Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2024 Mar 26;45(5):e26656. doi: 10.1002/hbm.26656

Gray matter atrophy and white matter lesions burden in delayed cognitive decline following carbon monoxide poisoning

Yanli Zhang 1,2,3,4, Tianhong Wang 5, Shuaiwen Wang 1,2,3,4, Xin Zhuang 1,2,3,4, Jianlin Li 1,2,3,4, Shunlin Guo 1,2,3,4, Junqiang Lei 1,2,3,4,✉
PMCID: PMC10964793  PMID: 38530116

Abstract

Gray matter (GM) atrophy and white matter (WM) lesions may contribute to cognitive decline in patients with delayed neurological sequelae (DNS) after carbon monoxide (CO) poisoning. However, there is currently a lack of evidence supporting this relationship. This study aimed to investigate the volume of GM, cortical thickness, and burden of WM lesions in 33 DNS patients with dementia, 24 DNS patients with mild cognitive impairment, and 51 healthy controls. Various methods, including voxel‐based, deformation‐based, surface‐based, and atlas‐based analyses, were used to examine GM structures. Furthermore, we explored the connection between GM volume changes, WM lesions burden, and cognitive decline. Compared to the healthy controls, both patient groups exhibited widespread GM atrophy in the cerebral cortices (for volume and cortical thickness), subcortical nuclei (for volume), and cerebellum (for volume) (p < .05 corrected for false discovery rate [FDR]). The total volume of GM atrophy in 31 subregions, which included the default mode network (DMN), visual network (VN), and cerebellar network (CN) (p < .05, FDR‐corrected), independently contributed to the severity of cognitive impairment (p < .05). Additionally, WM lesions impacted cognitive decline through both direct and indirect effects, with the latter mediated by volume reduction in 16 subregions of cognitive networks (p < .05). These preliminary findings suggested that both GM atrophy and WM lesions were involved in cognitive decline in DNS patients following CO poisoning. Moreover, the reduction in the volume of DMN, VN, and posterior CN nodes mediated the WM lesions‐induced cognitive decline.

Keywords: carbon monoxide poisoning, cognitive decline, delayed neurological sequelae, gray matter atrophy, neuroimaging analysis, white matter lesions

The delayed neurologic sequelae (DNS) patients with cognitive impairment following carbon monoxide poisoning showed a widespread gray matter (GM) atrophy in cortical and subcortical regions as well as the cerebellum. The subregions of the bilateral prefrontal and visual cortices, and cerebellar lobule VI atrophy mediate white matter lesions‐induced cognitive decline.

graphic file with name HBM-45-e26656-g002.jpg

Abbreviations

CO

carbon monoxide

DBM

deformation‐based morphometry

DNS

delayed neurologic sequelae

GM

gray matter

VBM

voxel‐based morphometry

WMH

white matter hyperintensities

1. INTRODUCTION

Cognitive decline is a prevalent symptom in patients suffering from delayed neurological sequelae (DNS) following carbon monoxide (CO) poisoning. The severity of cognitive decline ranges from dementia to mild cognitive impairment (MCI). A common imaging marker for diagnosing DNS is the presence of diffuse white matter hyperintensities (WMH) lesions in the centrum semiovale and periventricular area on T2‐weighted images (Beppu, 2014). Current research indicates a relationship between white matter (WM) lesions, WM disconnections, and cognitive decline in DNS patients (Chou et al., 2019; Fujiwara et al., 2012; Jiang et al., 2021; Parkinson et al., 2002). However, the cognitive impairment observed in DNS patients cannot be entirely accounted for by WM damage alone.

GM in the brain plays a crucial role in cognitive processes by facilitating information extraction, exchange, and integration. In a previous study utilizing voxel‐based morphometry (VBM) analysis, we observed a reduction in volume in the medial orbital superior frontal gyrus (SFG) and anterior cingulate cortex (ACC), which correlated with lower Mini‐Mental State Examination (MMSE) scores in patients with CO poisoning (Zhang et al., 2023). Another VBM study reported reduced gray matter (GM) volume in the bilateral basal ganglia, left postcentral gyrus, and left hippocampus, which was associated with decreased cognition in CO poisoning patients (Chen et al., 2013). Furthermore, our previous surface‐based morphometry (SBM) analysis demonstrated a significantly thinner cortical thickness in the SFG and rostral middle frontal gyrus (MFG), which was also related to lower MMSE scores in CO poisoning patients (Wang et al., 2023). These findings emphasize the crucial role of GM atrophy in the cognitive decline observed in patients with CO poisoning. However, the inconsistent results regarding GM atrophy in these studies may be attributed to inherent patient heterogeneity, variations in analysis methods and statistical thresholds, as well as differences in data pre‐processing pipelines. Currently, there is a lack of investigation combining GM volume and cortical thickness in DNS patients with dementia (DNS‐D) and MCI.

Two main methods for analyzing structural magnetic resonance imaging (MRI) data are voxel‐level and regional‐level analyses. The former includes techniques such as VBM, deformation‐based morphometry (DBM), and SBM. On the other hand, the latter relies on brain atlases, providing precise and direct information regarding GM alterations in specific subregions. Consequently, the regional‐level analysis offers excellent anatomical interpretability (Long et al., 2018). Combining these analysis methods at different levels can provide a comprehensive understanding of GM structural alterations. In the field of neuropsychiatric disorder research, two prominent brain atlases, the automated anatomical labeling (AAL) atlas (Rolls et al., 2020) and the Human Brainnetome (BN) Atlas (Fan et al., 2016), have gained increasing popularity. Studies have demonstrated that the BN atlas outperforms the AAL atlas in the classification of patients with neurological disorders, showcasing its superior classification performance (Shi et al., 2022).

Previous studies have provided evidence linking GM atrophy and WMH burden to cognitive decline in different populations affected by WMH, such as aging populations, patients with Alzheimer's disease (AD), or vascular dementia. Specifically, regional GM atrophy has been identified as a mediating factor contributing to WMH‐induced cognitive impairment (Jokinen et al., 2020; Rizvi et al., 2021; Tuladhar et al., 2015; Zhu et al., 2021). In the case of CO poisoning patients with WMH lesions, their GM atrophy may stem from direct neuronal hypoxic‐ischemia, secondary WM fiber disruption, or possibly both. The relationship among WM lesions burden, GM structural changes, and cognitive impairment in DNS patients following CO poisoning remains uncertain.

