Alterations of White Matter Integrity in Cerebral Small Vessel

Yifan Wang, Tianyao Wang, Zekuan Yu*, Bo Huang, Biao Liu, Xianwei Liu, Huabin Yin 4 , Jun Liu* 5 a Department of Radiology, Shanghai Fifth People’s Hospital, Fudan University, Shanghai, China 6 b Department of Radiology, Guigang City People’s Hospital, Guigang, China 7 c Department of Radiology, Tongren Hospital, Shanghai Jiao Tong University School of Medicine, 8 Shanghai, China 9 d Academy for Engineering and Technology, Fudan University, Shanghai, 200433, China 10 e Key Laboratory of Industrial Dust Prevention and Control & Occupational Health and Safety, 11 Ministry of Education 12 f Anhui Province Engineering Laboratory of Occupational Health and Safety 13 g Laboratory of Industrial dust deep reduction and occupational health and safety of Anhui Higher 14 Education Institutes 15


46
Cerebral small vessel disease (CSVD) is a group of clinical syndromes involving cerebral 47 arterioles, microarteries, capillaries and venules, which accounts for approximately 10%-30% of 48 global ischemic strokes [1] and is a major vascular contributor to cognitive deficits and dementia [2]. 49 White matter hyperintensity (WMH), as one the common imaging markers of CSVD has been 50 widely reported to be associated with cognitive decline and progression of cognitive impairment[3- injury as well as the mechanism of WMH-related cognitive impairment. 55 Diffusion tensor imaging (DTI) is an advanced technique for detecting changes in the 56 microstructure of WM [6], which is sensitive to the change of WM microstructure integrity [7,8]. It 57 can not only reflect the injury of WM in WMH area, but also detect the change of WM fiber tracts 58 which seem normal on traditional MRI [9]. Relevant studies have shown that injury of WM 59 microstructure is related to cognitive impairment [10,11]. Therefore, DTI can be used to explore the 60 characteristic of WM injury at the micro level and the neural mechanism of cognitive impairment 61 caused by WMH. Under usual circumstances, four main diffusion indicators including fractional 62 anisotropy (FA), mean diffusion (MD), axial diffusion (AD) and radial diffusion (RD) are applied 63 to provide more information on WM microstructure and its changes in relation to cognitive function 64 Among different methods used in DTI research, trace-based spatial statistics (TBSS) is a reliable 66 and optimized one that minimizes registration errors and personal evaluation biases, and is 67 considered to improve sensitivity, objectivity, and interpretability when applied to multiple diffused 68 assessment and multimodal MRI. The inclusion criteria were as follows: 1) patients aged older than 85 55 years, 2) no history of brain trauma or dementia, 3) MRI scan showed WMH imaging 86 manifestations. The exclusion criteria were as follows: 1) non-lacunar infarction in cerebral cortex 87 or cerebellum or brainstem, 2) have a history of hydrocephalus, cerebral tumor or space occupation, 88 3) unable to cooperate with this study independently or suffering from serious physical and mental 89 diseases, 4) MRI contraindications, 5) leukodystrophy caused by other causes (such as multiple 90 sclerosis, history of brain exposure, etc) 91 According to Fazekas grade scale [18], WMH in the patient's periventricular and deep white 92 matter were graded separately, and the two grades were added together to record the total score. 93 Finally, WMH patients were divided into two groups:(A) WMH score of 1-2 points, (b) WMH score 94 of 3-6 points. 95

Neuropsychological assessment 96
In this study, all subjects underwent neuropsychological cognitive assessment within one week of MRI examination. We performed a simple mental state examination (MMSE) along with 98 Montreal Cognitive Assessment (MOCA) for cognitive assessment and recorded the total score. 99

Image preprocessing 108
The steps of the DTI data preprocessing were as follows

Tract-Based Spatial Statistics (TBSS) 115
Firstly, FSL nonlinear image registration algorithm was used to align the FA map of each subject 116 to FMRIB58_FA standard space. Then, the mean FA image is generated. By refining the mean FA 117 image, the mean FA skeleton representing the core structure of WM domain is generated later. 118 Finally, individual subject FA images were projected onto the mean FA skeleton. These skeleton 119 projection factors are also applicable to MD, AD, and RD images [13]. 120

Statistical analysis 121
All statistical analyses were performed in SPSS26.0 statistical software [20]. Demographic, 122 clinical characteristics, medical history, and neuropsychological data were compared by t test, chi 123 square test, and nonparametric test. We used t-test to compare the difference of DTI-derived indexes 124 between the two groups. In order to control class I errors, false discovery rate (FDR) correction is 125 adopted. Then, linear regression analysis was used for age correction. P<0.01 was considered 126 statistically significant [21]. Partial correlation analysis was used to evaluate the relationship 127 between DTI-derived indexes and overall cognitive function. Age, gender, and education level were 128 considered as covariates in partial correlation analysis. P<0.05 was considered statistically 129

132
In terms of demographic data, no significant difference exists between two groups except the 133 age (p<0.05). Compared to subjects in Group A, subjects in Group B were characterized by an older 134 mean age significantly. In addition, no statistically difference was observed on the aspect of clinical 135 data as well as neuropsychological data. All relevant results are depicted in Table 1. 136

TBSS Analysis 139
After age correction, compared to the group of low WMH scores, the patient group with high    as well as some long association fibers was more obvious. We hypothesized that these WM fibers 192 were very sensitive to hemodynamic changes. The more severe the damage of WM microstructure 193 was, the more obvious the WMH presented on conventional MRI. 194 In addition to the results above, we also found the decreased FA in the bilateral tapetum along All of these suggest that with the increase of the severity of WMH, the impairment of 223 microstructure tends to occur on the WM fiber tracts which are closely related to cognition. Similarly, 224 researchers noticed that in the early stage of CSVD, WM microstructural injury mainly occurred in 225 the cognition-related WM fibers [35][36]. Therefore, we hypothesized that in CSVD patients, the 226 WM fiber tracts, which are closely related to cognition, are more susceptible to be injuried. This 227 may be the reason why CSVD patients often suffer from cognitive impairment. 228 However, in this study, we found no correlation between the DTI derived index and cognition, 229 which was inconsistent with previous studies [14,16,17,39]. This may be related to the fact that 230 our enrolled subjects are mostly preclinical patients. These subjects tended to show only 231 microstructure impairment but not obvious clinical symptoms such as cognitive decline. Previous 232 study indicated that in these non-dementia CSVD patients, only a few areas showed significant node 233 efficiency changes which contribute to cognition decline, despite extensive WM integrity 234 impairment [27]. For one reason, WMH represents loss of the myelin sheath and axon and does not 235 cause complete destruction of the fibers especially in the early stage of CSVD. For the other, reactive 236 structural plasticity such as gliosis is a common histopathological change in CSVD, which may lead 237 to the strengthening of interhemispheric connections [40]. Hence, we conclude that changes in WM 238 microstructure in CSVD patients predate cognitive decline. If we can detect the microstructural 239 changes in WM before the onset of cognitive decline and give some interventions, we may be able 240 to delay cognitive decline to some extent.
Several limitations need to be mentioned in current study. First of all, this is a cross-sectional 242 study which limits our observation on longitudinal effects of cerebral small vessel disease. Secondly, 243 patients with WMH were graded by visual observation, which is somewhat subjective and cannot 244 accurately reflect the severity of white matter lesions (WMLs). Finally, no health control group was 245 set up in our research. Therefore, we plan to include healthy subjects in further experimental study. 246 Besides, we will assess the WMH load by measuring the WMH volume as well as its location. 247

Conclusion 248
In CSVD patients, the WM fiber tracts that are closely related to cognitive function tend to be 249 more vulnerable to be injured, and the injury of these WM fiber tracts is more obvious with the 250 aggravation of WMH degree. In addition, changes in WM microstructure often predate changes of 251 cognition. Therefore, early detection of microstructural changes and timely intervention can delay 252 cognitive decline to some extent.

Availability of data and materials 261
The data sets in this study are available from the corresponding author on reasonable request.