Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 15.
Published before final editing as: Neurobiol Aging. 2026 Sep 21;169:100–109. doi: 10.1016/j.neurobiolaging.2026.09.009

Plasma inflammatory markers and brain white matter microstructure in late middle-aged and older adults

Siona Mishra a, Corinne Pettigrew b, Chidi Ugonna c, Nan-kuei Chen c, Jennifer B Frye d, Kristian P Doyle d, Lee Ryan e, Marilyn Albert b, Sara Grace Ho b, Abhay Moghekar b, Anja Soldan b, Elizabeth R Paitel b
PMCID: PMC13664638  NIHMSID: NIHMS2213021  PMID: 42777342

Abstract

Chronic inflammation is a common feature of aging and is observed across various age-related neurodegenerative diseases, including Alzheimer’s disease (AD). Diffusion magnetic resonance imaging (dMRI) metrics of white matter microstructure are hypothesized to be sensitive to inflammation, though few studies have assessed this relationship in non-clinical populations. This cross-sectional study of 457 dementia-free older adults (Mage=63.82) examined associations of plasma markers related to inflammation with white matter microstructure metrics from multi-shell dMRI Neurite Orientation Dispersion and Density Imaging (NODDI) and conventional, single-shell dMRI metrics. Plasma inflammatory markers included high-sensitivity C-reactive protein (CRP), a composite of pro-inflammatory cytokines, and glial fibrillary acidic protein (GFAP), a measure of astrocytic activation. Higher cytokine composite levels were associated with higher cerebral free water fraction (FWF) and higher cerebellar peduncle orientation dispersion index (ODI). Higher CRP was associated with higher cerebral and cerebellar ODI, which was non-significant after adjusting for additional demographics and vascular risk factors. Higher GFAP was associated with lower cerebral FWF and lower cerebral and cerebellar neurite density index (NDI). The only significant association with conventional DTI metrics was between higher cytokine composite levels and higher cerebral mean diffusivity (MD). Results were consistent after accounting for plasma AD biomarkers. Thus, higher levels of peripheral pro-inflammatory markers are associated with white matter microstructure (FWF, ODI, NDI, and MD), supporting the sensitivity of these dMRI-based metrics to inflammatory processes. Additionally, the sensitivity of dMRI-based measures to inflammation may differ by inflammatory marker types, with multi-shell approaches like NODDI providing information beyond conventional, single-shell dMRI.

Keywords: Aging, DTI, cytokines, CRP, GFAP, inflammation, NODDI

1.1. Introduction

Chronic neuroinflammation is regularly observed in aging (Baylis et al., 2013; Godbout & Johnson, 2009) and is a common feature of various age-related neurodegenerative diseases, including Alzheimer’s disease (AD; Chen et al., 2016; Heneka et al., 2015). Neuroinflammation is an inflammatory response in the brain that is largely mediated by glial cells (i.e., microglia, which are the principal immune cells in the central nervous system, and astrocytes). Activated microglia and astrocytes can release immune mediators, such as cytokines, which may serve protective immune responses in the short-term (for review see Lyman et al., 2014; Shabab et al., 2017). However, chronic neuroinflammation can interfere with neuronal signaling and damage axons, synapses, and myelin, with downstream consequences for white matter integrity (Oestreich & O’Sullivan, 2022; Rodrigue et al., 2019). The availability of measures of brain structure that reflect chronic inflammation would therefore be helpful in understanding brain and cognitive health among older adults.

While levels of systemic inflammation can be easily obtained from blood, measuring inflammation in the brain is difficult due to the highly selective blood-brain barrier (Galea, 2021). Efforts are therefore underway to identify non-contrast magnetic resonance imaging (MRI) measures of neuroinflammation. Although MRI cannot directly image microglia or astrocytes, several MRI metrics have been proposed to reflect brain tissue properties that are closely associated with neuroinflammation (Jelescu & Fieremans, 2023; Oestreich & O’Sullivan, 2022). One such approach is diffusion MRI (dMRI), which measures the diffusion of water in tissue and is used to investigate brain tissue microstructure.

It has been suggested that some dMRI-based metrics are sensitive to neuroinflammation; however, there is relatively little evidence directly linking central or peripheral inflammatory markers to measures of brain tissue microstructure in non-clinical populations (for review see Jelescu & Fieremans, 2023; Oestreich & O’Sullivan, 2022). Two studies among dementia-free older adults using conventional, single-shell diffusion tensor imaging (DTI) did not find significant relationships between levels of pro-inflammatory cytokines in blood (i.e., interleukin 6 [IL-6] and IL-1β) with whole-brain white matter fractional anisotropy (FA; Gu et al., 2017) or mean diffusivity (MD; Boots et al., 2020). In a large sample that was not evaluated for cognitive impairment, higher baseline levels of plasma glial fibrillary acidic factor (GFAP), a marker of astrocytic activation often associated with neuroinflammatory processes, were associated with a longitudinal increase in MD (Huang et al., 2025). A number of studies have examined C-reactive protein (CRP), which is primarily produced by the liver in response to inflammation (Du Clos, 2000; Sproston & Ashworth, 2018). While largely reflective of peripheral inflammation, high levels of CRP may disrupt the blood-brain barrier, thereby contributing to neuroinflammation (Alilou et al., 2024; Hsuchou et al., 2012). Studies among older adults without dementia showed that higher CRP levels or a composite of CRP and TNF-α (Arfanakis et al., 2013), a pro-inflammatory cytokine, were associated with lower white matter FA and/or higher MD (Alilou et al., 2024; Boots et al., 2020; Conole et al., 2021; Miralbell et al., 2012; O’Donovan et al., 2021; Walker et al., 2017; Walker et al., 2018; Wersching et al., 2010). Taken together, these prior studies suggest associations between higher levels of CRP and conventional DTI-based metrics of white matter microstructure (e.g., FA and MD) among dementia-free older adults, though studies examining associations between levels of pro-inflammatory cytokines or GFAP with these metrics are scarce.

Importantly, there is a dearth of research examining associations between markers related to inflammation and more advanced diffusion-based measures from neurite orientation dispersion and density imaging (NODDI). NODDI provides more details about microstructural tissue properties by separating the signal in each voxel into three compartments: within neurites (i.e., axons and dendrites), between neurites, and extracellular, unrestricted (i.e., “free”) water, such as interstitial fluid and cerebrospinal fluid (CSF; Zhang et al., 2012). Animal studies suggest that higher free water fraction (FWF), a NODDI measure reflecting elevated extracellular “free” water, is associated with higher GFAP levels (Garcia-Hernandez et al., 2022; Yi et al., 2019), while higher orientation density index (ODI), a NODDI measure reflecting variability in neurite orientation, is associated with microglial density (Kim et al., 2023; Yi et al., 2019). A recent UK Biobank study found that longitudinal increases in plasma GFAP were associated with decreases in NODDI neurite density index (NDI), a measure of how densely neurites are packed, but not with consistent changes in extracellular volume fraction or orientation dispersion (Huang et al., 2025).

Two previous studies using a bi-tensor model of free water (vs. three-compartment with NODDI) among individuals with AD dementia and other neurodegenerative diseases, reported a positive association between plasma GFAP and FWF (Kim et al., 2025; Sumra et al., 2025). It is unclear if similar associations exist among dementia-free individuals, as the above-mentioned plasma inflammatory markers and white matter NODDI metrics have not been examined within the context of aging.

To address these gaps, the current study examined the cross-sectional associations of multiple plasma measures related to inflammation (a composite of pro-inflammatory cytokines, GFAP, and CRP) with NODDI (FWF, ODI, NDI) and conventional DTI (MD, FA) measures of white matter microstructure in a large sample (N=457) of middle-aged and older dementia-free adults. Global white matter microstructure was evaluated in both the cerebrum and the cerebellar peduncles, as reports suggest that cerebellar microglia may have specialized immunogenic properties (Dukhinova et al., 2024; Stoessel & Majewska, 2021; Stowell et al., 2018). In exploratory analyses, we also examined whether these associations differed by levels of AD pathology (i.e., amyloid and tau protein levels in blood plasma), given that AD pathology accumulation is associated with neuroinflammation (Kamila et al., 2025). We hypothesized that higher levels of inflammatory markers would be associated with higher FWF and ODI and lower NDI in the global cerebral and cerebellar tracts and that these associations may be stronger in individuals with more abnormal AD biomarker levels. We expected the NODDI metrics to have stronger associations with inflammatory markers than the conventional DTI metrics (MD, FA).