Therefore, in this study, we hypothesized that GM atrophy and WM lesions collectively contribute to cognitive decline in DNS patients, with regional GM atrophy potentially serving as a mediating factor in WM lesions‐induced cognitive impairment. To comprehensively investigate the pattern of GM structural alterations in these patients and enable direct comparisons across different measurement methods, we conducted a battery of analyses. This included VBM and DBM for volume analysis, SBM for cortical thickness evaluation, and BN atlas‐based methods for both volume and cortical thickness assessments, using the same group of patients and controls. Additionally, we further validated the reproducibility of our main results by conducting a verification analysis using the AAL3 atlas.

2. MATERIALS AND METHODS

2.1. Patient enrollment

This retrospective observational study initially recruited a total of 118 adult individuals. Among them, 67 were consecutive patients, and the remaining 51 were healthy controls (HCs). To ensure comparability, the HCs were matched to the patients based on sex, age, and education level. Data for all participants were collected at the Department of Neurology, the first hospital of Lanzhou University, from May 2018 to December 2022. The study received approval from the Ethics Committee of the first hospital of Lanzhou University (ID: LDYYLL2018‐114), and informed consent was obtained from all participants involved.

All participants underwent clinical interviews and neuropsychological tests, including the MMSE, Montreal Cognitive Assessment (MoCA), and Barthel index (BI). Patients were selected based on the following criteria: (1) We selected a group of adults aged 18 to younger than 65 years, as studies have reported that the prevalence of MCI is 21.0% and dementia is 5.1% among elderly Chinese individuals aged 65 years or older (Li et al., 2016). (2) Patients had a clear history of recent CO exposure, with initial evidence of an elevated carboxyhemoglobin level greater than 10% upon first admission (Rose et al., 2017). (3) The diagnosis of DNS was confirmed by a neurologist (T.W., with 14 years of experience in neurology) based on recurrent neurological symptoms that occurred after an apparent period of recovery. These symptoms included cognitive dysfunction, mental deterioration, mutism, gait disturbance, urinary or fecal incontinence, psychosis, depression, and Parkinsonism (Jeon et al., 2018). (4) Patients had an initial MRI image taken during acute CO poisoning. (5) The presence of DNS‐related WMH lesions on T2‐weighted fluid‐attenuated inversion recovery (T2w‐FLAIR) images, as diagnosed by two radiologists (Y.Z., with 9 years of experience, and S.W., with 10 years of experience in radiology). DNS‐related WMH was defined as the identification of additional lesions on follow‐up MRI images compared to the initial magnetic resonance (MR) images taken during acute CO poisoning. (6) Patients exhibited symptoms of cognitive impairment. Dementia associated with DNS was diagnosed according to the criteria of the National Institute of Aging and Alzheimer's Association (McKhann et al., 2011) and an MMSE score <17 (illiterate subjects), <20 (grade school literacy), or <24 (junior high school and higher education literacy) (Li et al., 2016). MCI associated with DNS was defined as a MoCA score of less than 26, with essentially preserved daily and social functioning and no dementia. For patients with less than 12 years of education, one point was added to the final MoCA score, following the principle of education correction. The HCs had no history of neuropsychiatric disorders or cognitive complaints.

The exclusion criteria were as follows: (1) Left‐handedness. (2) History of nervous system disorders or other systemic diseases, including AD, Parkinsonism, depression, traumatic brain injury, cranial brain operation, stroke, infection, tumor, multiple sclerosis, metabolic disease, other poisoning, and long‐term drug treatment known to induce cognitive decline. (3) Patients with a total Fazekas score of ≥3 (sum of the deep WMH and periventricular WMH Fazekas score ≥3) on the initial MR images were excluded, as severe WMH with such a score in normal adults has been linked to poor cognitive performance (Fazekas et al., 1987; Zeng et al., 2020). (4) Patients who displayed insufficient cooperation during the neuropsychological tests were also excluded. In total, six patients were excluded: one had an unexpected arachnoid cyst on MRI, four had a total Fazekas score of ≥3 for WMH, and one was unable to cooperate with the clinical tests. Additionally, three HCs were excluded due to significant WMH with a total Fazekas score of ≥3. Consequently, the final sample consisted of 61 patients with DNS and 51 HCs for further data pre‐processing.

2.2. MRI protocols and data pre‐processing

All images were acquired using the same 3T MRI scanner (Magneton Skyra; Siemens, Erlangen, Germany) equipped with a 32‐channel phased‐array head coil. The MRI data included 3D T1‐weighted and T2w‐FLAIR images.

For analysis of the 3D T1‐weighted data, the computational anatomy toolbox 12 (CAT12) toolbox (Computational Anatomy Toolbox; http://www.neuro.uni-jena.de/cat) implemented in SPM12 (Statistical Parametric Mapping, Institute of Neurology, London, UK; http://www.fl.ion.ucl.ac.uk/spm/software/spm12/) was utilized. This voxel‐based analysis approach involved methods such as VBM, DBM, and SBM, running on MATLAB 2018a (MathWorks, Natick, MA, USA). The detailed MRI acquisition and data processing procedures are described in Supporting Information.

After the completion of the pre‐processing pipeline, a sample of 33 data from patients with DNS‐D, 26 data from patients with DNS patients with mild cognitive impairment (DNS‐MCI), and 51 data from HCs with a weighted average score of B or higher (based on quality reports generated by CAT12) were selected for subsequent statistical analysis. The WMH volume was manually segmented by two independent raters (Y. Z. and S.W.), who were blinded to the clinical information from the T2w‐FLAIR images of each subject using ITK‐SNAP 3.8.0 (http://www.itksnap.org/pmwiki/pmwiki.php). The total WMH volume was calculated as the sum of WMH across all image layers. The intraclass correlation efficiency of the two measurements was 0.953 (p < .001). The mean WMH volume from the two measurements was log‐transformed to achieve a normal distribution, with higher values indicating a greater burden of WMH. For each patient, all the segmented WMH from each image layer were saved as one lesion mask. The MRIcron toolbox (University of Nottingham School of Psychology, Nottingham, UK) was used to overlay the lesion masks of all patients onto Montreal Neurological Institute (MNI) space, producing lesion frequency heatmaps for each patient group.

2.3. Statistical analyses

2.3.1. Demographic, neuropsychological, and structural MRI data

Statistical analyses were conducted using SPSS 22 (IBM Corp., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation or as median and interquartile range, depending on the normality assessment using the Shapiro–Wilk test. Differences in age, total intracranial volume (TIV), and WM lesion volume were assessed using one‐way analysis of variance (ANOVA) with post hoc analysis. Differences in education level, MMSE score, MoCA score, and BI score were analyzed using the Kruskal–Wallis H test with Bonferroni correction for post hoc analysis. Two independent‐sample t‐tests were performed to examine differences in the time interval between CO poisoning and WMH burden. The chi‐square test was employed to evaluate the distribution of sex among the three groups. The threshold for statistical significance was set at p < .05.

