. Sensory dysfunction, microbial infections, and host responses in Alzheimer's disease. J Infect Dis. 2024;230(Supplement_2):S150‐s164. [DOI] [PubMed] [Google Scholar]
  • 42. Ai X, Cao Z, Ma Z, et al. Proteomic analysis reveals physiological activities of Aβ peptide for Alzheimer's disease. Int J Mol Sci. 2024;25(15):8336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Shao Y, Li F, Zou B, et al. Up‐regulation of myelin‐associated glycoprotein is associated with the ameliorating effect of omega‐3 polyunsaturated fatty acids on Alzheimer's disease progression in APP‐PS1 transgenic mice. Food Funct. 2024;15(22):11236‐11251. [DOI] [PubMed] [Google Scholar]
  • 44. Mathys H, Davila‐Velderrain J, Peng Z, et al. Single‐cell transcriptomic analysis of Alzheimer's disease. Nature. 2019;570(7761):332‐337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Ishii A, Pathoulas JA, MoustafaFathy Omar O, et al. Contribution of amyloid deposition from oligodendrocytes in a mouse model of Alzheimer's disease. Mol Neurodegener. 2024;19(1):83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Rossi SL, Bovenkamp DE. Are oligodendrocytes the missing link in Alzheimer's disease and related dementia research?. Mol Neurodegener. 2024;19(1):84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Canales‐Rodríguez EJ, Alonso‐Lana S, Verdolini N, et al. Age‐ and gender‐related differences in brain tissue microstructure revealed by multi‐component T(2) relaxometry. Neurobiol Aging. 2021;106:68‐79. [DOI] [PubMed] [Google Scholar]
  • 48. Rivas‐Fernández M, Bouhrara M, Canales‐Rodríguez EJ, et al. Brain microstructure alterations in subjective cognitive decline: a multi‐component T2 relaxometry study. Brain Commun. 2025;7(1):fcaf017. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

DAD2-17-e70192-s001.docx (14.1MB, docx)

Supporting Information

DAD2-17-e70192-s002.pdf (1.8MB, pdf)

Data Availability Statement

The data used in this study are publicly available from the ADNI database (https://adni.loni.usc.edu) upon registration and adherence to the data use agreement. The whole‐brain gene expression maps supporting the findings of this study are available in the Allen Human Brain Atlas at http://human.brain‐map.org. The list of cell‐specific gene sets compiled from all available large‐scale single‐cell studies of the adult human cortex can be found at https://static‐content.springer.com/esm/art%3A10.1038%2Fs41467‐020‐17051‐5/MediaObjects/41467_2020_17051_MOESM8_ESM.xlsx. The source data underlying figures are provided as Supplementary Data. The code and data supporting the findings of this study can be obtained from the corresponding author upon reasonable request.

Articles from Alzheimer's & Dementia : Diagnosis, Assessment & Disease Monitoring are provided here courtesy of Wiley

RESOURCES