The Neural Network Of Cognitive Impairments Associated With Periventricular White Matter Lesions(PWMLs)

Background:White matter lesions(WMLs), which are defined by areas in the cerebral white matter that appear hypo-intense on T1-weighted, but hyper-intense on T2-weighted or fluid attenuated inversion recovery(FLAIR) of magnetic resonance imaging(MRI) and hypo-dense on computed tomography, are commonly related to chronic microvascular diseases and cerebral hypo-perfusion in elderly people. The prevalence of MRI-detected WMLs in individuals over 60 years is higher than 30% and increases with age. There are several theories to explain the pathogenic mechanism of WMLs including injury caused by hypo-perfusion, blood-brain barrier damage, endothelial damage, and β amyloid deposition.Most researchers believe that WMLs are associated with chronic ischemia induced by hypo-perfusion. Traditionally, WMLs has been separated into periventricular(PWMLs) and deep white matter lesions(DWMLs). Periventricular is defined as 3- 13 mm from ventricular surface.There is increasing evidence that WMLs are associated with an increased risk of cognitive impairment. The findings of several studies suggest that the presence of PWMLs rather than DWMLs is more associated with the impairment of cognitive function. This may be because PWMLs affected the long association fibers that connect the more distant cortical areas. Until now, the neural mechanisms explaining how cognitive impairment is associated with WMLs have not been fully elucidated. Some researchers hypothesize that WMLs might impair cognitive functions by interfering with the speed or fidelity of neuronal signaling when the signal passes through the injured white matter regions.However, the direct evidence to support this hypothesis has been scant. Conversely, recent studies did not find that WMLs per se directly resulted in cognitive impairments but supported the idea that the cognitive decline associated with WMLs was mediated bycortical dysfunction secondary to WML progression. That means, the cognitive impairments associated with WMLs might be caused by a disruption of white matter tracts in WMLs and, consequently, abnormal function or even structure in the corresponding cortices, which are involved in cognitive processing. This hypothesis requires further confirmation. Therefore, it is necessary to simultaneously investigate the impacts of specific locations of white matter and the corresponding regions of cortex on cognitive function when addressing the neural mechanisms of cognitive impairment associated with WMLs.Methods:In the present study, subjects with “pure” PWMLs were recruited to investigate the neural mechanisms of cognitive impairment associated with WMLs, focusing on the contribution of specific locations of white matter and cortex to the observed deficits.Twenty-three PWMLs subjects were divided into cognitive impairment(PWMLs-CI) and cognitive normal(PWMLs-CN) grous according to their cognitive assessments by a series of neuropsychological tests. Twelve subjects with normal magnetic resonance imaging(MRI) and cognition were recruited as controls.Then, the fractional anisotropy(FA) values of the white matter from diffusion tension imaging(DTI) and the fractional amplitude of low-frequency fluctuation(f ALFF) values of cortex from resting-state functional MRI(rfMRI) were evaluated and comparedamong the PWMLs-CI, PWMLs-CN and control(CN)groups. Finally, correlations between FA values, fALFF values and cognitive assessment scores were analyzed in the PWMLs subjects.Brain regions with significant different fALFF value between two groups were defined as seed ROIs, Group comparisons of functional connectivity based on seed ROIs, The altered functional connectivity were further correlated with the MOCA scores.Results:We recruited 23 subjects with PWMLs and 12 control subjects. One subject in the PWMLs group was excluded from the final analysis due to excessive head motion during MR scanning(> 2 mm in translation). Of the remaining 22 PWMLs subjects, 11 subjects were allocated into the PWMLs-CI group, and the other 11 subjects were allocated into the PWMLs-CN group. There were no differences in demographic or clinical variables including age, gender, years of education, and vascular risk factors among the three groups.In the aspect of cognitive functions, compared with the PWMLs-CN group, the subjects in the PWMLs-CI group demonstrated lower Mo CA, TMT, SDMT and LMT scores, which indicated that the PWMLs-CI group had a significant decline in the domains of executive,attention and logical memory functions(all P < 0.05). Compared with the CN group, the subjects in the PWMLs-CN group demonstrated no differences in the neuropsychological test(all P > 0.05).Compared with the PWMLs-CN group, the subjects in the PWMLs-CI group presented with decreased FA values in the right splenium of the corpus callosum and the right posterior cingulum bundle(all P < 0.05, corrected by AlphaSim method).The FA values of the right splenium of the corpus callosum were positively correlated with the Mo CA and SDMT scores. The FA values of the right posterior cingulum bundle were positively correlated with the MoCA, SDMT and LMT scores but negatively correlated with the TMT scores.Compared with the PWMLs-CN group, the subjects in the PWMLs-CI group presented with decreased fALFF values in the left frontal middle gyrus, left precentral gyrus, right angular gyrus, and right precuneus. Additionally, the subjects in the PWMLs-CI group also presented increased fALFF values in the right mid cingulum cortex, right hippocampus amygdala, right cerebellar hemisphere and left vermis(all P < 0.05, corrected by AlphaSim method). The fALFF values were not significantly different between the PWMLs-CN and CN groups. The fALFF values of the left frontal middle gyrus and right precuneus were positively correlated with the MoCA, SDMT, and LMT scores but negatively correlated with the TMT scores. The fALFF values of the left precentral gyruswere positively correlated with the Mo CA scores. The fALFF values of the right angular gyrus were positively correlated with the Mo CA and SDMT scores but negatively correlated with the TMT scores. The f ALFF values of the right mid cingulum cortex, and left cerebellar vermis were positively correlated with the TMT scores but negatively correlated with the MoCA,SDMT, and LMT scores. The fALFF values of the right hippocampus amygdala were positively correlated with the TMT scores but negatively correlated with the Mo CA and LMT scores. The fALFF values of the right cerebellar hemisphere were negatively correlated with the MoCA scores.The averaged Z values of all voxels in areas with altered functional connectivitywere further correlated with the MOCA score. The results showed thatPrefrontal-Temporal,Angular-Temporal-Occipital,Precuneus-Temporal,Precuneus-Prefrontal,Cingulum_Mi d-Parietal-Postcentral,Amy_Hipp-Temporal,Cerebelum-Prefrontal,Cerebelum-Cingulum_M id-Precentral networks correlated with the MOCA score.Conclusions:Our study suggests that PWMLs impair executive, attention and logical memory functions, and these impairments are associated with specific locations of white matter lesion and cortical dysfunction. This may be the neural networks mechanism to explain the cognitive impairments associated with PWMLs.