2.3.2. Voxel‐, deformation‐, and vertex‐based analyses

Significant differences in GM volume determined through VBM and DBM analyses, or cortical thickness identified through SBM analysis, were evaluated using one‐way ANOVA. Post hoc t contrasts were conducted to perform specific pairwise comparisons between DNS‐D and HCs, DNS‐MCI and HCs, and DNS‐D and DNS‐MCI, within the CAT12 framework. Covariates included age, sex, years of education, and the time interval between CO poisoning and MRI scans (with the addition of TIV for GM volume analysis). All statistical analyses were conducted within the GM mask, and the results were assessed at a peak threshold of p < .05, after applying false discovery rate (FDR) correction for multiple comparisons and a cluster size threshold of 100 voxels.

2.3.3. Atlas‐based analysis

Statistical analyses using the region of interest method were conducted with the combined atlas comprised of the Human Brainnetome Atlas (consisting of 210 cortical and 36 subcortical subregions in the cerebrum) (Fan et al., 2016) and the Probabilistic Atlas of the Human Cerebellum (consisting of 28 subregions) (Diedrichsen et al., 2009) within the CAT12 framework. All 274 regions of GM volume and 210 regions of cortical thickness were examined. One‐way ANOVA and post hoc t‐tests (with FDR correction and a significance level of p < .05) were conducted to identify regional differences in GM volume and cortical thickness using CAT12. Age, sex, years of education, the time interval between CO poisoning and MRI scans, and TIV for VBM analysis were included as confounding variables. In cases where significant GM volume and/or cortical thickness abnormalities were found among the groups based on the Brainnetome atlas analysis, a complementary analysis using the AAL3‐based atlas (Rolls et al., 2020) was performed to validate the main results, utilizing the same statistical threshold.

2.3.4. Regression and mediation analyses

According to the results of volumetric analysis using BN‐based GM, the total volume of GM atrophy (TGMAV) was calculated for each patient by summing the volumes of all identified atrophic subregions when comparing between‐patient groups. To investigate the relationship between TGMAV, burden of WM lesions, and cognition, multiple linear regression analysis was conducted while controlling for confounding variables mentioned above. Model 1 considered only the burden of WM lesions as predictor variables and cognitive score (measured by MMSE or MoCA) as the dependent variable; model 2 considered only the TGMAV as a predictor variable. Finally, in model 3, WM lesions burden, TGMAV, and an interaction term (WM lesions burden + TGMAV) were included as predictor variables.

To examine the role of GM atrophy in the relationship between the burden of WM lesions and cognitive decline, mediation analyses were conducted using the PROCESS Macro in SPSS. Mediator variable “M” included altered GM measures (TGMAV and volume of each identified atrophic subregion) obtained from the analysis comparing the DNS‐D and DNS‐MCI groups. The independent variable “X” represented the total volume of white matter hyperintensities (WMH), while the dependent variable “Y” represented either MMSE or MoCA scores. The analysis controlled for the same confounding variables mentioned previously. Log transformation was applied to the total WMH volume for better normalization. Each mediation model assessed the relationships between WMH volume and cognitive performance (total effect, path c), WMH volume and GM measures (path a), and GM measures and cognitive performance (path b). The bootstrap method (n = 5000) was used to evaluate the mediation analyses, with significant indirect effects defined as a 95% confidence interval that did not include 0.

3. RESULTS

3.1. Participant characteristics

Demographic, structural MRI, and cognitive characteristics are summarized in Table 1. No significant differences were observed between the patient groups in terms of age, gender, education, the interval from CO poisoning to MR scans, and TIV (p > .05). However, the DNS‐D group exhibited a significantly higher burden of WM lesions than the DNS‐MCI and HCs groups (p < .05). Furthermore, both patient groups demonstrated worse cognitive performance compared to the HCs group (p < .05).

TABLE 1.

Demographic, structural MRI, and clinical features.

DNS‐D (n = 33) DNS‐MCI (n = 24) HCs (n = 51) Statistic value Over all p value
Demographic data
Age (years) 52.27 ± 9.62 48.63 ± 8.30 52.49 ± 8.58 F = 1.708 .186
Gender (M/F) 16/17 11/13 27/24 .893
Education (years) 8.00 (4.50) 8.00 (6.00) 11.00 (6.00) H = 5.768 .056
Time interval between CO poisoning to MR scans 48.29 ± 22.02** 53.29 ± 21.49 NA T = −0.806 <.001
Structural MRI features
TIV (cm3) 1422.28 ± 129.57 1478.84 ± 138.08 1420.90 ± 124.33 F = 2.058 .133
WM lesions volume (cm3) 147.71 ± 85.69* , ** 102.77 ± 43.46* 0.115 ± 0.153 F = 89.324 <.001
Cognitive performance
MMSE 17.00 (7.00)* , *** 26.00 (1.00)* 30.00 (0.00) H = 97.221 <.001
MoCA 17.00 (8.00)* , ** 24.50 (1.00)* 30.00 (2.00) H = 94.492 <.001
Barthel index 40.00 (40.00)* , *** 100.00 (10.00) 100.00 (0.00) H = 77.062 <.001
Open in a new tab

Note: Continuous variables are presented as mean ± standard deviation or as median and interquartile range.

Abbreviations: CO, carbon monoxide; DNS‐D, delayed neurological sequelae patients with dementia; DNS‐MCI, delayed neurological sequelae patients with mild cognitive impairment; HCs, healthy controls; MMSE, Mini‐Mental State Examination; MoCA, Montreal Cognitive Assessment; NA, no acquire; TIV, total intracranial volume; WHM, white matter hyperintensities; WM, white matter.

*

p < .001 with HCs group.

**

p < .05.

***

p < .001 with DNS‐MCI group.

3.2. WM lesions frequency heatmap

To visualize the distribution of WMH lesions, a frequency heatmap was generated using mask overlapping. The results revealed a preferential distribution of WMH lesions in various brain regions, including the centrum semiovale, periventricular area, frontal, parietal, occipital, and temporal lobes, as well as the genu and splenium of the corpus callosum, in both the DNS‐D and DNS‐MCI groups (Figure 1).

FIGURE 1.

FIGURE 1

Open in a new tab

WM lesions distribution heatmap for delayed neurological sequelae (DNS) patients with dementia (a) and DNS patients with mild cognitive impairment (b) patients. Voxels are color‐coded based on the number of subjects with white matter lesions in that voxel. The color ranges from purple to red indicates the number of subjects from 2 to 20 and above.

3.3. GM volume and cortical thickness alteration patterns

As depicted in Figure 2, both patient groups exhibited extensive cortical thinning in comparison to the HCs, with the most severe thinning primarily affecting the bilateral frontal, occipital, and parietal lobes (FDR‐corrected p < .05). However, no significant differences in cortical thickness were observed between the patient groups (FDR‐corrected p > .05).

