Skip to main content
NCBI home page
AJNR: American Journal of Neuroradiology logoLink to AJNR: American Journal of Neuroradiology
. 2024 Dec;45(12):1910–1918. doi: 10.3174/ajnr.A8481

Macro- and Microstructural White Matter Differences in Neurologic Postacute Sequelae of SARS-CoV-2 Infection

Erin E O’Connor a,✉, Rosangela Salerno-Goncalves b, Nikita Rednam a, Rory O’Brien c, Peter Rock d, Andrea R Levine e, Thomas A Zeffiro a
PMCID: PMC11630878  PMID: 39389778

In this cross-sectional study, SARS-CoV-2 infection effects on GM/WM volume, WM microstructure, cognitive performance, and peripheral cytokine concentrations were investigated. The study found that the NeuroPASC group had larger global and regional (prefrontal cortex and hippocampus) cerebral WM volume, more self-reported PASC symptoms, more headaches, lower psychometric performance, widespread higher mean kurtosis in WM microstructure, and different cytokine concentration profiles compared with controls. The authors conclude that immune dysregulation following SARS-CoV-2 infection might result in cerebral WM macrostructural and microstructural injury causing NeuroPASC clinical signs and symptoms.

Abstract

BACKGROUND AND PURPOSE:

Neuropsychiatric complications of SARS-CoV-2 infection, also known as neurologic postacute sequelae of SARS-CoV-2 infection (NeuroPASC), affect 10%–60% of infected individuals. There is growing evidence that NeuroPASC is a multi system immune dysregulation disease affecting the brain. The behavioral manifestations of NeuroPASC, such as impaired processing speed, executive function, memory retrieval, and sustained attention, suggest widespread WM involvement. Although previous work has documented WM damage following acute SARS-CoV-2 infection, its involvement in NeuroPASC is less clear. We hypothesized that macrostructural and microstructural WM differences in NeuroPASC participants would accompany cognitive and immune system differences.

MATERIALS AND METHODS:

In a cross-sectional study, we screened a total of 159 potential participants and enrolled 72 participants, with 41 asymptomatic controls (NoCOVID) and 31 NeuroPASC participants matched for age, sex, and education. Exclusion criteria included neurologic disorders unrelated to SARS-CoV-2 infection. Assessments included clinical symptom questionnaires, psychometric tests, brain MRI measures, and peripheral cytokine levels. Statistical modeling included separate multivariable regression analyses of GM/WM/CSF volume, WM microstructure, cognitive, and cytokine concentration between-group differences.

RESULTS:

NeuroPASC participants had larger cerebral WM volume than NoCOVID controls (β = 0.229; 95% CI: 0.017–0.441; t = 2.16; P = .035). The most pronounced effects were in the prefrontal and anterior temporal WM. NeuroPASC participants also exhibited higher WM mean kurtosis, consistent with ongoing neuroinflammation. NeuroPASC participants had more self-reported symptoms, including headache, and lower performance on measures of attention, concentration, verbal learning, and processing speed. A multivariate profile analysis of the cytokine panel showed different group cytokine profiles (Wald-type-statistic = 44.6, P = .046), with interferon (IFN)-λ1 and IFN-λ2/3 levels higher in the NeuroPASC group.

CONCLUSIONS:

NeuroPASC participants reported symptoms of lower concentration, higher fatigue, and impaired cognition compatible with WM syndrome. Psychometric testing confirmed these findings. NeuroPASC participants exhibited larger cerebral WM volume and higher WM mean kurtosis than NoCOVID controls. These findings suggest that immune dysregulation could influence WM properties to produce WM volume increases and consequent cognitive effects and headaches. Further work will be needed to establish mechanistic links among these variables.

Neuropsychiatric complications of SARS-CoV-2 infection, also known as neurologic postacute sequelae of SARS-CoV-2 infection (NeuroPASC), affect 10%–60% of infected individuals.1,2 These sequelae include fatigue, memory problems, attention deficits, sleep disorders, headaches, anosmia/hyposmia, depression, and anxiety, constituting a syndrome that can last for months or even years after infection.3,4 While the signs and symptoms of NeuroPASC are well-documented, its biologic basis is poorly understood.

The behavioral manifestations of NeuroPASC, such as impairments in processing speed, executive function, memory retrieval, and sustained attention, suggest the type of widespread WM injury associated with “white matter syndromes.”5–7 The psychometric profile characterizing WM syndromes includes cognitive slowing, executive dysfunction, memory retrieval deficits, sustained attention problems, visuospatial impairment, and emotional dysregulation.5,8 The behavioral importance of WM function is underscored by associations among WM injury, cognitive impairment, and depression in other diseases, including vascular dementia, major depressive disorder, multiple sclerosis, traumatic brain injury, and following chemotherapy.7,9–12 Deficits in executive function, processing speed, attention, and memory are common in these studies. Nevertheless, as distributed neuronal dysfunction could mimic many of these signs, we cannot exclude that possibility. Because the behavioral problems seen in NeuroPASC are consistent with WM syndrome, we focused on WM macrostructural and microstructural measures.

WM tracts comprise approximately one-half of total brain volume, connecting cortical and subcortical GM regions into functional neural networks. The wide distribution of these tracts lends itself to variation as to which neural networks can be impaired.13 As SARS-CoV-2 does not appear to have high specificity for any particular WM tract, high behavioral variability can be expected in the NeuroPASC phenotype.

In acute SARS-CoV-2 infection, cross-sectional studies find WM hyperintensities by using conventional FLAIR MRI.14–16 Although acute imaging studies can be confounded by infection severity and mechanical ventilation effects, WM injury has also been found following mild and moderate acute infection.17 By contrast, visual examination of clinical structural MRI frequently fails to detect abnormalities in NeuroPASC, even in the presence of neurologic symptoms and signs.