2.1. Methods

2.1.1. Study Design and Participant Recruitment

The Healthy Minds for Life (HML) study of the Precision Aging® Network (PAN; https://precisionagingnetwork.org/) was designed to investigate factors that predict individual brain health risks with the ultimate goal to optimize brain health across the lifespan and extend the cognitive healthspan (Ryan et al., 2019). This multi-site study includes four sites: the University of Arizona (Tucson, AZ), the University of Miami (Miami, FL), Emory University (Atlanta, GA), and the Johns Hopkins University (Baltimore, MD). Participants met the following inclusion criteria at study entry: 50-79 years of age; self-identify as White, Black, or Hispanic race/ethnicity; English language proficiency; no reported diagnosis of memory loss or dementia; no reported history of psychotic illness including schizophrenia or bipolar disorder; and no contraindications for blood draw or brain MRI. Enrollment for this study began in 2022 and is ongoing. Participants are recruited through multiple approaches, including online methods, in-person recruitment events, and referrals from site-specific research registries. The current analyses included N = 457 late middle-aged to older adult participants with diffusion MRI and blood inflammatory marker data. This study was approved by the WIRB-Copernicus Group Institutional Review Board and all participants provided written informed consent.

2.1.2. Magnetic Resonance Imaging Acquisition

MRI scans were acquired on a 3T Philips Achieva scanner (Eindhoven, The Netherlands) at Johns Hopkins University and on 3T Siemens scanners (Munich, Germany) at the other three sites: Skyra at the University of Arizona, Vida at the University of Miami, and Prisma at Emory University. The multi-modal protocol included a magnetization-prepared rapid gradient echo (MPRAGE) scan used for anatomical reference and image registration (Philips TR = 6500 ms, TE = 2.943ms, FA = 9 degree, resolution 1mm x 1mm x 1mm, size 256 x 256 x 211; Siemens TR = 2300ms, TE =2.98ms, FA = 9 degree, resolution 1mm x 1mm x 1mm, size 240 x 256 x 208). Diffusion-weighted images were acquired using a multi-shell acquisition scheme (resolution 2mm x 2mm x 2mm, TE/TR = 91/3700ms, multiband factor = 3, Philips size 128 x 128 x 69, Siemens size 116 x 116 x 69) with b-values of 500, 1000 and 2000 s/mm2. A total of 137 volumes were collected, consisting of 13 non-diffusion-weighted (b0) images and 124 diffusion-weighted images evenly distributed across the three shells. In addition, b0 images with reversed phase-encoding polarity were acquired to estimate and correct susceptibility-induced distortions in the diffusion MRI data. Detailed scanner-specific acquisition parameters are reported in Supplemental Table 1.

2.1.3. Image Processing

The diffusion-weighted images were pre-processed using QSIPREP version 0.21.4 (Cieslak et al., 2021). Processing included joint eddy current correction, head motion correction and susceptibility distortion correction. Noise level estimation and denoising were conducted using Marchenko-Pastur PCA from MRtrix ecosystem (Tournier et al., 2019; Veraart et al., 2016). White matter bundle segmentation for each participant was performed on the pre-processed QSIPREP outputs using Tractseg (https://github.com/MIC-DKFZ/TractSeg) version 2.8. The qsirecon module of QSIPREP was used to fit a white-matter NODDI model (amiconoddi) using default NODDI parameters. For NODDI modeling of white matter, including cerebellar white matter, the intrinsic parallel diffusivity was fixed at the default value of 1.7 × 10−3 mm2/s. Additional masking was performed using white matter masks from FSL’s FAST algorithm and Freesurfer’s surface pipeline to obtain mean values of FWF, ODI, NDI, MD, and FA in global cerebral white matter, which were used to calculate the cerebral white matter measures used in this study. Additionally, tract-level FWF, ODI, and MD values were computed for 38 tracts, from which a cerebellar peduncle composite was derived as the average of bilateral inferior, middle, and superior cerebellar peduncle values.

The primary analyses focused on the three NODDI measures (FWF, ODI, NDI), given the increased specificity of multi-compartment modeling compared to single-shell DTI. Auxiliary analyses with conventional, single shell metrics (MD, FA) are included for comparison with the existing literature, as these have been proposed to reflect, at least partially, processes related to neuroinflammation, as described above. MD reflects the average diffusivity of water in all directions (Pierpaoli et al., 1996), providing a measure of overall tissue density and microstructure. FA provides an index of uniformity in the directionality of water diffusion, with larger values indicating that movement is more restricted to one direction. Though their biophysical specificity is limited (Jones et al., 2013), lower MD and higher FA values are generally interpreted as reflecting ‘better’ microstructural integrity. From NODDI, higher FWF reflects elevated extracellular “free” water (Kim et al., 2025; Sumra et al., 2025). ODI quantifies variability in neurite orientation, with higher values generally reflecting less coherent organization in the direction of neurites. NDI reflects the density of neurites (axons and dendrites), with lower values reflecting less densely packed, lower volume, or looser clustering of axons and dendrites within brain tissue. Analyses focused on NODDI and DTI values for the global cerebral white matter as well as a cerebellar peduncle composite. All diffusion metric values were z-scored within study site prior to averaging.

2.1.4. Plasma Inflammatory, GFAP, and AD Biomarker Measures

Plasma was collected via EDTA tubes after an overnight fast; samples were centrifuged at 1500g for 15 minutes at 4°C and stored at −80°C in polypropylene vials prior to being shipped on dry ice to the coordinating site (University of Arizona) for processing. Cytokine concentrations were quantified using the MILLIPLEX® MAP Human Cytokine Magnetic Bead Panel (Millipore, Cat# HYCTA60K) according to the manufacturer’s instructions. Plasma samples were thawed on ice and undiluted samples were centrifuged at 10,000 × g for ten minutes to remove debris. Standards and quality controls were reconstituted and serially diluted as specified by the manufacturer. The assay was performed on a Luminex® LX200 instrument, equipped with xPONENT® software. A minimum of 100 beads per analyte were collected per well. Standard curves were generated using a 5-parameter logistic (5-PL) curve-fitting algorithm using Millipore’s Belysa Analysis Software. Cytokine concentrations were calculated by interpolation from the standard curves and expressed in pg/mL. Samples with values below the lower limit of detection (LLOD) were reported as the minimum detectable concentration reported by Millipore divided by two.

Inflammatory markers of interest included: IL-1α, IL-1β, IL-2, IL-6, IL-8, TNF-α, and TNF-β. Given the lack of a priori hypotheses regarding specific cytokines, markers were z-score transformed and then averaged to create a pro-inflammatory composite score. Preliminary analyses showed that these measures tended to show weak positive correlations with one another, supporting the approach of combining them into a single measure. While less specific than the individual markers, this pro-inflammatory cytokine composite may be a more sensitive measure of overall inflammation. High-sensitivity CRP was measured by Sonora Quest Laboratories using a chemiluminescent immunoassay. CRP was analyzed separate from the pro-inflammatory cytokine composite given biological differences in the production, time course, and function. CRP values were log-transformed to reduce skewness and then z-score normalized.

Plasma samples were analyzed for GFAP, p-tau181, and Aβ42 using the Single Molecule Array (Simoa) Neurology 4-Plex E, with p-tau181 assays on the HD-X instrument (Quanterix Corporation). Quanterix Simoa is an ultrasensitive assay that uses monoclonal antibodies and calibrators to allow for detection at very low concentrations. All assays were run in duplicate, and the mean value of the duplicate measurements was used for all analyses. Intra-assay CVs for all analytes were less than 10% and inter-assays CVs for all analytes were less than 20% for a native plasma sample run across all plates in addition to high and low kit controls supplied by the manufacturer. GFAP scores were log-transformed to reduce skewness and then z-scored. AD biomarker levels were quantified using the ratio of p-tau181/Aβ42 (Fowler et al., 2022), which was z-scored prior to analysis. Biomarker values were z-scored across all sites. The percentage of participants with abnormal levels of AD biomarkers (11% for Aβ42/Aβ40, 4.49% for p-tau181) were determined based on Aβ clinical cut-points (Quaresima et al., 2024) and company reported p-tau norms.

2.1.5. Vascular Risk Summary Score

A summary score reflecting vascular risk factors was computed from medical history interview and clinical assessments, reflecting the sum of four dichotomous vascular risk factors (Gottesman et al., 2017): self-reported hypertension, high cholesterol (based on total cholesterol ≥200 mg/dL), self-reported diabetes, and smoking (self-reported current or past use). Participants were categorized into three groups representing 0, 1, or ≥2 risk factors. Body mass index (BMI) was not included in the vascular risk score and was instead analyzed as a separate, continuous variable, given potential associations with plasma biomarker values (Lee et al., 2025).

2.1.6. Cognitive Assessment

Participants completed a short cognitive battery during their visit that included the Montreal Cognitive Assessment (MoCA), a screener providing a measure of global cognitive performance (Nasreddine et al., 2005). A previously validated MoCA cut-point of 24 was applied for sensitivity analyses to examine relationships among participants without cognitive impairment (Malek-Ahmadi & Nikkhahmanesh, 2024). A meta-analysis that evaluated the MoCA for detection of mild cognitive impairment demonstrated that a cut-point of 24 has good sensitivity (80%, range = 67–88%) and specificity 84% (range 75–90%; Islam et al., 2023). Values for dementia are reported to be even higher, with a very large, nation-wide study showing sensitivity 95.3% and specificity 92.67% (Ismail et al., 2025). Additional cognitive testing and surveys (not included in these analyses) were completed in-person as well as online via the MindCrowd.org website.

2.1.7. Statistical Analysis

Multiple linear regression models were used to examine the cross-sectional associations of measures related to inflammation (i.e., CRP, pro-inflammatory cytokine composite, and GFAP, the primary predictors of interest, evaluated in separate models) with the global cerebral and cerebellar peduncle white matter microstructure measures as outcomes. Separate models were run for each of the diffusion measures (FWF, ODI, and NDI from NODDI, as well as MD and FA). All models covaried age, sex, and study site. A set of exploratory follow-up models examined whether the observed associations between inflammatory levels and diffusion measures differed by AD biomarker levels. This was done by adding terms for p-tau181/Aβ42 and the inflammation*p-tau181/Aβ42 interaction to the primary models, as well as BMI, given reports that plasma AD biomarker levels are influenced by BMI (Lee et al., 2025). To address concerns regarding multiple comparisons, a False Discovery Rate (FDR) correction was applied for all analyses, adjusting for six comparisons (i.e., 2 ROIs [cerebral, cerebellar] x 3 NODDI metrics [FWF, ODI, NDI]) separately for the three plasma inflammatory measures, given the biological differences in the production, time course, and function of the different inflammatory markers (Benjamini & Hochberg, 1995). The same approach was applied to the DTI metrics, adjusting for 4 comparisons (i.e., 2 ROIs [cerebral, cerebellar] x 2 DTI metrics [MD, FA]).

A sensitivity analysis examined whether results changed when covarying BMI and a summary vascular risk score, given vascular health may influence both inflammatory markers and white matter microstructure (e.g., Jacobs et al., 2013; Kennedy & Raz, 2009; Moyse et al., 2022; Williams et al., 2019). A second sensitivity analysis covaried education level, ethnicity, and race, given previous associations between these variables with inflammation and white matter microstructure (e.g., Boots et al., 2020; Friedman & Herd, 2010; Fuller-Rowell et al., 2015). A third sensitivity analysis examined whether the results from primary models changed when excluding participants with a MoCA score below 24, which may be indicative of cognitive impairment (Malek-Ahmadi & Nikkhahmanesh, 2024). A final sensitivity analysis was conducted to assess robustness to effects of scanner vendor (Phillips at JHU vs. Siemens at the other sites) by excluding JHU from the analysis and covarying site, which additionally accounts for differences in scanner type (Skyra, Vida, and Prisma).

3.1. Results

3.1.1. Descriptive statistics

Descriptive statistics for the 457 participants included in the analyses (Atlanta n = 127, Baltimore n = 92, Miami n = 108, Tucson n = 130) are shown in Table 1. On average, participants were 63.8 years old, predominantly female (69%), and 37% from racial or ethnic groups traditionally underrepresented in research. The two systemic inflammatory measures were significantly correlated with one other (CRP-cytokines: r = 0.15, p < 0.01) but not with GFAP (CRP-GFAP: r = 0.04, p = 0.44, cytokines-GFAP: r = 0.08, p = 0.07). Nineteen participants out of the N = 457 included participants were missing CRP data, resulting in n = 438 in analyses with CRP. The individual cytokines within the pro-inflammatory cytokine composite were significantly correlated with each other (rs = 0.15 to 0.73, all ps < 0.01, see Supplemental Table 2), with good internal consistency (Cronbach’s α = 0.83). The correlations among dMRI metrics ranged from non-significant to moderate (r = 0.02 to 0.68, p = 0.66 to < 0.001, see Supplemental Table 3). Older age was associated with significantly higher GFAP levels (r = 0.52, p < 0.001), after accounting for sex and study site, but was not significantly associated with the cytokine composite or CRP levels (ps > 0.43).

Table 1.

Descriptive statistics for total sample.

N=457
Age 63.82 (7.63)
Sex, n (%) female 314 (68.7%)
Ethnicity (% Hispanic/Latino) 76 (17%)
Race
Black or African American, n (%) 91, 19.91%
White, n (%) 356, 77.90%
Other/muti-racial, n (%) 10, 2.19%
Education level
  High School diploma, n (%) 31 (6.8%)
  Some college or less, n (%) 89 (19.5%)
  Undergraduate degree, n (%) 173 (37.9%)
  Post-graduate degree, n (%) 164 (35.9%)
MoCA 26.31 (2.64)
  MoCA < 24, n (%)  71 (16%)
CRP 2.54 (4.37)
Cytokine composite 0.00 (0.71)
GFAP 135.00 (76.70)
p-tau181/Aβ42 3.81 (2.10)
BMI 28.00 (6.00)
Vascular risk summary score [/4 max] 1.11 (0.79)
Open in a new tab

Note: MoCA = Montreal Cognitive Assessment total score. Numbers indicate mean (SD), unless otherwise noted.

3.1.2. Associations of inflammatory markers with white matter microstructure measures

In the primary NODDI models, higher cytokine composite levels were associated with higher FWF in cerebral white matter (p < 0.05) and higher ODI in the cerebellar peduncle composite (p = 0.001; see Table 2, Figure 1). A positive association with cerebral ODI did not remain significant following correction for multiple comparisons. Higher CRP levels were associated with higher cerebral (p < 0.05) and cerebellar (p < 0.01) ODI. Higher GFAP levels were associated with lower cerebral FWF (p < 0.05) and both lower cerebral and cerebellar NDI (ps < 0.001; Figure 2). In models with conventional DTI metrics, only the association between higher cytokine composite levels and higher cerebral MD was significant (p < 0.05).

Table 2.

Results from multiple linear regression models assessing the associations of systemic inflammatory and GFAP levels with white matter microstructure, covarying age, sex, and study site.

NODDI
Cerebral Coeff β SE t p pFDR Cerebellar Coeff β SE t p p FDR
Cytokines Cytokines
FWF 0.158 0.112 0.065 2.418 0.016 0.048 FWF 0.014 0.012 0.058 0.245 0.807 0.807
ODI 0.140 0.099 0.066 2.131 0.034 0.068 ODI 0.193 0.151 0.060 3.241 0.001 0.006
NDI −0.026 −0.019 0.064 −0.413 0.680 0.807 NDI 0.039 0.029 0.062 0.625 0.532 0.798
CRP CRP
FWF −0.007 −0.007 0.048 −0.146 0.884 0.884 FWF 0.022 0.025 0.043 0.514 0.607 0.884
ODI 0.120 0.119 0.048 2.496 0.013 0.039 ODI 0.126 0.139 0.043 2.890 0.004 0.024
NDI −0.012 −0.012 0.046 −0.262 0.794 0.884 NDI 0.063 0.067 0.045 1.403 0.161 0.322
GFAP GFAP
FWF −0.127 −0.127 0.055 −2.287 0.023 0.046 FWF −0.096 −0.110 0.049 −1.957 0.051 0.066
ODI −0.089 −0.089 0.055 −1.602 0.110 0.110 ODI −0.098 −0.108 0.051 −1.924 0.055 0.066
NDI −0.231 −0.232 0.053 −4.401 <0.001 <0.001 NDI −0.218 −0.228 0.052 −4.211 <0.001 <0.001
DTI
Cerebral Coeff β SE t p pFDR Cerebellar Coeff β SE t p pFDR
Cytokines Cytokines
MD 0.155 0.110 0.062 2.508 0.012 0.050 MD 0.004 0.003 0.057 0.07 0.944 0.944
FA −0.097 −0.069 0.065 −1.494 0.136 0.195 FA −0.084 −0.068 0.058 −1.457 0.146 0.195
CRP CRP
MD 0.014 0.014 0.046 0.308 0.758 0.932 MD 0.004 0.005 0.041 0.105 0.917 0.932
FA 0.021 0.022 0.047 0.450 0.653 0.932 FA 0.004 0.004 0.042 0.085 0.932 0.932
GFAP GFAP
MD 0.001 0.090 0.052 1.708 0.088 0.173 MD 0.065 0.075 0.048 1.365 0.173 0.173
FA −0.086 −0.087 0.055 −1.566 0.118 0.173 FA −0.069 −0.079 0.049 −1.418 0.157 0.173
Open in a new tab

Note. Bold, un-corrected p < 0.05. pFDR = FDR-adjusted p value corrected for six comparisons (significant pFDR < 0.05). coeff = coefficient, FA = fractional anisotropy, FWF = free water fraction, MD = mean diffusivity, NDI = neurite density index, ODI = orientation dispersion index, SE = standard error.

Figure 1.

Figure 1.

Figure 1.

Open in a new tab

Higher levels of a pro-inflammatory cytokine composite (including IL-1α, IL-1β, IL-2, IL-6, IL-8, TNF-α, and TNF-β) are significantly associated with higher cerebral white matter free-water fraction (FWF, top left, β=0.112), orientation dispersion index (ODI, top right, β=0.099), mean diffusivity (MD, bottom left, β=0.110), and cerebellar peduncle ODI (bottom right, β=0.151). Diffusion MRI metrics are shown residualizing age, sex, and study site.