FIGURE 2.

FIGURE 2

Open in a new tab

Cortical thickness alteration patterns identified by atlas‐based and vertex‐based analyses in patients. (a) Delayed neurological sequelae (DNS) patients with dementia versus healthy controls (HCs); (b) DNS patients with mild cognitive impairment versus HCs. p < .05, false discovery rate corrected.

As illustrated in Figure 3a,b, the patient groups exhibited significant GM volume reduction, particularly in multiple subregions of the prefrontal cortices (PFC), ACC, occipital, temporal, and insular cortices, compared to the HCs. Additionally, both patient groups displayed decreased GM volume in various subregions of the bilateral thalamus, caudate nucleus, putamen, hippocampus, and posterior cerebellar lobe in the subcortical and cerebellar regions. Comparatively, the DBM analysis identified more atrophic subregions in the subcortical areas and fewer in the cortical areas compared to the VBM analysis in both patient groups.

FIGURE 3.

FIGURE 3

Open in a new tab

Gray matter volume change patterns in the cortical, subcortical, and cerebellum using different data analysis methods. (a) Delayed neurological sequelae patients with dementia (DNS‐D) versus healthy controls (HCs); (b) DNS patients with mild cognitive impairment (DNS‐MCI) versus HCs; (c) DNS‐D versus DNS‐MCI. BN, Brainnetome. p < .05, false discovery rate corrected.

In contrast to the DNS‐MCI group (Figure 3c), the DNS‐D group exhibited significant cortical GM volume reduction in 31 subregions, encompassing the bilateral PFC, right inferior parietal lobe, bilateral occipitotemporal cortices, and bilateral cerebellum (FDR‐corrected p < .05; Table S1). The BN atlas‐based analysis detected a higher number of regions with reduced GM volume across the whole brain compared to the VBM and DBM analyses.

Furthermore, when comparing different methods of measuring cortical atrophy, we observed a similar distribution pattern but more pronounced cortical thinning in both patient groups, consistent with the findings of volumetric alterations (Figure S1).

3.4. Control results

To validate the reliability of the volume atrophy results observed in the BN atlas‐based analysis, the AAL3 atlas was utilized. Notably, the AAL3 atlas‐based analyses confirmed consistent findings with similar GM volume alteration maps across the three groups (Figure S2). When comparing the patient groups, the atrophic regions identified in the BN atlas‐based analysis largely overlapped with the regions identified by the AAL3 atlas‐based analysis, with the exception of a few temporal subregions.

3.5. Relationship between GM volume alteration, WM lesions burden, and cognitive performance

Regression models 1 and 2 demonstrated that both the burden of WM lesions (β = −4.362, p = .032 for MMSE, and β = −5.438, p = .030 for MoCA) and TGMAV (β = 47.415, p < .001 for MMSE, and β = 59.356, p < .001 for MoCA) were risk factors for lower cognitive scores. Model 2 further revealed that TGMAV independently influenced lower MMSE and MoCA scores (both p < .001). However, no significant interaction was found between the two imaging indices (p = .362).

Subsequent mediation analysis demonstrated that TGMAV played a significant role in mediating the relationship between log‐transformed WM lesion volume and MMSE (with a total effect of β = −6.489, p < .05) and MoCA (with total effect of β = −7.928, p < .05). Specifically, TGMAV displayed a mediation effect (β = −5.020 for MMSE, and β = −6.474 for MoCA), while the direct effects (β = −1.469 for MMSE, and β = −1.454 for MoCA) were not statistically significant. This suggests a full mediation role of TGMAV in the relationship between WM lesion burden and cognitive impairment.

Furthermore, when individual subregions of GM volume reduction were considered as mediators (variable “M”), it was found that WM lesions exerted both direct and indirect effects on cognition scores (MMSE and MoCA). The indirect effects were mediated through 16 subregions of GM volume reduction, including 7 subregions of the prefrontal cortex, 5 subregions of the left occipital region, 2 subregions of the temporal cortex, and 2 subregions of the cerebellar lobes (Figure 4 and Table S2). Notably, the prefrontal subregions exhibited a fully mediating role (with a direct effect p > .05), while the occipital, temporal, and cerebellar subregions displayed a partial mediating role (with a direct effect p < .05) in the relationship between WM lesion burden and cognitive decline.

FIGURE 4.

FIGURE 4

Open in a new tab

Mediation effect of reduced gray matter (GM) volume on white matter hyperintensities (WMH) lesions‐induced cognition decline. Seven atrophy subregions showed in the prefrontal (a), five subregions in the left occipital (b), two subregions in the cerebellum (c) and temporal lobe (d), respectively. CI, confidence interval; FuG, fusiform gyrus; LOcC, lateral occipital cortex; MMSE, Mini‐Mental State Examination; MoCA, Montreal Cognitive Assessment; MFG, middle frontal gyrus; MTG, middle temporal gyrus; MVOcC, medioventral occipital cortex; OrG, orbital gyrus; SFG, superior frontal gyrus.

4. DISCUSSION

This comprehensive study aimed to investigate the patterns of GM atrophy and WM lesions burden in patients with DNS‐D and DNS‐MCI, as well as HCs. Furthermore, the study aimed to explore the causal relationship between these imaging indices and clinical cognitive performance. Three main findings were observed. Firstly, both DNS‐D and DNS‐MCI patients showed a significant burden of WM lesions, along with a similar GM atrophy distributed widely across bilateral cerebral cortical, subcortical, and cerebellar regions. Secondly, DNS‐D patients displayed a greater WM lesions burden and increased GM atrophy in specific regions, including the bilateral PFC, right inferior parietal lobe, bilateral occipitotemporal cortices, and bilateral cerebellum compared to DNS‐MCI patients. Thirdly, the TGMAV was found to be an independent factor influencing the severity of cognitive impairment. Furthermore, WM lesions exerted both direct and indirect effects on cognitive impairment, with the latter mediated by regional GM atrophy. Finally, our study highlighted the BN atlas‐based analysis as an effective method for investigating structural changes in GM.