To examine the contribution of WM dysfunction to cognitive impairment in NeuroPASC, we used quantitative analysis of high-resolution MRI to examine WM macrostructure and diffusion MRI (dMRI) to examine WM microstructure. For clinical characterization, we used postacute sequelae of SARS-CoV-2 infection (PASC) scores, a symptom framework used for clinical case identification that heavily weights NeuroPASC symptoms. Because there is growing evidence that persisting symptoms after SARS-CoV-2 infection may be related to immune dysregulation, we measured chemokine/cytokine concentrations to determine if cytokine differences accompanied the WM, cognitive performance, and clinical manifestations. Our central hypothesis is that self-reported clinical symptoms summarized by using the PASC score and associated psychometric deficits consistent with WM syndromes would be accompanied by WM structural effects. These effects will scale with NeuroPASC clinical severity.

We used a STROBE reporting checklist to guide this work.

MATERIALS AND METHODS

Ethics Statement

Study procedures were approved by our Institutional Review Board. We conducted procedures in accordance with principles outlined in the 1964 Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines. Before enrollment, we informed participants of research objectives and procedures. Written informed consent was obtained.

Participants

We recruited participants through medical center referrals, advertisements, and the university COVID-19 biorepository. Inclusion criteria included 1) men or women of any ethnicity, ages 18–70 years, 2) fluency in English, and 3) adequate visual and auditory acuity for testing. Recruitment took place between September 2021 and February 2023. SARS-CoV-2 infected participants were asked about the presence of “cognitive fog” or mental fatigue beginning after SARS-CoV-2 infection.

To increase generalizability, alcohol, marijuana, and nicotine use were permitted. Marijuana use was not permitted on the day of, or day before, participation. Participants were excluded if substance use or substance use disorder was reported in the past 180 days. We administered urine drug screens on the testing day.

Other exclusion criteria included 1) pregnancy, 2) claustrophobia, 3) MRI contraindications, 4) neurologic disorders unrelated to SARS-CoV-2 infection including seizure disorders, closed head injuries with loss of consciousness greater than 15 minutes, CNS neoplasm, multiple sclerosis, neurodegenerative disease or stroke, 5) unstable cardiac or pulmonary disease, and 6) chronic fatigue. All participants had a negative rapid antigen test to screen for acute SARS-CoV-2 infection.

To determine if cognitive fog or mental fatigue were present, participants were asked during initial screening if they experienced problems with remembering events, concentrating, sustaining mental effort, thinking and focus, slowed conversational responses, or timely word selection and word retrieval. All participants assigned to the NeuroPASC group responded yes to 1 or more of these questions.

After the initial screening, SARS-CoV-2 participants completed structured questionnaires to better characterize their self-reported neurologic and systemic symptoms. We administered the neuro-quality of life questions from the NIH Toolbox (NeuroQOL).18 It included self-reported measures of anxiety, depression, apathy, fatigue, sleep, executive function, and cognition. A separate headache questionnaire identified the presence of persistent headaches beginning after the acute SARS-CoV-2 infection. Another questionnaire probed the history of drug use. Participants also answered 8 items regarding any dates of acute infection, hospitalization for SARS-CoV-2 infection (COVID-19), treatments, and vaccination status. Depression was assessed by using the Center for Epidemiological Studies Depression Scale (CES-D).19

Using the questionnaires and NeuroQoL answers, we computed a composite, weighted symptom score that classified participants as having PASC on the basis of 12 pertinent symptoms, of which at least 6 reflected neurologic effects2 (Online Supplemental Data). The PASC score was developed by using a sample of over 5000 patients.2 The score was used as a continuous estimate of clinical severity to probe associations with WM volume and psychometric performance, as described below. Other studies characterizing patients with persistent symptoms after SARS-CoV-2 infection have had less stringent inclusion criteria.20–22 All our participants with persistent symptoms after SARS-CoV-2 infection had at least 1 neurologic symptom and met World Health Organization criteria for post-COVID syndrome.23

We confirmed prior SARS-CoV-2 infection with polymerase chain reaction and/or antibodies to SARS-CoV-2 nucleocapsid protein, which has sensitivity for at least 15 months after infection.24 Infected participants were enrolled at least 12 weeks after their initial date of infection. Uninfected control participants (NoCOVID) had no known history of SARS-CoV-2 infection, no neurologic complaints, and no detectable antibodies to SARS-CoV-2 nucleocapsid protein.

Psychometric Assessments

As patients’ self-reported symptoms can sometimes be at variance with objective testing results in neurologic disease, we further assessed cognitive signs by using psychometric tests that evaluated domains known to be abnormal in WM syndrome.

The d2 Test of Attention measured attention, concentration, and processing speed by requesting participants to differentiate visually similar stimuli by using a pencil to mark target characters (“d” with a total of 2 dashes placed above and/or below) occurring among nontarget characters (“d” with more or less than 2 dashes, and “p” characters with any number of dashes) in 14 consecutive 20-second trials. The Auditory Verbal Learning Test from the NIH Toolbox measured immediate memory by presenting 15 unrelated words, followed by attempted recall.