Figure 2.

Figure 2.

Open in a new tab

Higher plasma GFAP levels are significantly associated with lower cerebral (left, β=−0.232) and cerebellar (right, β=−0.228) neurite density index (NDI). Diffusion MRI metrics are shown residualizing age, sex, and study site.

For model covariates, older age was associated with higher cerebral FWF, higher cerebral ODI (in CRP model only), lower cerebral and cerebellar (cerebellar significant only in cytokine and CRP models) NDI, higher cerebral and cerebellar MD, and lower cerebral FA (ps < 0.05). For sex, cerebral ODI values were higher in men (ps < 0.01); men also had higher cerebellar ODI in the CRP model (ps < 0.05). Cerebellar NDI and FA values were higher in men (ps < 0.05). Study site was not a significant covariate in any model (ps > 0.524).

In the exploratory follow-up analyses including interactions with AD biomarker levels, there was one significant pro-inflammatory cytokine composite*p-tau181/Aβ42 interaction for cerebellar MD (p = 0.024), although this did not remain significant following FDR adjustment for multiple comparisons. All other interaction terms were non-significant (ps > 0.064, data not shown). When the non-significant interaction terms were removed, there were no significant associations between p-tau181/Aβ42 and white matter microstructure (ps > 0.064, data not shown).

In sensitivity analyses that additionally covaried BMI and the vascular risk summary score, the cytokine composite and GFAP results were comparable, although associations for CRP were no longer significant (Supplemental Table 4). In sensitivity analyses covarying education level, ethnicity, and race, the pattern of results was consistent with the primary models, though associations with cerebral ODI were no longer significant (Supplemental Table 5). For completeness, results of a fully adjusted model covarying age, sex, study site, education, ethnicity, race, BMI, and vascular risk are shown in Table 3. Sensitivity analyses also showed that the results of the primary models were consistent when excluding participants with a MoCA score below 24 (Supplemental Table 6). Supplemental Table 7 shows results using the individual cytokines, rather than the cytokine composite score. Overall, IL1α, IL6, and IL8 had the most robust and consistent associations with the white matter microstructure metrics. Analyses conducted within the sites with Siemens scanners demonstrated findings consistent with the primary analyses (Supplemental Table 8).

Table 3.

Results from fully adjusted multiple linear regression models assessing the associations of systemic inflammatory and GFAP levels with white matter microstructure, covarying age, sex, study site, education level, ethnicity, race, BMI, and vascular risk.

NODDI
Cerebral Coeff β SE t p pFDR Cerebellar Coeff β SE t p pFDR
Cytokines Cytokines
FWF 0.166 0.117 0.066 2.505 0.013 0.045 FWF 0.029 0.024 0.059 0.498 0.619 0.756
ODI 0.088 0.062 0.065 1.361 0.174 0.348 ODI 0.143 0.111 0.058 2.445 0.015 0.045
NDI −0.02 −0.014 0.064 −0.311 0.756 0.756 NDI 0.024 0.018 0.063 0.376 0.707 0.756
CRP CRP
FWF −0.01 −0.010 0.054 −0.185 0.853 0.988 FWF 0.059 0.067 0.049 1.213 0.226 0.988
ODI 0.001 0.001 0.053 0.015 0.988 0.988 ODI 0.025 0.028 0.048 0.524 0.601 0.988
NDI −0.008 −0.008 0.052 −0.161 0.872 0.988 NDI 0.022 0.023 0.051 0.429 0.668 0.988
GFAP GFAP
FWF −0.131 −0.131 0.056 −2.357 0.019 0.038 FWF −0.105 −0.120 0.049 −2.131 0.034 0.051
ODI −0.057 −0.057 0.054 −1.041 0.298 0.298 ODI −0.060 −0.067 0.049 −1.225 0.221 0.265
NDI −0.233 −0.234 0.053 −4.428 <0.001 <0.001 NDI −0.203 −0.212 0.052 −3.906 <0.001 <0.001
DTI
Cerebral Coeff β SE t p pFDR Cerebellar Coeff β SE t p pFDR
Cytokines Cytokines
MD 0.151 0.107 0.062 2.420 0.016 0.064 MD 0.021 0.017 0.058 0.365 0.715 0.715
FA −0.097 −0.069 0.065 −1.480 0.140 0.187 FA −0.087 −0.071 0.058 −1.491 0.137 0.187
CRP CRP
MD 0.001 0.001 0.052 0.023 0.982 0.982 MD 0.052 0.060 0.047 1.111 0.267 0.944
FA 0.032 0.032 0.053 0.600 0.549 0.944 FA −0.018 −0.021 0.048 −0.375 0.708 0.944
GFAP GFAP
MD 0.092 0.092 0.053 1.745 0.082 0.244 MD 0.053 0.060 0.048 1.089 0.277 0.277
FA −0.082 −0.083 0.055 −1.496 0.135 0.244 FA −0.066 −0.075 0.049 −1.334 0.183 0.244
Open in a new tab

Note. Bold, un-corrected p < 0.05. pFDR = FDR-adjusted p value corrected for six comparisons (significant pFDR < 0.05). coeff = coefficient, FA = fractional anisotropy, FWF = free water fraction, MD = mean diffusivity, NDI = neurite density index, ODI = orientation dispersion index, SE = standard error.

4.1. Discussion

The goal of this study was to examine cross-sectional associations between peripheral measures of inflammation and measures of white matter microstructure based on both NODDI and conventional dMRI. In late middle-aged and older adults, higher levels of a pro-inflammatory cytokine composite were associated with higher cerebral FWF and cerebellar ODI. The individual cytokines with the most robust and consistent associations with dMRI metrics were IL1α, IL6, and IL8. Higher CRP levels were associated with higher cerebral and cerebellar ODI, though these effects were notably not significant when accounting for vascular risk factors. Higher GFAP was associated with lower cerebral FWF and both lower cerebral and cerebellar NDI. In models with conventional DTI metrics, only the association between higher cytokine composite levels and higher cerebral MD was significant. These patterns were evident even after accounting for biomarkers of AD pathology, though the current sample had a low prevalence of clinically abnormal AD biomarker levels. Taken together, these findings suggest that systemic inflammation is linked to alterations in both cerebral and cerebellar white matter microstructure integrity, and that these diffusion measures partially reflect inflammatory processes.

Only the composite of pro-inflammatory cytokines was significantly associated with both NODDI (FWF, ODI) and conventional dMRI (MD) metrics of white matter microstructure. The few prior studies that examined plasma cytokine levels in relationship to dMRI measures of brain microstructure among dementia-free older adults used only single-shell models and did not find significant associations for the individual cytokines IL-6 or IL-1β with FA (Gu et al., 2017) or MD (Boots et al., 2020). In contrast, we found that higher cytokine composite levels were associated with higher cerebral MD. Our positive findings may be due to the use of a composite of various pro-inflammatory cytokines, which may provide greater sensitivity in assessing inflammation, as it captures a greater array of inflammatory responses. The individual cytokines with the most robust and consistent associations were IL-1α, IL-6, and IL-8. Thus, these may be important to include for analysis in future studies. The current study importantly expands on existing research with conventional DTI by showing that higher levels of the pro-inflammatory composite were associated with higher dMRI values based on NODDI. While MD represents the average of the diffusion signal in all directions, NODDI separates the signal in each voxel into three compartments, providing more specificity. The current findings demonstrated that higher cytokine levels were associated with higher FWF, reflecting elevated extra-cellular “free” water, in cerebral white matter and higher ODI, reflecting greater variability in neurite orientation, in cerebellar peduncle tracts. Thus, this newer multi-shell acquisition approach to the assessment of white matter microstructure may be particularly valuable for assessing relationships with inflammation.

The study also found that higher levels of CRP were associated with higher ODI in both cerebral white matter and the cerebellar peduncle composite. These findings expand upon previous studies that have primarily analyzed CRP in relationship to conventional, single-shell diffusion metrics (Boots et al., 2020; Conole et al., 2021; Miralbell et al., 2012; O’Donovan et al., 2021; Walker et al., 2017; Walker et al., 2018; Wersching et al., 2010), and suggest that higher CRP may be related to more dispersed or disorganized neurites among late middle-aged and older adults. However, these results did not remain significant when covarying vascular risk factors, suggesting that relationships between CRP and dMRI-based measures of brain microstructure may be confounded by levels of vascular risk. This finding underscores the importance of accounting for vascular risk in studies of inflammation and highlights an important area of future research – assessing potential mechanisms underlying the relationships between CRP, brain white matter microstructure, and vascular health. Additionally, unlike some previous longitudinal studies (Conole et al., 2021; O’Donovan et al., 2021; Walker et al., 2017; Walker et al., 2018), we did not find significant associations between CRP and MD. These differences may be attributable to the cross-sectional nature of the current study and sample characteristics, such as level of vascular risk, with most of the previous studies being comprised of participants with greater vascular and medical risk factors. Our results therefore highlight the importance of controlling for BMI and other vascular risk factors when investigating plasma markers related to inflammation and measures of brain microstructure. Future studies directly evaluating whether associations between DTI metrics and CRP differ by level of vascular risk, or are mediated by vascular risk, may provide valuable insight.

To our knowledge, only one previous study has examined relationships between plasma GFAP and white matter dMRI metrics in dementia-free participants. This previous study found that higher baseline GFAP was associated with a longitudinal increase in MD, and that longitudinal increases in GFAP were associated with lower NDI over time. There were no consistent associations with ODI or FWF (Huang et al., 2025). The present study found that higher GFAP was associated with lower cerebral and cerebellar NDI, indicative of a lower density of axons and dendrites, which is consistent with the longitudinal patterns reported by Huang and colleagues. The lack of significant associations with MD in the current analyses may be attributable to sample characteristics or the cross-sectional nature of the present study. In two studies of older participants across the AD spectrum, higher GFAP levels were associated with lower FA and higher MD (Bettcher et al., 2021; Kim et al., 2025), as well as higher free-water fractional volume using bi-tensor free water imaging (Kim et al., 2025). Thus, future longitudinal research is needed to determine whether associations between GFAP and dMRI are more robust in later stages of AD, which are characterized by higher levels of neurodegeneration and pathology.

The current study also found that higher GFAP was associated with lower cerebral FWF, which contrasts with the positive association between cerebral FWF and the cytokine composite. This pattern may be attributable to higher GFAP levels reflecting astrogliosis, a reactive overexpression of astrocytes, which is associated with both a loss of neurites, which would be reflected in lower NDI, and more restricted extracellular space, which would be evident in lower FWF (Leipp et al., 2024). The difference in patterns between GFAP and cytokines demonstrates the importance of examining multiple markers of systemic inflammation and astrocytic activation, which may capture different aspects of the inflammatory response. While cytokine levels are general, non-specific markers of inflammation, GFAP is thought to primarily reflect astrocytic activation, lending it more specificity to brain and central nervous system inflammation (Eng & Ghirnikar, 1994). However, GFAP levels measured in plasma (vs. cerebrospinal fluid) may reflect a mix of other neurodegenerative or inflammatory processes, particularly in the absence of blood-brain barrier damage (Youn et al., 2025). Future studies examining GFAP and cytokine levels in CSF are therefore needed to determine whether they are similarly related to microstructural properties.

Relative to conventional DTI, which revealed only one significant association, NODDI metrics had more robust and more consistent relationships with all three inflammatory markers, demonstrating the value of multi-compartment dMRI in analyzing relationships with inflammation. Higher values of FWF, ODI, and MD were associated with higher levels of a pro-inflammatory cytokine composite. In contrast, CRP and GFAP were each associated with metrics derived from the multi-shell acquisition, but not the standard metrics of MD or FA. Thus, NODDI metrics may provide more sensitivity for some markers related to inflammation and astrocytic activation. Yet, MD, which can be obtained using a conventional, single-shell acquisition protocol is still informative. This is consistent with suggestions that all of these metrics may reflect neuroinflammation to some degree, though these measures are not specific to inflammation and may reflect multiple biological processes including neurodegeneration and edema (Jelescu & Fieremans, 2023; Kim et al., 2025; Kim et al., 2023; Plank et al., 2025; Sumra et al., 2025; Yi et al., 2019). Future studies are needed to replicate these findings, and to further explore the mechanisms underlying the reported associations.

While all three of the NODDI metrics in cerebral white matter were associated with inflammatory markers, relationships between microstructure of the cerebellar peduncles and inflammatory markers were specific to ODI and NDI. These associations indicate that relationships between inflammation and cerebellar white matter microstructure are most evident with diffusion metrics that reflect characteristics of the NODDI tissue compartment, in contrast with FWF, which reflects extra-cellular “free” water. Specifically, higher pro-inflammatory cytokine levels were associated with higher ODI and higher GFAP was associated with higher NDI in a composite of the cerebellar peduncles, which are the white matter tracts that connect the cerebellum with the rest of the brain and spinal cord. Because the cerebellum is often omitted from neuroimaging research, this expands prior work. Although the cerebellum is important for cognitive functioning (Buckner, 2013; Schmahmann, 2019; Schmahmann et al., 2019; Stoodley, 2012) and changes with aging (Arleo et al., 2024; Bernard, 2022) and in various neurodegenerative diseases (Gellersen et al., 2021; Gellersen et al., 2017; Hoxha et al., 2018; Liu et al., 2024; Schmahmann, 2016), relatively little is known about markers of inflammation in the human cerebellum. Of possible relevance, higher ODI has also been suggested to correspond with microglial density in animal models (Kim et al., 2023; Yi et al., 2019). Animal models further suggest that activated cerebellar microglia release pro-inflammatory cytokines, including TNFα, IL-1β, and IL-6, and that the latter two cytokines mediate cerebellar neuronal activity (Dukhinova et al., 2024; Gyengesi et al., 2019; Yamamoto et al., 2019). Additionally, cerebellar microglia have unique features that have led some to consider them a separate subset compared to microglia in the cerebral cortex, including distinct morphology, distribution, functional roles, and molecular structure (Dukhinova et al., 2024; Stoessel & Majewska, 2021; Stowell et al., 2018).

It is noteworthy that the cerebellar associations with GFAP are consistent with those in cerebral white matter (i.e., lower FWF, lower NDI). In contrast, for the cytokine composite, there were relationships with both ODI and FWF in cerebral white matter (i.e., higher FWF, higher ODI) but only with ODI in the cerebellar peduncles. This relationship with cerebellar ODI but not FWF may suggest greater resilience to the effects of systemic inflammation in the cerebellum vs. cerebrum. Specifically, in early phases of tissue degeneration, structural disorganization, which would be reflected by higher ODI, precedes tissue shrinkage, which would be reflected by higher FWF. Higher ODI but stable FWF in the cerebellar peduncles may be indicative of more disorganization in the fibers but no loss of fibers or tissue shrinkage, while in cerebral white matter both higher ODI and FWF were observed. Future research is needed to replicate these findings and further assess the potential resilience of the cerebellar peduncles to inflammation.

In this study, the observed relationships between inflammatory markers with white matter microstructure were largely independent of biomarkers of AD pathology. While some previous studies have reported relationships between diffusion metrics and AD biomarkers in cognitively unimpaired samples and theorized that the patterns may be attributable to inflammation, these previous studies did not include measures of inflammation (Benitez et al., 2022; Collij et al., 2021; Dong et al., 2020; Hoy et al., 2017; Sun et al., 2024). A number of factors may have contributed to the lack of associations of AD biomarkers with GFAP and the inflammatory measures in the current study. Notably, our sample included many middle-aged participants and was largely cognitively unimpaired. This resulted in a small percentage of participants with clinically abnormal levels of AD biomarkers (11% for Aβ42/Aβ40, 4.49% for p-tau181). Interactions between AD biomarkers and measures of inflammation may be more evident in samples that have greater AD pathology accumulation, particularly those in early symptomatic stages of AD.

This study has several strengths, including a large, multi-site cohort of participants with multiple inflammation-related plasma measures, multi-compartment diffusion metrics, and AD biomarker measures. Limitations should also be noted. First, this was a cross-sectional, observational study. We therefore cannot infer directionality of the relationships between these plasma and diffusion measures, or causal relationships. The relationships observed herein may also be attributable to shared upstream factors, such as vascular risk factors or other non-AD pathological process that give rise to both inflammation and white matter microstructure alterations. Second, a low proportion of the current sample had clinically abnormal AD biomarker levels; future studies with participants who have higher levels of AD pathology are necessary to evaluate potential interactions between inflammation and AD pathology. Third, the present study used plasma measures of inflammation. It is uncertain to what degree these peripheral measures reflect CNS neuroinflammation (e.g., Youn et al., 2025). Thus, future studies are needed examining measures of inflammation in CSF and their associations with dMRI metrics. A further limitation concerns the fixed intrinsic parallel diffusivity assumed by the NODDI model. We used the default value of 1.7 × 10−3 mm2/s for cerebellar white matter; however, the optimal value may vary across brain regions and has not been well established for the cerebellum. Consequently, the absolute cerebellar NODDI parameter estimates may be biased by this model assumption and should be interpreted primarily as relative, model-derived measures rather than direct quantitative estimates of underlying histological properties. This, however, does not affect estimates of relationships between the NODDI measures and the inflammatory markers. Additionally, an explicit correction for magnitude noise-floor bias was not applied prior to AMICO-NODDI fitting. Residual noise-floor effects, particularly in the lower-SNR, high-b-value data, may therefore have introduced some bias into the estimated NODDI parameters. Finally, the current analyses investigated global cerebral white matter tracts and the cerebellar peduncles. Future analyses interrogating regionally specific white matter microstructure may reveal patterns that are unique to different regions and tracts, particularly given potential heterogeneity across tracts.

Supplementary Material

1

Highlights:

  • Higher inflammation was associated with cerebral dMRI values: FWF, ODI, NDI, MD

  • Higher cytokine composite and GFAP values were associated with cerebellar ODI, NDI

  • Inflammation-microstructure associations were independent of Alzheimer’s biomarkers

Acknowledgements.

We thank the Precision Aging® Network for their contribution to the research. The members of the Precision Aging® Network are A. Aldabergenova, A. Bilgin, A. Bonfitto, A. Box, A. Dolby, A. Feal Rodriguez, A. Glinka, A. Moghekar, A. Sidhu, A. Sokan, A. Soldan, A.I. Levey, A.J.B. Lee, B. Aimagamabetova, B. Levin, B. Najafi, B. Nursal, B. Stark, B.J. LaFleur, C.A. Barnes, C.A. Pettigrew, C. Anderson, C. Babbitt, C. Camargo, C. Carrasco, C.H. Na, C. Ugonna, C. Ye, C.S. Mitchell, D. Cabral, D. Chambers, D. Coon, D. Kartchner, D. Metz, D. Sama-Borbon, E. Burrows, E.M. Sternberg, E. Paitel, F. Taguinod, G. Leito, G. Pfaff, H. Venkatachalam, J. Don, J. Pekar, J.R. Runyon, J. Simon, J. Sloan, J. Xie, J. Zhou, J.B. Frye, J.G. Varelo Saboria, J.J Lah, K.A. Delgado, K. Ellingson, K. Johnson, K. Norton, K.P. Doyle, K. Sanders, L.F. Gladulich, L. Gossa, L. Ryan, L. White, M. Albert, M. Dehghan-Rouzi, M. Fan, M. Hay, M. Johnson, M. Lee, M. M. Crespo, M.R. Mehl, M. Modjeski, M. Naymik, M. Swartzlander, M. Taylor John, M.D. De Both, M.J. Huentelman, N. Bhadra, N. Merchant, N. Schork, NK. Chen, P. F. Worley, P. Pattany, R.D. Brinton, R. Tandon, S. Beres, S. Callahan, S. Fox-Rosellini, S. Han, S. Hoscheidt, S. Matijevic, S. Merritt, S. Scott, S. Sharma, S. Sockanathan, S. Soto, S.-E. Roh, S.-Y. Kim, T. James, T. Nuno, T. Rundek, T. Smith, T. Trouard, T. Yuhas, T.K. Zepeda, V.D. Calhoun, V.M. Dotson, V. Pfeifer, W. Degnan III, X. Sun, Y. Yang, Y.F. Bolla, and Z. Chen.

Funding:

This work was supported by the grants from the National Institutes of Health [U19-AG065169, T32-AG027668].

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Submission declaration. We confirm that the work herein has not before been published nor is it currently under consideration elsewhere. The manuscript has been approved by all authors, and the authors have no conflicts of interest to report.

Data statement.

Data used in these analyses are available through standard application procedures described on the PAN website (precisionagingnetwork.org).

References

  1. Alilou S, Valizadeh P, KamaliZonouzi S, Salabat D, Jannatdoust P, Amanollahi M, Rashidi F, Rezaie S, Delavari S, & Aarabi MH. (2024). The association between c-reactive protein and human brain microstructure: A systematic review of diffusion imaging studies. Brain Behav Immun Integr, 8, 100092. 10.1016/j.bbii.2024.100092 [DOI] [Google Scholar]
  2. Arfanakis K, Fleischman DA, Grisot G, Barth CM, Varentsova A, Morris MC, Barnes LL, & Bennett DA. (2013). Systemic inflammation in non-demented elderly human subjects: brain microstructure and cognition. PLoS One, 8(8), e73107. 10.1371/journal.pone.0073107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arleo A, Bareš M, Bernard JA, Bogoian HR, Bruchhage MM, Bryant P, Carlson ES, Chan CC, Chen L-K, & Chung C-P. (2024). Consensus paper: cerebellum and ageing. The Cerebellum, 23(2), 802–832. 10.1007/s12311-023-01577-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baylis D, Bartlett DB, Patel HP, & Roberts HC. (2013). Understanding how we age: insights into inflammaging. Longev Healthspan, 2(1), 8. 10.1186/2046-2395-2-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Benitez A, Jensen JH, Thorn K, Dhiman S, Fountain-Zaragoza S, Rieter WJ, Spampinato MV, Hamlett ED, Nietert PJ, & Falangola M. d. F. (2022). Greater diffusion restriction in white matter in preclinical Alzheimer disease. Ann Neurol, 91(6), 864–877. 10.1002/ana.26353 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Benjamini Y, & Hochberg Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol, 57(1), 289–300. 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
  7. Bernard JA. (2022). Don’t forget the little brain: A framework for incorporating the cerebellum into the understanding of cognitive aging. Neurosci Biobehav Rev, 137, 104639. 10.1016/j.neubiorev.2022.104639 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bettcher BM, Olson KE, Carlson NE, McConnell BV, Boyd T, Adame V, Solano DA, Anton P, Markham N, & Thaker AA. (2021). Astrogliosis and episodic memory in late life: higher GFAP is related to worse memory and white matter microstructure in healthy aging and Alzheimer’s disease. Neurobiol Aging, 103, 68–77. 10.1016/j.neurobiolaging.2021.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Boots EA, Castellanos KJ, Zhan L, Barnes LL, Tussing-Humphreys L, Deoni SC, & Lamar M. (2020). Inflammation, cognition, and white matter in older adults: an examination by race. Front Aging Neurosci, 12, 553998. 10.3389/fnagi.2020.553998 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Buckner RL. (2013). The cerebellum and cognitive function: 25 years of insight from anatomy and neuroimaging. Neuron, 80(3), 807–815. 10.1016/j.neuron.2013.10.044 [DOI] [PubMed] [Google Scholar]
  11. Chen WW, Zhang X, & Huang WJ. (2016). Role of neuroinflammation in neurodegenerative diseases. Mol Med Report, 13(4), 3391–3396. 10.3892/mmr.2016.4948 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cieslak M, Cook PA, He X, Yeh F-C, Dhollander T, Adebimpe A, Aguirre GK, Bassett DS, Betzel RF, & Bourque J. (2021). QSIPrep: an integrative platform for preprocessing and reconstructing diffusion MRI data. Nat Methods, 18(7), 775–778. 10.1038/s41592-021-01185-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Collij LE, Ingala S, Top H, Wottschel V, Stickney KE, Tomassen J, Konijnenberg E, Ten Kate M, Sudre C, & Lopes Alves I. (2021). White matter microstructure disruption in early stage amyloid pathology. Alzheimers Dement (Amst), 13(1), e12124. 10.1002/dad2.12124 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Conole EL, Stevenson AJ, Muñoz Maniega S, Harris SE, Green C, Valdés Hernández MDC, Harris MA, Bastin ME, Wardlaw JM, & Deary IJ. (2021). DNA methylation and protein markers of chronic inflammation and their associations with brain and cognitive aging. Neurology, 97(23), e2340–e2352. 10.1212/WNL.0000000000012997 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dong JW, Jelescu IO, Ades-Aron B, Novikov DS, Friedman K, Babb JS, Osorio RS, Galvin JE, Shepherd TM, & Fieremans E. (2020). Diffusion MRI biomarkers of white matter microstructure vary nonmonotonically with increasing cerebral amyloid deposition. Neurobiol Aging, 89, 118–128. 10.1016/j.neurobiolaging.2020.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Du Clos TW. (2000). Function of C-reactive protein. Ann Med, 32(4), 274–278. 10.3109/07853890009011772 [DOI] [PubMed] [Google Scholar]
  17. Dukhinova MS, Guo J, Shen E, Liu W, Huang W, Shen Y, & Wang L. (2024). Cerebellar microglia: On the edge between neuroinflammation and neuroregulation. Neural Regen Res, 21(1), 156–172. 10.4103/NRR.NRR-D-24-00550 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Eng LF, & Ghirnikar RS. (1994). GFAP and astrogliosis. Brain Pathol, 4(3), 229–237. 10.1111/j.1750-3639.1994.tb00838.x [DOI] [PubMed] [Google Scholar]
  19. Fowler CJ, Stoops E, Rainey-Smith SR, Vanmechelen E, Vanbrabant J, Dewit N, Mauroo K, Maruff P, Rowe CC, & Fripp J. (2022). Plasma p-tau181/Aβ1-42 ratio predicts Aβ-PET status and correlates with CSF-p-tau181/Aβ1-42 and future cognitive decline. Alzheimers Dement (Amst), 14(1), e12375. 10.1002/dad2.12375 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Friedman EM, & Herd P. (2010). Income, education, and inflammation: differential associations in a national probability sample (The MIDUS study). Biopsychosoc Sci Med, 72(3), 290–300. 10.1097/PSY.0b013e3181cfe4c2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fuller-Rowell TE, Curtis DS, Doan SN, & Coe CL. (2015). Racial disparities in the health benefits of educational attainment: A study of inflammatory trajectories among African American and white adults. Biopsychosoc Sci Med, 77(1), 33–40. 10.1097/PSY.0000000000000128 [DOI] [PubMed] [Google Scholar]
  22. Galea I. (2021). The blood–brain barrier in systemic infection and inflammation. Cell Mol Immunol, 18(11), 2489–2501. 10.1038/s41423-021-00757-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Garcia-Hernandez R, Cerdan Cerda A, Trouve Carpena A, Drakesmith M, Koller K, Jones DK, Canals S, & De Santis S. (2022). Mapping microglia and astrocyte activation in vivo using diffusion MRI. Sci Adv, 8(21), eabq2923. 10.1126/sciadv.abq2923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gellersen HM, Guell X, & Sami S. (2021). Differential vulnerability of the cerebellum in healthy ageing and Alzheimer’s disease. NeuroImage Clin, 30, 102605. 10.1016/j.nicl.2021.102605 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Gellersen HM, Guo CC, O’Callaghan C, Tan RH, Sami S, & Hornberger M. (2017). Cerebellar atrophy in neurodegeneration—a meta-analysis. J Neurol Neurosurg Psychiatry, 88(9), 780–788. 10.1136/jnnp-2017-315607 [DOI] [PubMed] [Google Scholar]
  26. Godbout JP, & Johnson RW. (2009). Age and neuroinflammation: a lifetime of psychoneuroimmune consequences. Immunol Allergy Clin North Am, 29(2), 321–337. 10.1016/j.iac.2009.02.007 [DOI] [PubMed] [Google Scholar]
  27. Gottesman RF, Albert MS, Alonso A, Coker LH, Coresh J, Davis SM, Deal JA, McKhann GM, Mosley TH, & Sharrett AR. (2017). Associations between midlife vascular risk factors and 25-year incident dementia in the Atherosclerosis Risk in Communities (ARIC) cohort. JAMA Neurol, 74(10), 1246–1254. 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Gu Y, Vorburger R, Scarmeas N, Luchsinger JA, Manly JJ, Schupf N, Mayeux R, & Brickman AM. (2017). Circulating inflammatory biomarkers in relation to brain structural measurements in a non-demented elderly population. Brain Behav Immun, 65, 150–160. 10.1016/j.bbi.2017.04.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gyengesi E, Rangel A, Ullah F, Liang H, Niedermayer G, Asgarov R, Venigalla M, Gunawardena D, Karl T, & Münch G. (2019). Chronic microglial activation in the GFAP-IL6 mouse contributes to age-dependent cerebellar volume loss and impairment in motor function. Front Neurosci, 13, 303. 10.3389/fnins.2019.00303 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Heneka MT, Carson MJ, El Khoury J, Landreth GE, Brosseron F, Feinstein DL, Jacobs AH, Wyss-Coray T, Vitorica J, & Ransohoff RM. (2015). Neuroinflammation in Alzheimer’s disease. Lancet Neurol, 14(4), 388–405. 10.1016/S1474-4422(15)70016-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hoxha E, Lippiello P, Zurlo F, Balbo I, Santamaria R, Tempia F, & Miniaci MC. (2018). The emerging role of altered cerebellar synaptic processing in Alzheimer’s disease. Front Aging Neurosci, 10, 396. 10.3389/fnagi.2018.00396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hoy AR, Ly M, Carlsson CM, Okonkwo OC, Zetterberg H, Blennow K, Sager MA, Asthana S, Johnson SC, & Alexander AL. (2017). Microstructural white matter alterations in preclinical Alzheimer’s disease detected using free water elimination diffusion tensor imaging. PLoS One, 12(3), e0173982. 10.1371/journal.pone.0173982 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Hsuchou H, Kastin AJ, Mishra PK, & Pan W. (2012). C-reactive protein increases BBB permeability: implications for obesity and neuroinflammation. Cell Physiol Biochem, 30(5), 1109–1119. 10.1159/000343302 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Huang X, Han X, Chang H, Yu T, Dong Y, Mao M, Ding J, Li X, Liu H, & Tian M. (2025). Associations between trajectories of plasma biomarkers for Alzheimer’s disease, brain structures, and cognitive function: a prospective cohort study in the UK Biobank. Mol Psychiatry, 1–12. 10.1038/s41380-025-03166-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Islam N, Hashem R, Gad M, Brown A, Levis B, Renoux C, Thombs BD, & McInnes MD. (2023). Accuracy of the Montreal Cognitive Assessment tool for detecting mild cognitive impairment: A systematic review and meta-analysis. Alzheimers Dement, 19(7), 3235–3243. 10.1002/alz.13040 [DOI] [PubMed] [Google Scholar]
  36. Ismail YA, Auf HA, Sadik SA, Ahmed NM, & Ali Y. (2025). Sensitivity and specificity of the Montreal cognitive assessment using US National alzheimer coordinating centre uniform data set: a retrospective analysis of 16,309 participants. BMC Neurol, 25(1), 381. 10.1186/s12883-025-04190-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Jacobs HI, Leritz EC, Williams VJ, Van Boxtel MP, Elst W. v. d., Jolles J, Verhey FR, McGlinchey RE, Milberg WP, & Salat DH. (2013). Association between white matter microstructure, executive functions, and processing speed in older adults: the impact of vascular health. Hum Brain Mapp, 34(1), 77–95. 10.1002/hbm.21412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Jelescu IO, & Fieremans E. (2023). Sensitivity and specificity of diffusion MRI to neuroinflammatory processes. In Advances in magnetic resonance technology and applications (Vol. 9, pp. 31–50). Elsevier. 10.1016/B978-0-323-91771-1.00010-1 [DOI] [Google Scholar]
  39. Jones DK, Knösche TR, & Turner R. (2013). White matter integrity, fiber count, and other fallacies: the do’s and don’ts of diffusion MRI. Neuroimage, 73, 239–254. 10.1016/j.neuroimage.2012.06.081 [DOI] [PubMed] [Google Scholar]
  40. Kamila P, Kar K, Chowdhury S, Chakraborty P, Dutta R, Singh A, & Prajapati BG. (2025). Effect of neuroinflammation on the progression of Alzheimer’s disease and its significant ramifications for novel anti-inflammatory treatments. IBRO Neurosci Rep, 18, 771–782. 10.1016/j.ibneur.2025.05.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Kennedy KM, & Raz N. (2009). Pattern of normal age-related regional differences in white matter microstructure is modified by vascular risk. Brain Res, 1297, 41–56. 10.1016/j.brainres.2009.08.058 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kim B-H, Shin D, Kang SH, Jang H, Yun J, Park Y, Zetterberg H, Blennow K, Gonzalez-Ortiz F, & Ashton NJ. (2025). Mapping Alzheimer’s disease pathology using free water through integrated analysis of plasma biomarkers, microstructural DTI metrics, and macrostructural MRI measures. Sci Rep, 15(1), 30702. 10.1038/s41598-025-14200-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Kim E, Figueiredo IC, Simmons C, Randall K, Gonzalez LR, Wood T, Ranieri B, Sureda-Gibert P, Howes O, & Pariante C. (2023). Mapping acute neuroinflammation in vivo with diffusion-MRI in rats given a systemic lipopolysaccharide challenge. Brain Behav Immun, 113, 289–301. 10.1016/j.bbi.2023.07.010 [DOI] [PubMed] [Google Scholar]
  44. Lee EH, Kang SH, Shin D, Kim YJ, Zetterberg H, Blennow K, Gonzalez-Ortiz F, Ashton NJ, Cheon BK, & Yoo H. (2025). Plasma Alzheimer’s disease biomarker variability: amyloid-independent and amyloid-dependent factors. Alzheimers Dement, 21(1), e14368. 10.1002/alz.14368 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Leipp F, Vialaret J, Mohaupt P, Coppens S, Jaffuel A, Niehoff AC, Lehmann S, & Hirtz C. (2024). Glial fibrillary acidic protein in Alzheimer’s disease: a narrative review. Brain Commun, 6(6), fcae396. 10.1093/braincomms/fcae396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Liu G, Yang C, Wang X, Chen X, Cai H, & Le W. (2024). Cerebellum in neurodegenerative diseases: Advances, challenges, and prospects. Iscience, 27(11). 10.1016/j.isci.2024.111194 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lyman M, Lloyd DG, Ji X, Vizcaychipi MP, & Ma D. (2014). Neuroinflammation: the role and consequences. Neurosci Res, 79, 1–12. 10.1016/j.neures.2013.10.004 [DOI] [PubMed] [Google Scholar]
  48. Malek-Ahmadi M, & Nikkhahmanesh N. (2024). Meta-analysis of Montreal cognitive assessment diagnostic accuracy in amnestic mild cognitive impairment. Front Psychol, 15, 1369766. 10.3389/fpsyg.2024.1369766 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Miralbell J, Soriano JJ, Spulber G, López-Cancio E, Arenillas JF, Bargalló N, Galán A, Barrios MT, Cáceres C, & Alzamora MT. (2012). Structural brain changes and cognition in relation to markers of vascular dysfunction. Neurobiol Aging, 33(5), 1003.e1009–1003.e1017. 10.1016/j.neurobiolaging.2011.09.020 [DOI] [PubMed] [Google Scholar]
  50. Moyse E, Krantic S, Djellouli N, Roger S, Angoulvant D, Debacq C, Leroy V, Fougere B, & Aidoud A. (2022). Neuroinflammation: a possible link between chronic vascular disorders and neurodegenerative diseases. Front Aging Neurosci, 14, 827263. 10.3389/fnagi.2022.827263 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, & Chertkow H. (2005). The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc, 53(4), 695–699. 10.1111/j.1532-5415.2005.53221.x [DOI] [PubMed] [Google Scholar]
  52. O’Donovan A, Bahorik A, Sidney S, Launer LJ, & Yaffe K. (2021). Relationships of inflammation trajectories with white matter volume and integrity in midlife. Brain Behav Immun, 91, 81–88. 10.1016/j.bbi.2020.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Oestreich LK, & O’Sullivan MJ. (2022). Transdiagnostic in vivo magnetic resonance imaging markers of neuroinflammation. Biol Psychiatry Cogn Neurosci Neuroimaging, 7(7), 638–658. 10.1016/j.bpsc.2022.01.003 [DOI] [PubMed] [Google Scholar]
  54. Pierpaoli C, Jezzard P, Basser PJ, Barnett A, & Di Chiro G. (1996). Diffusion tensor MR imaging of the human brain. Radiology, 201(3), 637–648. 10.1148/radiology.201.3.8939209 [DOI] [PubMed] [Google Scholar]
  55. Plank JR, Morgan CA, Dell’Acqua F, Sundram F, Hoeh NR, Muthukumaraswamy S, & Lin JC. (2025). Mapping neuroinflammation with diffusion-weighted MRI: randomized crossover study. Biol Psychiatry Cogn Neurosci Neuroimaging. 10.1016/j.bpsc.2025.05.002 [DOI] [PubMed] [Google Scholar]
  56. Quaresima V, Pilotto A, Trasciatti C, Tolassi C, Parigi M, Bertoli D, Mordenti C, Galli A, Rizzardi A, & Caratozzolo S. (2024). Plasma p-tau181 and amyloid markers in Alzheimer’s disease: A comparison between Lumipulse and SIMOA. Neurobiol Aging, 143, 30–40. 10.1016/j.neurobiolaging.2024.08.007 [DOI] [PubMed] [Google Scholar]
  57. Rodrigue AL, Knowles EE, Mollon J, Mathias SR, Koenis MM, Peralta JM, Leandro AC, Fox PT, Sprooten E, & Kochunov P. (2019). Evidence for genetic correlation between human cerebral white matter microstructure and inflammation. Hum Brain Mapp, 40(14), 4180–4191. 10.1002/hbm.24694 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Schmahmann JD. (2016). Cerebellum in Alzheimer’s disease and frontotemporal dementia: not a silent bystander. Brain, 139(5), 1314–1318. 10.1093/brain/aww064 [DOI] [PubMed] [Google Scholar]
  59. Schmahmann JD. (2019). The cerebellum and cognition. Neurosci Lett, 688, 62–75. 10.1016/j.neulet.2018.07.005 [DOI] [PubMed] [Google Scholar]
  60. Schmahmann JD, Guell X, Stoodley CJ, & Halko MA. (2019). The theory and neuroscience of cerebellar cognition. Annu Rev Neurosci, 42(1), 337–364. 10.1146/annurev-neuro-070918-050258 [DOI] [PubMed] [Google Scholar]
  61. Shabab T, Khanabdali R, Moghadamtousi SZ, Kadir HA, & Mohan G. (2017). Neuroinflammation pathways: a general review. Int J Neurosci, 127(7), 624–633. 10.1080/00207454.2016.1212854 [DOI] [PubMed] [Google Scholar]
  62. Sproston NR, & Ashworth JJ. (2018). Role of C-reactive protein at sites of inflammation and infection. Front Immunol, 9, 754. 10.3389/fimmu.2018.00754 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Stoessel MB, & Majewska AK. (2021). Little cells of the little brain: microglia in cerebellar development and function. Trends Neurosci, 44(7), 564–578. 10.1016/j.tins.2021.04.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Stoodley CJ. (2012). The cerebellum and cognition: evidence from functional imaging studies. The Cerebellum, 11, 352–365. 10.1007/s12311-011-0260-7 [DOI] [PubMed] [Google Scholar]
  65. Stowell RD, Wong EL, Batchelor HN, Mendes MS, Lamantia CE, Whitelaw BS, & Majewska AK. (2018). Cerebellar microglia are dynamically unique and survey Purkinje neurons in vivo. Dev Neurobiol, 78(6), 627–644. 10.1002/dneu.22572 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Sumra V, Hadian M, Dilliott AA, Farhan SM, Frank AR, Lang AE, Roberts AC, Troyer A, Arnott SR, & Marras C. (2025). Regional free-water diffusion is more strongly related to neuroinflammation than neurodegeneration. J Neurol, 272(7), 478. 10.1007/s00415-025-13201-1 [DOI] [PubMed] [Google Scholar]
  67. Sun X, Zhao C, Chen SY, Chang Y, Han YL, Li K, Sun HM, Wang ZF, Liang Y, & Jia JJ. (2024). Free water MR imaging of white matter microstructural changes is a sensitive marker of amyloid positivity in Alzheimer’s disease. J Magn Reson Imaging, 60(4), 1458–1469. 10.1002/jmri.29189 [DOI] [PubMed] [Google Scholar]
  68. Tournier J-D, Smith R, Raffelt D, Tabbara R, Dhollander T, Pietsch M, Christiaens D, Jeurissen B, Yeh C-H, & Connelly A. (2019). MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation. Neuroimage, 202, 116137. 10.1016/j.neuroimage.2019.116137 [DOI] [PubMed] [Google Scholar]
  69. Veraart J, Novikov DS, Christiaens D, Ades-Aron B, Sijbers J, & Fieremans E. (2016). Denoising of diffusion MRI using random matrix theory. Neuroimage, 142, 394–406. 10.1016/j.neuroimage.2016.08.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Walker KA, Power MC, Hoogeveen RC, Folsom AR, Ballantyne CM, Knopman DS, Windham BG, Selvin E, Jack CR Jr, & Gottesman RF. (2017). Midlife systemic inflammation, late-life white matter integrity, and cerebral small vessel disease: the atherosclerosis risk in communities study. Stroke, 48(12), 3196–3202. 10.1161/STROKEAHA.117.018675 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Walker KA, Windham BG, Power MC, Hoogeveen RC, Folsom AR, Ballantyne CM, Knopman DS, Selvin E, Jack CR Jr, & Gottesman RF. (2018). The association of mid-to late-life systemic inflammation with white matter structure in older adults: The Atherosclerosis Risk in Communities Study. Neurobiol Aging, 68, 26–33. 10.1016/j.neurobiolaging.2018.03.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Wersching H, Duning T, Lohmann H, Mohammadi S, Stehling C, Fobker M, Conty M, Minnerup J, Ringelstein E, & Berger K. (2010). Serum C-reactive protein is linked to cerebral microstructural integrity and cognitive function. Neurology, 74(13), 1022–1029. 10.1212/WNL.0b013e3181d7b45b [DOI] [PubMed] [Google Scholar]
  73. Williams OA, An Y, Beason-Held L, Huo Y, Ferrucci L, Landman BA, & Resnick SM. (2019). Vascular burden and APOE ε4 are associated with white matter microstructural decline in cognitively normal older adults. Neuroimage, 188, 572–583. 10.1016/j.neuroimage.2018.12.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Yamamoto M, Kim M, Imai H, Itakura Y, & Ohtsuki G. (2019). Microglia-triggered plasticity of intrinsic excitability modulates psychomotor behaviors in acute cerebellar inflammation. Cell Rep, 28(11), 2923–2938. e2928. 10.1016/j.celrep.2019.07.078 [DOI] [PubMed] [Google Scholar]
  75. Yi SY, Barnett BR, Torres-Velázquez M, Zhang Y, Hurley SA, Rowley PA, Hernando D, & Yu J-PJ. (2019). Detecting microglial density with quantitative multi-compartment diffusion MRI. Front Neurosci, 13, 81. 10.3389/fnins.2019.00081 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Youn W, Yun M, Lee CJ, & Schöll M. (2025). Cautions on utilizing plasma GFAP level as a biomarker for reactive astrocytes in neurodegenerative diseases. Mol Neurodegener, 20(1), 54. 10.1186/s13024-025-00846-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Zhang H, Schneider T, Wheeler-Kingshott CA, & Alexander DC. (2012). NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. Neuroimage, 61(4), 1000–1016. 10.1016/j.neuroimage.2012.03.072 [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS2213021-supplement-1.docx (67KB, docx)

Data Availability Statement

Data used in these analyses are available through standard application procedures described on the PAN website (precisionagingnetwork.org).

RESOURCES