4.1. Cortical GM atrophy

According to a previous study (Chen et al., 2015), patients with cognitive impairment at the chronic stage of CO poisoning exhibit widespread volumetric loss in the bilateral cerebral cortex. Furthermore, greater atrophy is accompanied by a higher cognitive decline rating. Similarly, we observed a significant reduction in GM volume in both patient groups, particularly in the bilateral PFC, occipito‐temporo‐insular cortices, and cerebellum. Expanding on this finding, we also identified severe cortical thinning, mainly affecting the bilateral frontal, occipital, and parietal lobes. It is worth noting that GM is particularly susceptible to hypoxia due to the high activity and blood demand of neurons in this region. Previous studies on human and rat cases of DNS after CO poisoning (Lapresle & Fardeau, 1967; Piantadosi et al., 1997) have shown necrosis, microbleeds, apoptosis of the cerebral cortex, as well as spongy and necrotic demyelination of white matter in the centrum ovale. Therefore, the direct damage caused by hypoxia may explain cortical atrophy in WMH and subsequent cognitive impairment. Furthermore, in comparison to the DNS‐MCI group, the DNS‐D group exhibited a significant reduction in GM volume in 31 subregions, including the bilateral PFC, right inferior parietal lobe, bilateral occipitotemporal cortices, and cerebellum. These regions are known to be part of cognitive‐related networks, such as the DMN, VN, and posterior CN. Our findings suggested that the structural atrophy of these cognitive network nodes contributed to the neuropathological basis for dementia in DNS patients.

Interestingly, the SBM analysis demonstrated a higher sensitivity in detecting cortical structural changes when comparing the patient groups to the HCs. However, when comparing the patient groups to each other, the SBM analysis showed a lower sensitivity compared to VBM and DBM. It is important to note that cortical thickness analysis examines the interplay of GM morphology among different brain regions using 2D cortical surface data, whereas VBM and DBM utilize 3D volumetric data (Gupta et al., 2019). A reduction in cortical volume can be attributed to a decrease in thickness, area, or both (Landin‐Romero et al., 2017). However, the extent to which these data‐driven analyses correspond to the specific pathological features underlying between‐group comparisons is still unknown.

4.2. Subcortical GM and cerebellar atrophy

In addition to cortical atrophy, we observed significant atrophy in the caudate nucleus, putamen, thalamus, and hippocampus in both groups of DNS patients. Previous multimodality neuroimaging studies have also reported these patterns of subcortical GM atrophy and hypometabolism in DNS patients with cognitive impairment (Chen et al., 2013, 2015). Alterations in the glucose metabolism of the ACC, thalamus, ventral striatum, and superior frontal cortex have been linked to verbal memory function in patients with CO poisoning along the presynaptic dopamine pathway. Likewise, the uptake ratio of the putamen has been associated with visual memory (Chang et al., 2016). Thalamic and hippocampal atrophy has been correlated with lower MMSE scores in CO poisoning patients with mixed symptoms (Zhang et al., 2023). The thalamus, acting as a hub region for brain functional integration and coordination, has been suggested to be associated with cognitive decline in patients with WMH (Zhu et al., 2021). Moreover, we found no significant atrophy in the deep GM nuclei when comparing patient groups, indicating a lesser role of these structures in the severity of cognitive impairment. DNS patients commonly experience a combination of symptoms like anxiety, depression, fatigue, motor disorders, and mutism (Liu et al., 2020; Zhang et al., 2023). Structural and functional abnormalities in the thalamus, caudate nucleus, and putamen have been linked to emotional and motor decline in CO poisoning patients (Chen et al., 2013; Sun et al., 2018). In addition to being key regions within cognitive networks, these deep GM nuclei and the cerebral cortex form the cortex‐striatum circuit, which is implicated in emotion and motor control (George & Das, 2023). Therefore, it is possible that these deep GM nuclei also play a significant role in the manifestation of these mixed symptoms.

A significant decrease in GM volume was observed in the bilateral cerebellum, particularly in the posterior cerebellar lobes, in both patient groups. Additionally, we found that GM atrophy in the bilateral lobule VI mediated the cognitive decline induced by WMH. This finding is consistent with recent claims suggesting that lesions in the cerebellar posterior lobe contribute to non‐motor symptoms, and that the bilateral cerebellar lobule VI and its adjacent regions play a crucial role in cognitive circuits (Guell et al., 2018; Schmahmann et al., 2019). Although there is debate regarding the advantages of cerebellum‐specific normalization compared to the whole brain MNI normalization procedure, it is important to note that caution is still necessary when interpreting our cerebellar results due to the lack of cerebellum‐specific registration processing in the current study (Abdelgabar et al., 2019; Draganova et al., 2022).

4.3. The relationship among GM atrophy, WM lesions burden, and cognitive impairment

Our previous study identified greater GM atrophy in the limbic lobes, motor and visual cortices, and DMN in patients with CO poisoning and T2 hyperintense lesions compared to those without these lesions (Zhang et al., 2023). Building upon these findings, we aimed to investigate and compare GM structural alterations in patients with DNS based on different burdens of WM lesions. Our results indicated that patients with DNS and a higher burden of WM lesions exhibited more pronounced GM atrophy, which is correlated with poorer cognitive function. However, we did not explore any significant interaction effect between these two structural indices. Furthermore, total GM atrophy volume, serving as a reliable measure, independently influences the severity of cognitive impairment, regardless of WM lesion burden. Additionally, we discovered that total GM atrophy volume fully mediated the relationship between WM lesions and cognitive functioning.

Pathologically, the delayed formation of WM lesions is associated with demyelination, which occurs due to the absence of myelin‐related lipids and proteins (Law‐Ye et al., 2018; Meyer, 2013). The presence of WMH has long been considered indicative of the neuropathology underlying delayed cognitive, motor, psychiatric, and other abnormal symptoms. Our results provide additional evidence that insidious GM atrophy contributes to cognitive decline in patients with DNS, independently of extensive visible WM lesions. Notably, prior imaging studies have demonstrated improvements in cognitive symptoms accompanying GM structural, metabolic, and functional normalization in patients with CO poisoning (Chang et al., 2010; Hansen et al., 2014; Tamura et al., 2021). In contrast, even when cognitive impairment improves 1 month after DNS treatment, WM lesions detected on brain MRI persist and may continue beyond 1 year (Hayashi et al., 2022; Inagaki et al., 1997).

Furthermore, we used the volume of each atrophic subregion as a mediator variable “M” and found both direct and indirect effects of WMH lesions on cognitive decline. The indirect effect was mediated through the 16 GM atrophy subregions. This suggests that, in addition to possible direct GM damage caused by hypoxia, GM atrophy induced by WM lesions is also responsible for subsequent cognitive impairment. Among these atrophic subregions, the SFG, MFG, orbital gyrus, and middle temporal gyrus are involved in regulating language, memory, execution, and emotion. In contrast, the visual cortices subregions, including the right fusiform gyrus, left medioventral occipital cortex, and left lateral occipital cortex, are associated with controlling perception, vision, and motion (Fan et al., 2016). Additionally, we observed a significant direct effect of WMH lesions on cognitive impairment in the posterior cerebral and cerebellum regions, rather than the anterior cerebral region. This result suggests that the impact of WM lesions on cognition is influenced by the specific location. In future investigations, it will be necessary to take WM lesions distribution into account or explore the role of tract‐specific changes in WMH in cognitive impairment among DNS patients to gain a better understanding of the characteristic damage to WM.