Image Acquisition

Using a 3T Prisma Scanner (Siemens) we collected 1) high-resolution (1 mm3) T1-weighted MPRAGE structural scans to derive global and regional estimates of GM/WM/CSF volume, 2) multi-shell dMRI (1.7 mm3, fractional anisotropy [FA] = 90, TE = 89, TR = 4200, multiband = 3, with 96 directions and b = 0, 500, 1000, 2000, and 3000) to assess WM microstructure, and 3) high-resolution (1 mm3) FLAIR images to assess WM hyperintensities. Processing of T1-weighted structural images with CAT1225 included spatial normalization, tissue segmentation, and smoothing (8 mm full width at half maximum), followed by computation of regional maps of GM, WM, and CSF volume.26 CAT12 provides voxelwise estimates as well as total estimates of GM, WM, CSF and total intracranial volume based on the tissue segmentations. Processing of multi shell dMRI images with PyDesigner included tools from FSL and MRtrix3 for DICOM conversion, denoising, Gibbs ringing correction, eddy current motion correction, brain masking, image smoothing, and Rician bias correction.27

Plasma Cytokine/Chemokines

Because there is growing evidence that persisting symptoms after SARS-CoV-2 infection may be related to immune dysregulation, we investigated a select panel of chemokines/cytokines. Blood was collected and centrifuged at the study visit. Plasma was stored at −80°C. Using the LEGENDplex Human Anti-Virus Response Panel, a bead-based multiplex flow cytometric assay (Biolegend), we measured 13 cytokines/chemokines, including interleukin (IL)-1β, IL-6, IL-8, IL-10, IL-12p70, interferon (IFN)-α2, IFN-β, IFN-λ1, IFN-λ2/3, IFN-γ, TNF-α, interferon-inducible protein-10, and GM-CSF.20 Beads were acquired by flow cytometry on an LSR-II instrument (BD Biosciences). Samples were run in duplicate, and 4000 beads per analyte were acquired per sample. Data were analyzed by using Qognit LEGENDplex software (BioLegend).

Statistical Analysis

Univariate and multivariate statistical analysis was done by using R 4.4.1 and included separate analyses of Fazekas scores, brain global GM/WM volumes, psychometric test results, and cytokine concentration differences. Voxelwise analysis of regional GM/WM was done by using SPM12/CAT12. Voxelwise analysis of dMRI mean kurtosis measures was done by using PyDesigner. Models were adjusted for age and sex.

Between-group differences in psychometric test results were investigated by using multivariable regression and reported as standardized parameter estimates and 95% confidence intervals.

To assure that differences in the WM segmentations were not influenced by WM hyperintensities, the study neuroradiologist quantified WM lesions from the FLAIR images by using the Fazekas score, a widely used ordinal scale ranging from 0–3 to visually rate the severity of hyperintense WM signal abnormalities in MRI data.28,29 Group differences in the Fazekas score were examined by using multivariable regression.

To examine brain macrostructural effects, tissue volumes were examined for 1) total cerebral WM/GM/CSF volume differences by using multivariable regression and 2) regional cerebral WM/GM volume differences by using voxelwise multivariable regression adjusted for age and sex, in which the critical thresholds for the voxelwise statistics were corrected for multiple tests by using Gaussian random field theory (family-wise error [FWE], P < .05).

After processing of dMRI, voxelwise FA was estimated from the tensor fit and nonlinearly registered with FSL tools to an FA template in MNI-152 space.30 Diffusional kurtosis imaging (DKI) analysis used the same multi shell diffusion images, allowing estimates of voxelwise mean kurtosis.31,32 Tract-based spatial statistics (TBSS; https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/TBSS) identifies WM voxels with high FA by registering all participants to a common reference, then creating an FA skeleton by erosion for voxelwise statistical modeling. The dMRI parameter maps were modeled as outcomes and group as a predictor, adjusting for age and sex. Critical thresholds were corrected for multiple tests by using Gaussian random field theory (FWE, P < .05).

A multivariate repeated measures cytokine profile analysis with 1 within-subject factor (cytokine) and 3 between-subject factors (group, age, and sex) was used to examine group cytokine effects, by using a Wald-type statistic (WTS; with critical threshold of P < .05.

RESULTS

Participants

We screened a total of 159 potential participants and enrolled 72 participants, including 41 asymptomatic controls (NoCOVID) and 31 NeuroPASC participants, with groups balanced for age, sex, and education (Table). No participant reported more than 1 SARS-CoV-2 infection. Participant substance use, demographic, and clinical characteristics are shown in Online Supplemental Data, and Tables 2–4).

Participant characteristics

NoCOVID NeuroPASC SMD
n 41 31
PASC score2 1.6 (3.1) 13.0 (10.8) 1.43
Age in years (mean [SD]) 46.8 (15.6) 45.0 (13.5) 0.12
Sex = F (%) 19 (46.3) 14 (45.2) 0.024
Education years (SD) 16.7 (3.1) 16.7 (3.9) 0.009
CES-D score (mean [SD]) 6.4 (5.4) 14.3 (11.0) 0.913
Headache 6 (14.6) 17 (54.8) 0.931
Vaccination = Yes (%) 40 (100.0) 29 (93.5) –
Hospitalization = Yes (%) – 9 (29.0) –
Supplemental oxygen = Yes (%) – 7 (23.3) –
Mechanical ventilation = Yes (%) – 3 (10.0) –
Days from infection (mean [SD]) – 322.13 (179.52) –
Open in a new tab

Note:—SD indicates standard deviation; SMD, standardized mean difference.

Quality of Life Self-Report

In the NeuroPASC group, NeuroQoL symptoms in all domains were significantly higher compared with NoCOVID participants, with higher scores reflecting more impairment (Fig 1).

FIG 1.

FIG 1.

Open in a new tab

In NeuroPASC compared with NoCOVID participants, NeuroQoL symptoms were higher in all domains, reflecting greater disability. Group differences are shown as a forest plot by using θ scores, with values of 1 representing 1 standard deviation (SD) of difference.

Psychometric Tests

In the d2 Test of Attention, NeuroPASC participants had significantly lower processing speed (t = −2.66), concentration (t = −2.28), and accuracy (t = −2.55) (Fig 2). The NeuroPASC group also exhibited significantly fewer correct items in Auditory Verbal Learning test performance (β = −0.51; 95% CI: −0.94–−0.07; t = −2.32).

FIG 2.

FIG 2.

Open in a new tab

On the d2 Test of Attention between-group comparisons, NeuroPASC participants showed significantly: (A) lower processing speed (β = −0.30; 95% CI: −0.51–−0.09; t = −2.80), (B) concentration (β = −0.22; 95% CI: −0.43–−0.01; t = −2.09), and (C) accuracy (β = −0.27; 95% CI: −0.48–−0.06; t = −2.60).

WM Hyperintensities

Using a linear model adjusted for age and sex, we did not find evidence for group differences in WM hyperintensities. FLAIR image Fazekas scores (β = 0.22; 95% CI: −0.14–0.57; t = 1.21) showed that clinically apparent WM changes were not associated with the observed WM macrostructural effects. WM hyperintensity scores increased with age (β = 0.69; 95% CI: 0.51–0.86; t = 7.81) and were lower in women (β = −0.51; 95% CI: −0.86–−0.17; t = −2.95) (Online Supplemental Data).

Total Tissue Volumes

CAT12 provides ROI estimates of total GM, WM, CSF, and total intracranial volume. Adjusting for age, sex, and total intracranial volume, total cerebral WM volume was 18.1 mL higher in the NeuroPASC compared with the NoCOVID group (95% CI: 1.34–34.85; t = 2.16). Total GM volume was 11.4 mL higher (95% CI: −12.627–34.911; t = 0.94). CSF volume was 29.2 mL lower (95% CI: −62.592–4.121; t = −1.75) (Fig 3, left). As expected, we observed no between-group differences in total intracranial volume.

FIG 3.

FIG 3.

Open in a new tab

The total intracranial tissue compartments show interrelated changes consistent with maintenance of constant intracranial volume for each participant. In NeuroPASC participants, cerebral global WM volume is significantly higher, GM is higher, and CSF is lower (left). A voxelwise map of WM volume revealed widespread, predominantly frontal and subcortical, higher regional WM volume in the NeuroPASC group (red-yellow; P = .005, by using TFCE FWE-correction (right).

WM Regional Cerebral Volume

A related voxelwise analysis of WM volume revealed predominantly frontal higher regional WM volume in the NeuroPASC group by using threshold-free cluster enhancement (TFCE) FWE-corrected, P < .005 (Fig 3, right).

GM Regional Cerebral Volume

GM voxelwise analysis revealed clusters of higher GM volume in the NeuroPASC group, including orbitofrontal cortex and the anterior cingulate (TFCE FWE-corrected, P < .05) (Fig 4).

FIG 4.

FIG 4.

Open in a new tab

In NeuroPASC participants, regional GM volume was higher in the prefrontal cortex. Voxelwise analysis of group differences in GM volume (red-yellow) are superimposed on the group mean T1-weighted structural images. Suprathreshold voxels that represent areas of higher GM volume in NeuroPASC compared with NoCOVID (P = .005, TFCE FWE-corrected) are shown in coronal slices with MNI coordinates and participant left on viewer left. The color bar shows T statistic values.

WM Regional Microstructure

To investigate possible WM microstructural effects consistent with neuroinflammation, we used TBSS voxelwise analysis, finding that WM mean kurtosis was higher in the NeuroPASC group. Areas of higher mean kurtosis (red-yellow) superimposed on the DTI fractional anisotropy skeleton (blue) represent areas of higher WM mean kurtosis in NeuroPASC compared with NoCOVID group (TFCE FWE-corrected, P < .05) (Fig 5, axial view—left, coronal view—right).

FIG 5.

FIG 5.

Open in a new tab

In NeuroPASC participants, WM mean kurtosis is significantly higher. A voxelwise analysis demonstrates group differences. Areas of higher mean kurtosis (red-yellow) superimposed on the DTI fractional anisotropy skeleton (blue) represent areas of higher WM mean kurtosis in NeuroPASC compared with NoCOVID (left) axial view, and (right) coronal view. The critical threshold was (P = .05, TFCE FWE-corrected).

We also investigated the relationship between psychometric attention measures and self-reported clinical severity in both groups. Higher PASC scores were negatively associated with psychometric measures, including significantly lower processing speed, lower concentration, and fewer correct responses (Fig 6).

FIG 6.

FIG 6.

Open in a new tab

Higher clinical PASC scores were associated with significantly lower processing speed (A: β = −0.30; 95% CI: −0.51–−0.09; t = −2.80), lower concentration (B: β = −0.22; 95% CI: −0.43–−0.01; t = −2.09), and fewer correct responses (C: β = −0.27; 95% CI: −0.48–−0.06; t = −2.60).

Plasma Cytokine/Chemokine Profiles

Previous studies have demonstrated different cytokine/chemokine profiles in PASC, also known as long-COVID.20,33 Nevertheless, it remains unclear whether these cytokine/chemokine profile differences are seen in NeuroPASC. To investigate this, we measured a panel of plasma cytokines. A multivariate repeated measures profile analysis with 1 within-subject factor (cytokine) and 3 between-subject factors (group, age, and sex) revealed a main effect of cytokine (WTS = 542.09, P < .001) and a group x cytokine interaction (WTS = 44.6, P = .046). To facilitate visualization of the group profile differences, normalized concentration values are shown in Fig 7. These results show higher IFN-λ1 and IFN-λ2/3 in NeuroPASC.

FIG 7.

FIG 7.

Open in a new tab

Normalized cytokine profiles differ between the NeuroPASC and NoCOVID groups (P <.05). IFN-λ1, IFN-λ2/3, and IL-1β concentrations were higher in the NeuroPASC group.

DISCUSSION

Summary of Results

In a sample of NeuroPASC and NoCOVID participants, we investigated SARS-CoV-2 effects on GM/WM/CSF volume, WM microstructure, cognitive performance, and peripheral cytokine concentrations. We have 4 main findings. First, our NeuroPASC group had larger global and regional (voxelwise) cerebral WM volume compared with the NoCOVID group. Second, the observed GM differences were of lesser magnitude. Third, the NeuroPASC participants exhibited more self-reported PASC symptoms, more headaches, and lower psychometric performance on measures of attention, concentration, verbal learning, and processing speed. The WM volume differences were largest in the frontal lobe, consistent with the cognitive findings. Fourth, the NeuroPASC group showed widespread higher mean diffusional kurtosis in WM, indicating an increase in microstructural complexity possibly related to neuroinflammatory effects of SARS-CoV-2 infection. Five cytokine concentration profiles differed between the groups, suggesting that immune dysregulation following SARS-CoV-2 infection might result in cerebral WM macrostructural and microstructural injury causing NeuroPASC clinical signs and symptoms. NeuroPASC participants were seen a mean of 322 days after infection, demonstrating the persistent nature of these effects.

Total Cerebral WM Volume Differences

Our NeuroPASC group had larger total WM and GM volumes that were accompanied by smaller CSF volumes. The observed larger total WM volume and the higher mean kurtosis in NeuroPASC may reflect greater cellularity in the setting of chronic neuroinflammation, with mildly higher WM/GM volume compensated for by mildly lower CSF volume.34

Regional Cerebral WM Volume Differences

Our NeuroPASC group had a pattern of larger regional WM volume in frontal cortex, including WM in orbitofrontal and primary olfactory cortices. The regional specificity of effects in these areas might be explained by entry of SARS-CoV-2 via the olfactory bulb with subsequent local axonal spread to neighboring cortex.35

Regional Cerebral GM Volume Differences

Larger GM volume and higher cortical thickness in the frontotemporal lobes have been reported in NeuroPASC. Like previous reports in long-COVID,36,37 we also observed larger GM volumes in the NeuroPASC group. Although other studies have reported lower GM volume in acute SARS-CoV-2 infection, these studies did not focus on patients with persistent symptoms after SARS-CoV-2 infection.38,39

WM Microstructure

Using diffusional kurtosis methods, we found widespread higher kurtosis in the WM, indicating greater microstructural complexity. Higher kurtosis has also been reported in the basal ganglia and substantia nigra in Parkinson disease, in various WM regions after concussion or repetitive sub-concussive impacts, following acute ischemia, and in higher grade gliomas.40–45 Potential physiologic mechanisms underlying higher kurtosis include increased cellularity, neurite beading, and cellular swelling.46,47 Greater microstructural complexity has been reported with microglial and astrocytic proliferation occurring after brain injury.48,49 Astrocytic and microglial activation has been found on postmortem examination of patients with COVID-19, albeit not specifically in patients with NeuroPASC.50,51

DKI is sensitive to microstructural changes associated with reactive astrogliosis, which may be missed by standard DTI measures.49 Microglia, the brain’s resident immune cells, have shown pronounced sensitivity to the systemic immune changes in SARS-CoV-2 infection, particularly in the WM. Preclinical evidence includes WM selective microglial reactivity seen 7 weeks after mild SARS-CoV-2 infection in mice.52 Postmortem human brain specimens reveal elevated subcortical and hippocampal WM microglial activity following mild/moderate SARS-CoV-2 infection.52 In vivo evidence includes WM microstructure alterations detected with diffusion MRI in the subacute to chronic phase after SARS-CoV-2 infection, even in mild/moderate acute infection.37,38,53,54 Notably, WM injury has been linked to depression, cognitive decline, and fatigue in other neuropsychiatric conditions.9–12,55,56

Our findings agree with other studies that found evidence of fluid shifts in patients with persistent neuropsychiatric symptoms following SARS-CoV-2 infection. Using diffusion MRI, Rau et al22 found widespread fluid shifts from the intra- and extra-axonal space into the free water fraction in the supratentorial WM in patients with NeuroPASC. These WM fluid shifts accompanied cognitive impairment.

Peripheral Cytokines

There is growing evidence that persisting signs and symptoms after SARS-CoV-2 infection may be related to immune dysregulation, with cytokine profile differences reported up to 8 months after infection.20,21,52 Cytokine concentration differences, including IFN-λ1 and IFN-λ2/3, have been reported in long-COVID/postacute sequelae of COVID-19.20,21 ILs and INFs are mainly secreted by monocytes, macrophages, and helper T-cells following inflammation. These proteins then activate processes critical for controlling viral infections, including antigen presentation, NK cell activity, and T-cell expansion, differentiation, and function.57 The elevated cytokine concentrations observed in acute SARS-CoV-2 infection are more pronounced compared with other viral infections.20

Treatment Implications

Identifying WM effects of SARS-CoV-2 infection suggests potential biologic mechanisms contributing to the NeuroPASC syndrome. If the WM microstructural concomitants of NeuroPASC can be isolated, their nature may suggest underlying cellular mechanisms. If so, experimental treatments could be directed toward specific neuropathophysiologic mechanisms. For example, recovery of WM is far more likely when axons are preserved.58 Targeted treatments have improved damaged WM macro- and microstructure in diseases such as multiple sclerosis, systemic lupus erythematosus, AIDS dementia complex, vitamin B12 deficiency, traumatic brain injury, and hypertension.59–62 Remediating the effects of WM damage has potential ramifications for treatment and prevention of cognitive dysfunction.

Limitations

Our study has some limitations. First, some participants were referred from a post-COVID clinic with clinically diagnosed long-COVID. This may have led to higher observed clinical severity in our sample. Second, our limited sample size may have led to type II errors in detecting associations with some of the tested cytokines. Third, although we found associations among WM volume, microstructure, and PASC clinical severity, stronger evidence for causal effects of infection must await future longitudinal study. Fourth, we did not collect an asymptomatic infected control group for comparison. Finally, cytokine concentrations were likely affected by unaccounted interplay among environmental factors and genetic variation. These limitations and the generalizability of the main results can best be addressed with a larger, longitudinal sample.

CONCLUSIONS

NeuroPASC participants have larger WM volumes than NoCOVID controls and these larger WM volumes are associated with higher PASC scores and greater psychometric deficits. The WM effects may be related to neuroinflammation, with attendant inflammatory microstructural changes and extracellular fluid shifts, a conjecture that can be confirmed with further studies employing free water measures obtained by using multi shell diffusion and myelin water imaging.

Supplementary Material

0449.pdf
0449.pdf (520.9KB, pdf)

Acknowledgments

The authors thank the participants for their time and effort.

The authors would like to acknowledge the support of University of Maryland Medicine (UMM) Biorepository, Baltimore, Maryland, in conducting this study.

ABBREVIATIONS:

COVID

SARS-CoV-2 infection

DKI

diffusional kurtosis imaging

dMRI

diffusion MRI

FA

fractional anisotropy

FWE

family-wise error

IL

interleukin

IFN

interferon

NeuroPASC

neurologic postacute sequelae of SARS-CoV-2 infection

NeuroQOL

neuro-quality of life

PASC

postacute sequelae of SARS-CoV-2 infection

SD

standard deviation

TFCE

threshold-free cluster enhancement

WTS

Wald-type statistic

Footnotes

Erin E. O’Connor and Thomas A. Zeffiro contributed equally to this work.

This work was supported by K23MH118070l University of Maryland Baltimore, Institute for Clinical & Translational Research, and the National Center for Advancing Translational Sciences (NCATS) CTSA grant number 1UL1TR003098.

Disclosure forms provided by the authors are available with the full text and PDF of this article at www.ajnr.org.

References

  • 1.Taquet M, Geddes JR, Husain M, et al. 6-month neurological and psychiatric outcomes in 236 379 survivors of COVID-19: a retrospective cohort study using electronic health records. Lancet Psychiatry 2021;8:416–27 10.1016/S2215-0366(21)00084-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Thaweethai T, Jolley SE, Karlson EW; RECOVER Consortium, et al. Development of a definition of postacute sequelae of SARS-CoV-2 infection. JAMA 2023;329:1934–46 10.1001/jama.2023.8823 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Schou TM, Joca S, Wegener G, et al. Psychiatric and neuropsychiatric sequelae of COVID-19—a systematic review. Brain Behav Immun 2021;97:328–48 10.1016/j.bbi.2021.07.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Taquet M, Sillett R, Zhu L, et al. Neurological and psychiatric risk trajectories after SARS-CoV-2 infection: an analysis of 2-year retrospective cohort studies including 1 284 437 patients. Lancet Psychiatry 2022;9:815–27 10.1016/S2215-0366(22)00260-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Filley CM. White matter dementia. Ther Adv Neurol Disord 2012;5:267–77 10.1177/1756285612454323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Filley CM. The behavioral neurology of cerebral white matter. Neurology 1998;50:1535–40 10.1212/wnl.50.6.1535 [DOI] [PubMed] [Google Scholar]
  • 7.Gibson EM, Nagaraja S, Ocampo A, et al. Methotrexate chemotherapy induces persistent tri-glial dysregulation that underlies chemotherapy-related cognitive impairment. Cell 2019;176:43–55.e13 10.1016/j.cell.2018.10.049 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Schmahmann JD, Smith EE, Eichler FS, et al. Cerebral white matter: neuroanatomy, clinical neurology, and neurobehavioral correlates. Ann N Y Acad Sci 2008;1142:266–309 10.1196/annals.1444.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Kinnunen KM, Greenwood R, Powell JH, et al. White matter damage and cognitive impairment after traumatic brain injury. Brain 2011;134:449–63 10.1093/brain/awq347 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ungvari Z, Toth P, Tarantini S, et al. Hypertension-induced cognitive impairment: from pathophysiology to public health. Nat Rev Nephrol 2021;17:639–54 10.1038/s41581-021-00430-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.van Velzen LS, Kelly S, Isaev D, et al. White matter disturbances in major depressive disorder: a coordinated analysis across 20 international cohorts in the ENIGMA MDD working group. Mol Psychiatry 2020;25:1511–25 10.1038/s41380-019-0477-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Hickie I, Scott E, Wilhelm K, et al. Subcortical hyperintensities on magnetic resonance imaging in patients with severe depression–a longitudinal evaluation. Biol Psychiatry 1997;42:367–74 10.1016/S0006-3223(96)00363-0 [DOI] [PubMed] [Google Scholar]
  • 13.Catani M, Ffytche DH. The rises and falls of disconnection syndromes. Brain 2005;128:2224–39 10.1093/brain/awh622 [DOI] [PubMed] [Google Scholar]
  • 14.Radmanesh A, Derman A, Lui YW, et al. COVID-19-associated diffuse leukoencephalopathy and microhemorrhages. Radiology 2020;297:e223–27 10.1148/radiol.2020202040 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Wang Y, Wang Y, Huo L, et al. SARS-CoV-2-associated acute disseminated encephalomyelitis: a systematic review of the literature. J Neurol 2022;269:1071–92 10.1007/s00415-021-10771-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Klironomos S, Tzortzakakis A, Kits A, et al. Nervous system involvement in coronavirus disease 2019: results from a retrospective consecutive neuroimaging cohort. Radiology 2020;297:e324–34 10.1148/radiol.2020202791 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Marcic L, Marcic M, Kojundzic SL, et al. Personalized approach to patient with MRI brain changes after SARS-CoV-2 infection. J Pers Med 2021;11:442 10.3390/jpm11060442 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Gershon RC, Lai JS, Bode R, et al. Neuro-QOL: quality of life item banks for adults with neurological disorders: item development and calibrations based upon clinical and general population testing. Qual Life Res 2012;21:475–86 10.1007/s11136-011-9958-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. App Psychol Meas 1977;1:385–401 10.1177/014662167700100306 [DOI] [Google Scholar]
  • 20.Phetsouphanh C, Darley DR, Wilson DB, et al. Immunological dysfunction persists for 8 months following initial mild-to-moderate SARS-CoV-2 infection. Nat Immunol 2022;23:210–16 10.1038/s41590-021-01113-x [DOI] [PubMed] [Google Scholar]
  • 21.Schultheiß C, Willscher E, Paschold L, et al. The IL-1β, IL-6, and TNF cytokine triad is associated with post-acute sequelae of COVID-19. Cell Rep Med 2022;3:100663 10.1016/j.xcrm.2022.100663 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Rau A, Schroeter N, Blazhenets G, et al. Widespread white matter oedema in subacute COVID-19 patients with neurological symptoms. Brain 2022;145:3203–13 10.1093/brain/awac045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Soriano JB, Murthy S, Marshall JC, WHO Clinical Case Definition Working Group on Post-COVID-19 Condition, et al. A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect Dis 2022;22:e102–07 10.1016/S1473-3099(21)00703-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Swartz MD, DeSantis SM, Yaseen A, et al. Antibody duration after infection from SARS-CoV-2 in the Texas coronavirus antibody response survey. J Infect Dis 2023;227:193–201 10.1093/infdis/jiac167 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gaser C, Dahnke R, Thompson PM, et al. CAT–a computational anatomy toolbox for the analysis of structural MRI data. Gigascience 2024;13:giae049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.O’Connor EE, Zeffiro TA, Lopez OL, et al. Differential effects of an AIDS defining illness and chronic HIV infection on gray matter volume. Clin Infect Dis 2021;73:e2303–107 10.1093/cid/ciaa1552 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Dhiman S, Hickey RE, Thorn KE, et al. PyDesigner v1.0: a pythonic implementation of the DESIGNER pipeline for diffusion magnetic resonance imaging. J Vis Exp 2024;207. 10.3791/66397 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Fazekas F, Chawluk JB, Alavi A, et al. MR signal abnormalities at 1.5 T in Alzheimer’s dementia and normal aging. AJR Am J Roentgenol 1987;149:351–56 10.2214/ajr.149.2.351 [DOI] [PubMed] [Google Scholar]
  • 29.Ding T, Cohen AD, O’Connor EE, et al. An improved algorithm of white matter hyperintensity detection in elderly adults. Neuroimage Clin 2020;25:102151 10.1016/j.nicl.2019.102151 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Mori S, Oishi K, Jiang H, et al. Stereotaxic white matter atlas based on diffusion tensor imaging in an ICBM template. NeuroImage 2008;40:570–82 10.1016/j.neuroimage.2007.12.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Jensen JH, Helpern JA, Ramani A, et al. Diffusional kurtosis imaging: the quantification of non-Gaussian water diffusion by means of magnetic resonance imaging. Magn Reson Med 2005;53:1432–40 10.1002/mrm.20508 [DOI] [PubMed] [Google Scholar]
  • 32.Jensen JH, Helpern JA. MRI quantification of non-Gaussian water diffusion by kurtosis analysis. NMR Biomed 2010;23:698–710 10.1002/nbm.1518 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Low RN, Low RJ, Akrami A. A review of cytokine-based pathophysiology of long COVID symptoms. Front Med (Lausanne) 2023;10:1011936 10.3389/fmed.2023.1011936 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ho ML, Rojas R, Eisenberg RL. Cerebral edema. AJR Am J Roentgenol 2012;199:W258–73 10.2214/AJR.11.8081 [DOI] [PubMed] [Google Scholar]
  • 35.de Melo GD, Perraud V, Alvarez F, et al. Neuroinvasion and anosmia are independent phenomena upon infection with SARS-CoV-2 and its variants. Nat Commun 2023;14:4485 10.1038/s41467-023-40228-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Besteher B, Machnik M, Troll M, et al. Larger gray matter volumes in neuropsychiatric long-COVID syndrome. Psychiatry Res 2022;317:114836 10.1016/j.psychres.2022.114836 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Lu Y, Li X, Geng D, et al. Cerebral micro-structural changes in COVID-19 patients - an MRI-based 3-month follow-up study. EClinicalMedicine 2020;25:100484 10.1016/j.eclinm.2020.100484 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Qin Y, Wu J, Chen T, et al. Long-term microstructure and cerebral blood flow changes in patients recovered from COVID-19 without neurological manifestations. J Clin Invest 2021;131:e147329 10.1172/JCI147329 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Douaud G, Lee S, Alfaro-Almagro F, et al. SARS-CoV-2 is associated with changes in brain structure in UK Biobank. Nature 2022;604:697–707 10.1038/s41586-022-04569-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Davenport EM, Apkarian K, Whitlow CT, et al. Abnormalities in diffusional kurtosis metrics related to head impact exposure in a season of high school varsity football. J Neurotrauma 2016;33:2133–46 10.1089/neu.2015.4267 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Chung S, Chen J, Li T, et al. Investigating brain white matter in football players with and without concussion using a biophysical model from multishell diffusion MRI. AJNR Am J Neuroradiol 2022;43:823–28 10.3174/ajnr.A7522 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Alves R, Henriques RN, Kerkelä L, et al. Correlation tensor MRI deciphers underlying kurtosis sources in stroke. Neuroimage 2022;247:118833 10.1016/j.neuroimage.2021.118833 [DOI] [PubMed] [Google Scholar]
  • 43.Wang JJ, Lin WY, Lu CS, et al. Parkinson disease: diagnostic utility of diffusion kurtosis imaging. Radiology 2011;261:210–17 10.1148/radiol.11102277 [DOI] [PubMed] [Google Scholar]
  • 44.Rong Y, Xu Z, Zhu Y, et al. Combination of quantitative susceptibility mapping and diffusion kurtosis imaging provides potential biomarkers for early-stage Parkinson’s disease. ACS Chem Neurosci 2022;13:2699–708 10.1021/acschemneuro.2c00321 [DOI] [PubMed] [Google Scholar]
  • 45.Shi W, Qu C, Wang X, et al. Diffusion kurtosis imaging combined with dynamic susceptibility contrast-enhanced MRI in differentiating high-grade glioma recurrence from pseudoprogression. Eur J Radiology 2021;144:109941 10.1016/j.ejrad.2021.109941 [DOI] [PubMed] [Google Scholar]
  • 46.Budde MD, Frank JA. Neurite beading is sufficient to decrease the apparent diffusion coefficient after ischemic stroke. Proc Natl Acad Sci U S A 2010;107:14472–77 10.1073/pnas.1004841107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Raab P, Hattingen E, Franz K, et al. Cerebral gliomas: diffusional kurtosis imaging analysis of microstructural differences. Radiology 2010;254:876–81 10.1148/radiol.09090819 [DOI] [PubMed] [Google Scholar]
  • 48.Steven AJ, Zhuo J, Melhem ER. Diffusion kurtosis imaging: an emerging technique for evaluating the microstructural environment of the brain. AJR Am J Roentgenol 2014;202:W26–33 10.2214/AJR.13.11365 [DOI] [PubMed] [Google Scholar]
  • 49.Zhuo J, Xu S, Proctor JL, et al. Diffusion kurtosis as an in vivo imaging marker for reactive astrogliosis in traumatic brain injury. NeuroImage 2012;59:467–77 10.1016/j.neuroimage.2011.07.050 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Thakur KT, Miller EH, Glendinning MD, et al. COVID-19 neuropathology at Columbia University Irving Medical Center/New York Presbyterian Hospital. Brain 2021;144:2696–708 10.1093/brain/awab148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Matschke J, Lütgehetmann M, Hagel C, et al. Neuropathology of patients with COVID-19 in Germany: a post-mortem case series. Lancet Neurol 2020;19:919–29 10.1016/S1474-4422(20)30308-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Fernández-Castañeda A, Lu P, Geraghty AC, et al. Mild respiratory COVID can cause multi-lineage neural cell and myelin dysregulation. Cell 2022;185:2452–2468.e16 10.1016/j.cell.2022.06.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Benedetti F, Palladini M, Paolini M, et al. Brain correlates of depression, post-traumatic distress, and inflammatory biomarkers in COVID-19 survivors: a multimodal magnetic resonance imaging study. Brain Behav Immun Health 2021;18:100387 10.1016/j.bbih.2021.100387 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Petersen M, Nägele FL, Mayer C, et al. Brain imaging and neuropsychological assessment of individuals recovered from a mild to moderate SARS-CoV-2 infection. Proc Natl Acad Sci U S A 2023;120:e2217232120 10.1073/pnas.2217232120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Shan ZY, Kwiatek R, Burnet R, et al. Progressive brain changes in patients with chronic fatigue syndrome: a longitudinal MRI study. J Magn Reson Imaging 2016;44:1301–11 10.1002/jmri.25283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Papadopoulou A, Müller-Lenke N, Naegelin Y, et al. Contribution of cortical and white matter lesions to cognitive impairment in multiple sclerosis. Mult Scler 2013;19:1290–96 10.1177/1352458513475490 [DOI] [PubMed] [Google Scholar]
  • 57.Dowling JW, Forero A. Beyond good and evil: molecular mechanisms of type I and III IFN functions. J Immunol 2022;208:247–56 10.4049/jimmunol.2100707 [DOI] [PubMed] [Google Scholar]
  • 58.Medana IM, Esiri MM. Axonal damage: a key predictor of outcome in human CNS diseases. Brain 2003;126:515–30 10.1093/brain/awg061 [DOI] [PubMed] [Google Scholar]
  • 59.Mulero P, Midaglia L, Montalban X. Ocrelizumab: a new milestone in multiple sclerosis therapy. Ther Adv Neurol Disord 2018;11:1756286418773025 10.1177/1756286418773025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Xu J, Cheng Y, Chai P, et al. White-matter volume reduction and the protective effect of immunosuppressive therapy in systemic lupus erythematosus patients with normal appearance by conventional magnetic resonance imaging. J Rheumatol 2010;37:974–86 10.3899/jrheum.090967 [DOI] [PubMed] [Google Scholar]
  • 61.Blasko I, Hinterberger M, Kemmler G, et al. Conversion from mild cognitive impairment to dementia: influence of folic acid and vitamin B12 use in the VITA cohort. J Nutr Health Aging 2012;16:687–94 10.1007/s12603-012-0051-y [DOI] [PubMed] [Google Scholar]
  • 62.Mayer AR, Ling J, Mannell MV, et al. A prospective diffusion tensor imaging study in mild traumatic brain injury. Neurology 2010;74:643–50 10.1212/WNL.0b013e3181d0ccdd [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

0449.pdf
0449.pdf (520.9KB, pdf)

Articles from AJNR: American Journal of Neuroradiology are provided here courtesy of American Society of Neuroradiology

RESOURCES