The impact of WM lesions on the GM structure and subsequent cognitive function in DNS patients remains unclear. We observed WM lesions in two patient groups, which were widely distributed in the centrum semiovale, periventricular, frontal, occipital, temporal, and corpus callosum regions. Consequently, we speculate that these subcortical WM lesions might disrupt the subcortical–cortical connections, leading to secondary damage to the axonal cytoskeleton, cortical degeneration, and volume reduction (Du et al., 2005). Specifically, certain WM fiber bundles, such as the corpus callosum, frontal‐occipital fasciculus, superior and inferior longitudinal fasciculus, and corticopontine frontal tract and occipital tract, traverse these WMH lesions and are directly or indirectly connected to GM regions. The disruption of these fiber bundles may contribute to neurodegeneration, neuronal atrophy, and subsequent cognitive impairments.

4.4. Advantages of different GM structural analysis methods

Using the same statistical threshold, we observed greater sensitivity in the BN atlas‐based volume analysis compared to other analyses when exploring GM structural alterations among DNS groups. Additionally, we confirmed the reproducibility of this main finding through a control analysis utilizing the AAL3 atlas. Moreover, our DBM analysis revealed a greater number of identified subcortical regions but fewer cortical regions compared to the VBM analysis in the DNS groups. This aligns with previous research suggesting that DBM may possess heightened sensitivity for detecting subcortical atrophy in contrast to VBM (Zeighami et al., 2015). As different methods focus on distinct aspects to extract structural information from the brain, they may exhibit varying sensitivities in detecting different types of structural changes. Our findings underscore the importance of employing multiple analysis methods in conjunction and support the notion that comparing results across studies employing different analytical methods is not straightforward.

4.5. Limitation

This study has several limitations. Firstly, due to the absence of an ideal automated segmentation and quantification tool in the initial attempt, we resorted to a manual method for quantifying WMH, which introduced inter‐observer variability. Secondly, the retrospective study design inherently lacks control over pre‐existing cognitive states. Although we controlled for variables such as age, history of AD, and other potential causes of cognitive impairment, the limitations associated with a retrospective study design persist. Thirdly, we did not perform specific cognitive field‐related mediation analyses, such as attention, memory, executive function, and verbal fluency testing. Consequently, our investigation was limited to assessing alterations of cognitive field‐related imaging indices. Lastly, the small sample size and cross‐sectional design may have affected statistical power and rendered our findings insufficient in capturing the evolution of WM and GM structural damage.

4.6. Conclusion

Overall, our findings demonstrated that DNS patients with dementia and MCI often exhibited both diffuse GM atrophy and WM lesions. The TGMAV independently influenced lower cognitive scores, while WM lesions impacted cognitive decline partially through a mediation effect involving regional GM atrophy. These results contributed to a deeper comprehension of the underlying neuromechanisms responsible for delayed cognitive impairment in CO poisoning patients.

CONFLICT OF INTEREST STATEMENT

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

Supporting information

Data S1. Supporting Information.

HBM-45-e26656-s001.docx (9.4MB, docx)

ACKNOWLEDGMENTS

The authors wish to thank the patients and their caregivers for their time and commitment to this research. The authors would also like to thank Xiaoming Chen, Dapeng Liang, Hang Guo, and Jiang Nan for their assistance in recruitment of participants and MRI scan. Last, the authors wish to thank Xiping Sheng and Yu Zhu for their assistance in statistical analysis and image data processing scrutiny.

Zhang, Y. , Wang, T. , Wang, S. , Zhuang, X. , Li, J. , Guo, S. , & Lei, J. (2024). Gray matter atrophy and white matter lesions burden in delayed cognitive decline following carbon monoxide poisoning. Human Brain Mapping, 45(5), e26656. 10.1002/hbm.26656

DATA AVAILABILITY STATEMENT

The data supporting the findings of this study are available from the corresponding author upon reasonable request with the execution of necessary data‐use agreements.

REFERENCES

  1. Abdelgabar, A. R. , Suttrup, J. , Broersen, R. , Bhandari, R. , Picard, S. , Keysers, C. , De Zeeuw, C. I. , & Gazzola, V. (2019). Action perception recruits the cerebellum and is impaired in patients with spinocerebellar ataxia. Brain, 142(12), 3791–3805. 10.1093/brain/awz337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beppu, T. (2014). The role of MR imaging in assessment of brain damage from carbon monoxide poisoning: A review of the literature. American Journal of Neuroradiology, 35(4), 625–631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chang, C. C. , Hsu, J. L. , Chang, W. N. , Huang, S. H. , Huang, C. W. , Chang, Y. T. , Chen, N. C. , Lui, C. C. , Lee, C. C. , & Hsu, S. W. (2016). Metabolic covariant network in relation to nigrostriatal degeneration in carbon monoxide intoxication‐related parkinsonism. Frontiers in Neuroscience, 10, 187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chang, D. C. , Lee, J. T. , Lo, C. P. , Fan, Y. M. , Huang, K. L. , Kang, B. H. , Hsieh, H. L. , & Chen, S. Y. (2010). Hyperbaric oxygen ameliorates delayed neuropsychiatric syndrome of carbon monoxide poisoning. Undersea & Hyperbaric Medicine, 37(1), 23–33. [PubMed] [Google Scholar]
  5. Chen, H. L. , Chen, P. C. , Lu, C. H. , Hsu, N. W. , Chou, K. H. , Lin, C. P. , Wu, R. W. , Li, S. H. , Cheng, Y. F. , & Lin, W. C. (2013). Structural and cognitive deficits in chronic carbon monoxide intoxication: A voxel‐based morphometry study. BMC Neurology, 1(13), 129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen, N. C. , Huang, C. W. , Huang, S. H. , Chang, W. N. , Chang, Y. T. , Lui, C. C. , Lin, P. H. , Lee, C. C. , Chang, Y. H. , & Chang, C. C. (2015). Cognitive severity‐specific neuronal degenerative network in charcoal burning suicide‐related carbon monoxide intoxication: A multimodality neuroimaging study in Taiwan. Medicine (Baltimore), 94(19), e783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chou, M. C. , Li, J. Y. , & Lai, P. H. (2019). Longitudinal white matter changes following carbon monoxide poisoning: A 9‐month follow‐up voxelwise diffusional kurtosis imaging study. American Journal of Neuroradiology, 40(3), 478–482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Diedrichsen, J. , Balsters, J. H. , Flavell, J. , Cussans, E. , & Ramnani, N. (2009). A probabilistic MR atlas of the human cerebellum. NeuroImage, 46(1), 39–46. [DOI] [PubMed] [Google Scholar]
  9. Draganova, R. , Konietschke, F. , Steiner, K. M. , Elangovan, N. , Gümüs, M. , Göricke, S. M. , Ernst, T. M. , Deistung, A. , van Eimeren, T. , Konczak, J. , & Timmann, D. (2022). Motor training‐related brain reorganization in patients with cerebellar degeneration. Human Brain Mapping, 43(5), 1611–1629. 10.1002/hbm.25746 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Du, A. T. , Schuff, N. , Chao, L. L. , Kornak, J. , Ezekiel, F. , Jagust, W. J. , Kramer, J. H. , Reed, B. R. , Miller, B. L. , Norman, D. , Chui, H. C. , & Weiner, M. W. (2005). White matter lesions are associated with cortical atrophy more than entorhinal and hippocampal atrophy. Neurobiology of Aging, 26(4), 553–559. [DOI] [PubMed] [Google Scholar]
  11. Fan, L. , Li, H. , Zhuo, J. , Zhang, Y. , Wang, J. , Chen, L. , Yang, Z. , Chu, C. , Xie, S. , Laird, A. R. , Fox, P. T. , Eickhoff, S. B. , Yu, C. , & Jiang, T. (2016). The human brainnetome atlas: A new brain atlas based on connectional architecture. Cerebral Cortex, 26(8), 3508–3526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fazekas, F. , Chawluk, J. B. , Alavi, A. , Hurtig, H. I. , & Zimmerman, R. A. (1987). MR signal abnormalities at 1.5 T in Alzheimer's dementia and normal aging. American Journal of Roentgenology, 149(2), 351–356. [DOI] [PubMed] [Google Scholar]
  13. Fujiwara, S. , Beppu, T. , Nishimoto, H. , Sanjo, K. , Koeda, A. , Mori, K. , Kudo, K. , Sasaki, M. , & Ogasawara, K. (2012). Detecting damaged regions of cerebral white matter in the subacute phase after carbon monoxide poisoning using voxel‐based analysis with diffusion tensor imaging. Neuroradiology, 54(7), 681–689. [DOI] [PubMed] [Google Scholar]
  14. George, K. , & Das, J. M. (2023). Neuroanatomy, thalamocortical radiations. StatPearls Publishing. [PubMed] [Google Scholar]
  15. Guell, X. , Schmahmann, J. D. , Gabrieli, J. , & Ghosh, S. S. (2018). Functional gradients of the cerebellum. Elife, 7, e36652. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gupta, C. N. , Turner, J. A. , & Calhoun, V. D. (2019). Source‐based morphometry: A decade of covarying structural brain patterns. Brain Structure & Function, 224(9), 3031–3044. [DOI] [PubMed] [Google Scholar]
  17. Hansen, M. B. , Kondziella, D. , Danielsen, E. R. , Larsen, V. A. , Jansen, E. C. , & Hyldegaard, O. (2014). Cerebral proton magnetic resonance spectroscopy demonstrates reversibility of N‐acetylaspartate/creatine in gray matter after delayed encephalopathy due to carbon monoxide intoxication: A case report. Journal of Medical Case Reports, 8, 211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hayashi, M. , Otsuki, K. , Miura, S. , Mihara, Y. , Abe, S. , & Inagaki, M. (2022). Delayed encephalopathy after carbon monoxide poisoning treated with corticosteroid monotherapy: Case report. Psychiatry and Clinical Neurosciences, 76(11), 600–602. [DOI] [PubMed] [Google Scholar]
  19. Inagaki, T. , Ishino, H. , Seno, H. , Umegae, N. , & Aoyama, T. (1997). A long‐term follow‐up study of serial magnetic resonance images in patients with delayed encephalopathy after acute carbon monoxide poisoning. Psychiatry and Clinical Neurosciences, 51(6), 421–423. [DOI] [PubMed] [Google Scholar]
  20. Jeon, S. B. , Sohn, C. H. , Seo, D. W. , Oh, B. J. , Lim, K. S. , Kang, D. W. , & Kim, W. Y. (2018). Acute brain lesions on magnetic resonance imaging and delayed neurological sequelae in carbon monoxide poisoning. JAMA Neurology, 75(4), 436–443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Jiang, W. , Zhao, Z. , Wu, Q. , Wang, L. , Zhou, L. , Li, D. , He, L. , & Tan, Y. (2021). Study on brain structure network of patients with delayed encephalopathy after carbon monoxide poisoning: Based on diffusion tensor imaging. La Radiologia Medica, 126(1), 133–141. [DOI] [PubMed] [Google Scholar]
  22. Jokinen, H. , Koikkalainen, J. , Laakso, H. M. , Melkas, S. , Nieminen, T. , Brander, A. , Korvenoja, A. , Rueckert, D. , Barkhof, F. , Scheltens, P. , Schmidt, R. , Fazekas, F. , Madureira, S. , Verdelho, A. , Wallin, A. , Wahlund, L. O. , Waldemar, G. , Chabriat, H. , Hennerici, M. , … Erkinjuntti, T. (2020). Global burden of small vessel disease‐related brain changes on MRI predicts cognitive and functional decline. Stroke, 51(1), 170–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Landin‐Romero, R. , Canales‐Rodríguez, E. J. , Kumfor, F. , Moreno‐Alcázar, A. , Madre, M. , Maristany, T. , Pomarol‐Clotet, E. , & Amann, B. L. (2017). Surface‐based brain morphometry and diffusion tensor imaging in schizoaffective disorder. The Australian and New Zealand Journal of Psychiatry, 51(1), 42–54. [DOI] [PubMed] [Google Scholar]
  24. Lapresle, J. , & Fardeau, M. (1967). The central nervous system and carbon monoxide poisoning. II. Anatomical study of brain lesionss following intoxication with carbon monixide (22 cases). Progress in Brain Research, 24, 31–74. [DOI] [PubMed] [Google Scholar]
  25. Law‐Ye, B. , Dodet, P. , Hermann, B. , Trunet, S. , Dormont, D. , Pyatigorskaya, N. , & Leclercq, D. (2018). Progressive white‐matter demyelination in delayed CO poisoning encephalopathy. Journal of Neuroradiology, 45(1), 59–62. 10.1016/j.neurad.2017.09.008 [DOI] [PubMed] [Google Scholar]
  26. Li, H. , Jia, J. , & Yang, Z. (2016). Mini‐mental state examination in elderly Chinese: A population‐based normative study. Journal of Alzheimer's Disease, 53(2), 487–496. [DOI] [PubMed] [Google Scholar]
  27. Liu, J. , Si, Z. , Liu, J. , Lin, Y. , Yuan, J. , Xu, S. , He, Y. , Zhang, T. , & Wang, A. (2020). Clinical and imaging prognosis in patients with delayed encephalopathy after acute carbon monoxide poisoning. Behavioural Neurology, 2020, 1719360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Long, Z. , Huang, J. , Li, B. , Li, Z. , Li, Z. , Chen, H. , & Jing, B. (2018). A comparative atlas‐based recognition of mild cognitive impairment with voxel‐based morphometry. Frontiers in Neuroscience, 12, 916. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McKhann, G. M. , Knopman, D. S. , Chertkow, H. , Hyman, B. T. , Jack, C. R., Jr. , Kawas, C. H. , Klunk, W. E. , Koroshetz, W. J. , Manly, J. J. , Mayeux, R. , Mohs, R. C. , Morris, J. C. , Rossor, M. N. , Scheltens, P. , Carrillo, M. C. , Thies, B. , Weintraub, S. , & Phelps, C. H. (2011). The diagnosis of dementia due to Alzheimer's disease: Recommendations from the National Institute on Aging‐Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement, 7(3), 263–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Meyer, M. A. (2013). Delayed post‐hypoxic leukoencephalopathy: Case report with a review of disease pathophysiology. Neurology International, 5(3), e13. 10.4081/ni.2013.e13 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Parkinson, R. B. , Hopkins, R. O. , Cleavinger, H. B. , Weaver, L. K. , Victoroff, J. , Foley, J. F. , & Bigler, E. D. (2002). White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning. Neurology, 58(10), 1525–1532. [DOI] [PubMed] [Google Scholar]
  32. Piantadosi, C. A. , Zhang, J. , Levin, E. D. , Folz, R. J. , & Schmechel, D. E. (1997). Apoptosis and delayed neuronal damage after carbon monoxide poisoning in the rat. Experimental Neurology, 147(1), 103–114. 10.1006/exnr.1997.6584 [DOI] [PubMed] [Google Scholar]
  33. Rizvi, B. , Lao, P. J. , Chesebro, A. G. , Dworkin, J. D. , Amarante, E. , Beato, J. M. , Gutierrez, J. , Zahodne, L. B. , Schupf, N. , Manly, J. J. , Mayeux, R. , & Brickman, A. M. (2021). Association of regional white matter hyperintensities with longitudinal Alzheimer‐like pattern of neurodegeneration in older adults. JAMA Network Open, 4(10), e2125166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rolls, E. T. , Huang, C. C. , Lin, C. P. , Feng, J. , & Joliot, M. (2020). Automated anatomical labelling atlas 3. NeuroImage, 206, 116189. [DOI] [PubMed] [Google Scholar]
  35. Rose, J. J. , Wang, L. , Xu, Q. , McTiernan, C. F. , Shiva, S. , Tejero, J. , & Gladwin, M. T. (2017). Carbon monoxide poisoning: Pathogenesis, management, and future directions of therapy. American Journal of Respiratory and Critical Care Medicine, 195(5), 596–606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Schmahmann, J. D. , Guell, X. , Stoodley, C. J. , & Halko, M. A. (2019). The theory and neuroscience of cerebellar cognition. Annual Review of Neuroscience, 42, 337–364. [DOI] [PubMed] [Google Scholar]
  37. Shi, D. , Yao, X. , Li, Y. , Zhang, H. , Wang, G. , Wang, S. , & Ren, K. (2022). Classification of Parkinson's disease using a region‐of‐interest‐ and resting‐state functional magnetic resonance imaging‐based radiomics approach. Brain Imaging and Behavior, 16(5), 2150–2163. [DOI] [PubMed] [Google Scholar]
  38. Sun, T. K. , Chen, Y. Y. , Huang, S. H. , Hsu, S. W. , Lee, C. C. , Chang, W. N. , Huang, C. W. , Lui, C. C. , Lien, C. Y. , Cheng, J. L. , & Chang, C. C. (2018). Neurotoxicity of carbon monoxide targets caudate‐mediated dopaminergic system. Neurotoxicology, 65, 272–279. [DOI] [PubMed] [Google Scholar]
  39. Tamura, T. , Sugihara, G. , & Takahashi, H. (2021). Memory impairment and hippocampal volume after carbon monoxide poisoning. Archives of Clinical Neuropsychologist, 36(1), 145–148. [DOI] [PubMed] [Google Scholar]
  40. Tuladhar, A. M. , Reid, A. T. , Shumskaya, E. , de Laat, K. F. , van Norden, A. G. , van Dijk, E. J. , Norris, D. G. , & de Leeuw, F. E. (2015). Relationship between white matter hyperintensities, cortical thickness, and cognition. Stroke, 46(2), 425–432. [DOI] [PubMed] [Google Scholar]
  41. Wang, T. , Zhang, Y. , Nan, J. , Li, J. , Lei, J. , & Guo, S. (2023). Surface‐based morphometry study of brain in patients with carbon monoxide poisoning. European Journal of Radiology, 160, 110711. [DOI] [PubMed] [Google Scholar]
  42. Zeighami, Y. , Ulla, M. , Iturria‐Medina, Y. , Dadar, M. , Zhang, Y. , Larcher, K. M. , Fonov, V. , Evans, A. C. , Collins, D. L. , & Dagher, A. (2015). Network structure of brain atrophy in de novo Parkinson's disease. eLife, 4, e08440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zeng, W. , Chen, Y. , Zhu, Z. , Gao, S. , Xia, J. , Chen, X. , Jia, J. , & Zhang, Z. (2020). Severity of white matter hyperintensities: Lesions patterns, cognition, and microstructural changes. Journal of Cerebral Blood Flow and Metabolism, 40(12), 2454–2463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhang, Y. , Wang, T. , Wang, S. , Gao, Y. , Wang, S. , Guo, S. , & Lei, J. (2023). Early gray matter atrophy and neurological deficits in patients with carbon monoxide poisoning. Neuroradiology, 65(2), 245–256. [DOI] [PubMed] [Google Scholar]
  45. Zhu, W. , Huang, H. , Yang, S. , Luo, X. , Zhu, W. , Xu, S. , Meng, Q. , Zuo, C. , Liu, Y. , Wang, W. , & Alzheimer's Disease Neuroimaging Initiative . (2021). Cortical and subcortical grey matter abnormalities in white matter hyperintensities and subsequent cognitive impairment. Neuroscience Bulletin, 37(6), 789–803. [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

Data S1. Supporting Information.

HBM-45-e26656-s001.docx (9.4MB, docx)

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request with the execution of necessary data‐use agreements.

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES