Acta Neurol Scand DOI: 10.1111/ane.12262
© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA NEUROLOGICA SCANDINAVICA
Cognitive impairments associated with periventricular white matter hyperintensities are mediated by cortical atrophy Zi W, Duan D, Zheng J. Cognitive impairments associated with periventricular white matter hyperintensities are mediated by cortical atrophy. Acta Neurol Scand: DOI: 10.1111/ane.12262. © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd. Background – Previous studies have shown that white matter lesions (WMLs) is an important risk factor for cognitive impairment, but the underlying mechanisms have not been clarified. Objective – We tested the hypothesis that the cognitive impairments associated with periventricular white matter hyperintensities (PWMHs) on magnetic resonance imaging (MRI) would be mediated by the cortical thinning of corresponding area. Method – Sixteen stroke- and dementia-free subjects with PWMHs and 16 healthy control subjects were enrolled in this study. All participants underwent an examination of cognition, MRI-based cortical thickness measurement and a MRI-DTI scan. Then, the possible relationships among cognitive impairments, PWMHs and the topography of cortical thinning were analyzed. Results – Comparing with the controls, the cognitive tests of the subjects with PWMHs showed significant decline in the domains of verbal fluency and executive function. After accounting for age, gender, years of education, and treatable vascular risk factors related to cognitive performance, cortical thickness had an independent influence on the cognitive impairments, especially in the frontal pole, orbitofrontal cortex, superior and middle frontal gyrus, superior and middle temporal gyrus, insula, and cuneus. Conclusions – Our results suggest that the association between PWMHs and cognitive impairments is mediated by cortical thinning. Introduction
White matter hyperintensities (WMHs), which are commonly observed on MRI scans of elderly people, have been found to increase in prevalence and severity with age (1). They are associated with cognitive dysfunction in patients with mild cognitive impairment or dementia and have recently been identified as a risk factor for dementia (2). Although WMHs may tend to be associated with cognition decline, the independent associations of periventricular white matter hyperintensities (PWMHs) with cognition are inconsistent among studies. In a longitudinal cohort study of normal elderly subjects, PWMHs were associated with the level of cognitive impairment (3), whereas some other studies indicated that
W. Zi, D. Duan, J. Zheng Department of Neurology, Xinqiao Hospital, Third Military Medical University, Chongqing, China
Key words: periventricular white matter hyperintensities; cortical thinning; cognitive impairments; diffusion tension imaging J. Zheng, Department of neurology, Xinqiao Hospital, Third Military Medical University, 183 Xinqiao Street, Chongqing 400037, China Tel./Fax: +86 023 68763203 e-mail: [email protected] Accepted for publication April 16, 2014
PWMHs are not associated with levels of cognitive decline when cortical volume loss is accounted for (4). Therefore, some researchers presumed that the correlation between cognitive impairment and PWMHs is mediated by cortical atrophy. However, the previous analyses regarding the association of PWMHs or cortical thickness with cognitive function have not excluded the possible impacts of other types of WMHs. Therefore, it is necessary to further investigate the association of PWMHs and cortical volume loss with cognitive impairments in the subjects with ‘pure’ PWMHs. Conventional structural MRI techniques, such as T2-weighted and fluid-attenuated inversion recovery (FLAIR) images, provide little information about the severity of the underlying pathological changes in (5). Diffusion tension image 1
Zi et al. (DTI) is a quantitative MRI technique that measures the directionality and mobility of water diffusion in tissue and provides increased sensitivity to detect the structural integrity of white matter fibers (6). Fractional anisotropy (FA) is a measurement of the diffusion consistency in all direction (7). The present study evaluated PWMHS with magnetic resonance DTI, the topography of cortical thickness with surface-based morphometry, and cognitive impairments with neuropsychological tests in the stroke- and dementia-free subjects. The relationships among these variables are evaluated to further clarify the association of PWMHs and cortical thinning with cognitive impairments. Methods Subjects
Sixteen stroke- and dementia-free subjects with PWMHS and 16 controls without PWMHS on MRI were consecutively enrolled in this study between June 8, 2012 and August 30, 2013. All of them were clients who accepted regular health checks in the health examination center in Xinqiao Hospital, Third Military Medical University and voluntarily participated into this study. This study was approved by the ethics committee of our hospital, and all participants gave their informed consent. All participants were required to be within the age range of 50–75 years and underwent a comprehensive clinical examination by two experienced neurologists, including medical history, physical, and neurological assessments. Subjects PWMHS were evaluated with conventional structural MRI techniques, including T1-weighted, T2weighted, and FLAIR images. Periventricular regions were defined as regions between 3 and 13 mm from the ventricular surface (7). No participants had the disorders that might have confounded their current cognitive state, such as metabolic encephalopathy, thyroid disease, or syphilis, No participants had current or past somatic, psychiatric, or neurological disorders that might have caused the cognitive impairment, such as stroke, schizophrenia, epilepsy, severe head trauma, encephalitis, brain tumors, alcohol abuse, severe depression, or neurodegenerative diseases such as Parkinson’s disease. Control subjects, matched with the PWMHS subjects for age, gender, degree of education, and vascular risk factors, showed no WMH on MRI imaging, no neurological disorders, and no any deficits on the neuropsychological test battery. 2
Acquisition of MR images
Images were acquired with a 3.0-Tesla GE Excite MRI system (GE, Milwaukee, WI, USA). We performed routine whole brain MRI scans for every subject, including axial T1-weighted MRI (acquisition matrix = 320 9 320, TR = 5600 ms, TE = 2.46 ms, slice thickness = 5 mm), axial T2weighted MRI (acquisition matrix = 320 9 320, TR = 5600 ms, TE = 90 ms, slice thickness = 5 mm), and axial FLAIR (acquisition matrix = 320 9 320, TR = 9000 ms, TI = 2250 ms, TE = 85 ms, slice thickness = 5 mm). DTI measurement and analysis
Post-processing of the DTI data was performed on a GE workstation (FuncTool, Advantage Workstation 4.2, GE, Medical System, Milwaukee, WI, USA). First, fractional anisotropy (FA) maps and color-coded directionality maps of diffusion were created in individual images. The color-coded directional maps provided easy visualization of the white matter fiber tracts. Brickshaped regions of interest (ROIs) were drawn based on the identification of white matter tracts on the color-coded maps. Then FA maps were overlaid on these color-coded directional maps. We placed 12 ovoid ROIs in the white matter regions on two slices, which included the temporal deep white matter adjacent to the temporal horns, the genu and splenium of the corpus callosum, the anterior deep white matter, the anterior periventricular white matter, and the posterior deep white matter and the posterior periventricular white matter (see Fig. 1). The non-midline regions were measured on both sides. For most regions, we used the ROI designations described by Fellgiebel et al. (8). The size of the ROIs was 8–12 voxels, measuring 1.878 mm 9 1.875 mm 9 5 mm each (9). Cortical thickness analysis
The automatic cortical thickness analysis of the MRI was conducted using the pipelining method developed at Montreal Neurological Institute (10). First, intensity-based inuniformities, which resulted from the inhomogeneities in the magnetic field, were corrected with the N3 algorithm and the brain tissue was classified into gray matter, white matter, and cerebrospinal fluid, using a 3D stereotaxic brain mask and the intensity-normalized stereotaxic environment for classification of tissues algorithm (11). Next, the brain was automatically divided into two separate hemispheres
The correlation between PWMH and cognitive impairments
Figure 1. Topographic illustration of ROIs measurements in fractional anisotropy maps. Twelve spots for ROIs measurements were selected. Brain regions are indicated as: 1 and 2 = Right and left temporal deep white matter, 3 and 4 = genu and splenium of corpus callosum, 5 and 6 = right and left anterior medial deep white matter, 9 and 10 = right and left posterior deep white matter, 7 and 8 = right and left anterior periventricular white matter, and 11 and 12 = right and left posterior periventricular white matter.
and the inner and outer surfaces of the cortex were automatically extracted, using the Constrained Laplacian-Based Automated Segmentation With proximities (CLASP) algorithm. The CLASP algorithm extracts the boundaries of the cortex by using a dense polygon mesh that is first expanded inside the white matter and fitted into the surface between the white matter and the inner surface of cortex. The outer surface of the cortex is detected by expanding the polygon mesh further into the surface between the gray matter and cerebrospinal fluid. Cortical thickness was defined as the Euclidean distance between the linked vertices of the outer and inner surfaces. Finally, the cortical thickness maps were smoothened using a surface-based 20 mm WHM diffusion kernel that has been demonstrated to increase the signal-to-noise ratio and the statistical power (12). The cortical thickness was calculated in the native space through applying an inverse transformation matrix to cortical surfaces and reconstructing in the native space (13). Neuropsychological tests
All participants completed an extensive neuropsychological test battery. Cognitive measurements were focused on five cognitive domains: executive function, attention, verbal memory, visuospatial skills and verbal fluency. Executive function was assessed with Trail-Making Test, including A and B parts. Both part of the test consists of 25 circles distributed over a sheet of paper. The circles are numbered from 1 to 25 in part A, whereas the circles include both numbers (from 1 to 13) and letters (from A to L) in part B. Participants were required to link all the numbers or letters in digit sequence (part A) or alternating digit
sequence and alphabetical order (part B) as fast as possible. The final score for analysis in this study was calculated by part B score subtracting part A score (14, 15). Attention abilities were examined with Digit Span Forward from the Wechsler Adult Intelligence Scale 3rd edition (WAIS-III) (16, 17). Verbal memory was assessed with Huashan Auditory Verbal Learning Test (HAVLT) in Chinese (18), which was modified from Rey Auditory Verbal Learning Test (19) and California Verbal Learning Test (20) to be suited to Chinese characters. The total learning score for each subject, that is, the mean score of three learning-free recall trials of 12 words, was collected and statistcally analyzed in this study. Visuospatial skills were evaluated with Clock Drawing Test (0–5 clock drawing test rating scale) (19). In the verbal fluency task, participants were instructed to name as many vegetables, animals, and fruits as possible in 1 min for each category (20). Besides, Montreal Cognitive Assessment (MOCA) was adopted in this study to investigate other aspects of cognitive functions and evaluate the global cognitive abilities (21). Statistical analysis
Data analyses were performed using the statistical software package SPSS 13.0 for Windows (SPSS, Chicago, IL, USA). Descriptive statistics were run for neuropsychological tests, DTI measurements, clinical and demographic data. The mean (standard deviation) or the median (interquartile range) was used to express the participants’ characteristics for continuous variables. Categorical variables were expressed as frequencies and/or percentages, and were evaluated with the Chi-squared test or the Fisher’s exact test as appropriate. Continuous 3
Zi et al. Table 1 Demographic and clinical data
Age (years)1 Sex (n (%) female)2 Education (years)3 Vascular risk factors Hyperlipidemia (n (%))2 Hypertension (n (%))2 Diabetes mellitus (n (%))4 Current smoker (n (%))4
Table 2 Performance of cognitive function tests
PWMHS (n = 16)
Controls (n = 16)
P
62.00 (4.89) 9 (56.3) 8 (6.25–10.25)
61.46 (3.68) 9 (56.2) 9 (6–12)
0.12 0.24 0.23
8 10 3 3
(50.0) (62.5) (18.8) (18.8)
8 12 4 4
(50.1) (60) (20.2) (20)
0.32 0.68 0.28 0.28
PWMHS, periventricular white matter hyperintensities; Values are means (standard deviations) in Student’s t-test or medians (interquartile range) in Mann-Whitney test for continuous variables. Values are n (%) for categorical variables in chi-square test and Fisher’s exact test. P value is the result of statistical comparison between participants with and without periventricular white matter lesions. 1 Student’s t-test. 2 Chi-square test. 3 Mann–Whitney U-test. 4 Fisher’s exact test.
variables were analyzed with Student’s t-test for independent samples or the Mann–Whitney U-test as appropriate. Associations of clinical and demographic variables with cognitive function were measured using Student’s t-test, Pearson or Spearman rank correlation coefficients to clarify potential confounding variables that might influence the association of the DTI measurements with cognitive function. Pearson correlations were calculated to examine the associations of the DTI measurements or cortical thickness with cognitive function, respectively. To explore the relationships among PWMHS, cortical thickness, and cognitive performance, we performed a multiple linear regression analysis, after controlling for age, gender, years of education and treatable vascular risk factors.
Executive function TMT (B–A) Attention Digit span forward test Verbal fluency Category fluency test Visuospatial skills Clock drawing test Verbal memory Total learning score of HAVLT
PWMHS (n = 16)
Controls (n = 16)
P
88.0 (10.1)
51.2 (12.5)
0.001*
5.0 (0.9)
5.5 (0.5)
0.124
23.6 (3.5)
37.8 (6.7)
0.003*
2.9 (0.8)
3.0 (0.7)
0.226
16.68 (4.2)
17.64 (4.6)
0.112
TMT (B-A): Score difference of Part B subtracting Part A in Trail-Making Test; HAVLT, Huashan Auditory Verbal Learning Test; Data are presented as mean (standard deviation); P value is the result of statistical comparison by Student’s t-test between participants with and without periventricular white matter lesions. *P < 0.05. Table 3 Measurements of fraction anisotropy of region-of-interest PWMHS (n = 16)
Controls (n = 16)
Diffusion tensor measurements in periventricular white matter Anterior right 0.31 (0.04) 0.39 (0.07) Anterior left 0.31 (0.03) 0.40 (0.06) Posterior right 0.45 (0.05) 0.48 (0.07) Posterior left 0.44 (0.05) 0.48 (0.08) Diffusion tensor measurements in deep white matter Anterior right 0.43 (0.04) 0.44 (0.06) Anterior left 0.48 (0.04) 0.47 (0.08) Posterior right 0.43 (0.04) 0.44 (0.05) Posterior left 0.42 (0.04) 0.43 (0.04) Temporal right 0.45 (0.03) 0.46 (0.07) Temporal left 0.49 (0.05) 0.48 (0.05) Diffusion tensor measurements in corpus callosum Genu 0.69 (0.04) 0.79 (0.06) Splenium 0.74 (0.06) 0.74 (0.06)
P 0.01* 0.01* 0.04* 0.03* 0.24 0.32 0.48 0.38 0.56 0.61 0.03* 0.26
FA, fraction anisotropy (standard deviation); Values are means (standard deviations); P value is the result of statistical comparison by Student’s t-test between participants with and without periventricular white matter lesions. *P < 0.05.
Results Demographic and clinical data
The demographics and relevant clinical information, including mean age, gender, years of education, and vascular risk factors, were listed in Table 1. There were no differences in demographic and clinical variables between groups (Table 1). Cognitive function
Comparing with the controls, the cognitive tests of the subjects with PWMHs showed significant decline in the domains of verbal fluency and executive function (P < 0.05) (Table 2). In addition, the PWMHs subjects had significantly lower MOCA scores for global cognitive function 4
compared with the controls, which were mainly contributed by the domain of verbal fluency and executive function. DTI measurement
The DTI measurements showed a significant intergroup difference with decreased FA in the genu of the corpus callosum and in the bilateral anterior and posterior PWM in the subjects with PWMHS (P < 0.05) (Table 3). Cortical thickness measurement
Comparison of the cortical mantle between the two groups showed significant cortical thinning
The correlation between PWMH and cognitive impairments Table 4 Brain regions demonstrating statistically significant difference in cortical thickness between subjects with PWMHS and controls PWMHS
© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA NEUROLOGICA SCANDINAVICA
Cognitive impairments associated with periventricular white matter hyperintensities are mediated by cortical atrophy Zi W, Duan D, Zheng J. Cognitive impairments associated with periventricular white matter hyperintensities are mediated by cortical atrophy. Acta Neurol Scand: DOI: 10.1111/ane.12262. © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd. Background – Previous studies have shown that white matter lesions (WMLs) is an important risk factor for cognitive impairment, but the underlying mechanisms have not been clarified. Objective – We tested the hypothesis that the cognitive impairments associated with periventricular white matter hyperintensities (PWMHs) on magnetic resonance imaging (MRI) would be mediated by the cortical thinning of corresponding area. Method – Sixteen stroke- and dementia-free subjects with PWMHs and 16 healthy control subjects were enrolled in this study. All participants underwent an examination of cognition, MRI-based cortical thickness measurement and a MRI-DTI scan. Then, the possible relationships among cognitive impairments, PWMHs and the topography of cortical thinning were analyzed. Results – Comparing with the controls, the cognitive tests of the subjects with PWMHs showed significant decline in the domains of verbal fluency and executive function. After accounting for age, gender, years of education, and treatable vascular risk factors related to cognitive performance, cortical thickness had an independent influence on the cognitive impairments, especially in the frontal pole, orbitofrontal cortex, superior and middle frontal gyrus, superior and middle temporal gyrus, insula, and cuneus. Conclusions – Our results suggest that the association between PWMHs and cognitive impairments is mediated by cortical thinning. Introduction
White matter hyperintensities (WMHs), which are commonly observed on MRI scans of elderly people, have been found to increase in prevalence and severity with age (1). They are associated with cognitive dysfunction in patients with mild cognitive impairment or dementia and have recently been identified as a risk factor for dementia (2). Although WMHs may tend to be associated with cognition decline, the independent associations of periventricular white matter hyperintensities (PWMHs) with cognition are inconsistent among studies. In a longitudinal cohort study of normal elderly subjects, PWMHs were associated with the level of cognitive impairment (3), whereas some other studies indicated that
W. Zi, D. Duan, J. Zheng Department of Neurology, Xinqiao Hospital, Third Military Medical University, Chongqing, China
Key words: periventricular white matter hyperintensities; cortical thinning; cognitive impairments; diffusion tension imaging J. Zheng, Department of neurology, Xinqiao Hospital, Third Military Medical University, 183 Xinqiao Street, Chongqing 400037, China Tel./Fax: +86 023 68763203 e-mail: [email protected] Accepted for publication April 16, 2014
PWMHs are not associated with levels of cognitive decline when cortical volume loss is accounted for (4). Therefore, some researchers presumed that the correlation between cognitive impairment and PWMHs is mediated by cortical atrophy. However, the previous analyses regarding the association of PWMHs or cortical thickness with cognitive function have not excluded the possible impacts of other types of WMHs. Therefore, it is necessary to further investigate the association of PWMHs and cortical volume loss with cognitive impairments in the subjects with ‘pure’ PWMHs. Conventional structural MRI techniques, such as T2-weighted and fluid-attenuated inversion recovery (FLAIR) images, provide little information about the severity of the underlying pathological changes in (5). Diffusion tension image 1
Zi et al. (DTI) is a quantitative MRI technique that measures the directionality and mobility of water diffusion in tissue and provides increased sensitivity to detect the structural integrity of white matter fibers (6). Fractional anisotropy (FA) is a measurement of the diffusion consistency in all direction (7). The present study evaluated PWMHS with magnetic resonance DTI, the topography of cortical thickness with surface-based morphometry, and cognitive impairments with neuropsychological tests in the stroke- and dementia-free subjects. The relationships among these variables are evaluated to further clarify the association of PWMHs and cortical thinning with cognitive impairments. Methods Subjects
Sixteen stroke- and dementia-free subjects with PWMHS and 16 controls without PWMHS on MRI were consecutively enrolled in this study between June 8, 2012 and August 30, 2013. All of them were clients who accepted regular health checks in the health examination center in Xinqiao Hospital, Third Military Medical University and voluntarily participated into this study. This study was approved by the ethics committee of our hospital, and all participants gave their informed consent. All participants were required to be within the age range of 50–75 years and underwent a comprehensive clinical examination by two experienced neurologists, including medical history, physical, and neurological assessments. Subjects PWMHS were evaluated with conventional structural MRI techniques, including T1-weighted, T2weighted, and FLAIR images. Periventricular regions were defined as regions between 3 and 13 mm from the ventricular surface (7). No participants had the disorders that might have confounded their current cognitive state, such as metabolic encephalopathy, thyroid disease, or syphilis, No participants had current or past somatic, psychiatric, or neurological disorders that might have caused the cognitive impairment, such as stroke, schizophrenia, epilepsy, severe head trauma, encephalitis, brain tumors, alcohol abuse, severe depression, or neurodegenerative diseases such as Parkinson’s disease. Control subjects, matched with the PWMHS subjects for age, gender, degree of education, and vascular risk factors, showed no WMH on MRI imaging, no neurological disorders, and no any deficits on the neuropsychological test battery. 2
Acquisition of MR images
Images were acquired with a 3.0-Tesla GE Excite MRI system (GE, Milwaukee, WI, USA). We performed routine whole brain MRI scans for every subject, including axial T1-weighted MRI (acquisition matrix = 320 9 320, TR = 5600 ms, TE = 2.46 ms, slice thickness = 5 mm), axial T2weighted MRI (acquisition matrix = 320 9 320, TR = 5600 ms, TE = 90 ms, slice thickness = 5 mm), and axial FLAIR (acquisition matrix = 320 9 320, TR = 9000 ms, TI = 2250 ms, TE = 85 ms, slice thickness = 5 mm). DTI measurement and analysis
Post-processing of the DTI data was performed on a GE workstation (FuncTool, Advantage Workstation 4.2, GE, Medical System, Milwaukee, WI, USA). First, fractional anisotropy (FA) maps and color-coded directionality maps of diffusion were created in individual images. The color-coded directional maps provided easy visualization of the white matter fiber tracts. Brickshaped regions of interest (ROIs) were drawn based on the identification of white matter tracts on the color-coded maps. Then FA maps were overlaid on these color-coded directional maps. We placed 12 ovoid ROIs in the white matter regions on two slices, which included the temporal deep white matter adjacent to the temporal horns, the genu and splenium of the corpus callosum, the anterior deep white matter, the anterior periventricular white matter, and the posterior deep white matter and the posterior periventricular white matter (see Fig. 1). The non-midline regions were measured on both sides. For most regions, we used the ROI designations described by Fellgiebel et al. (8). The size of the ROIs was 8–12 voxels, measuring 1.878 mm 9 1.875 mm 9 5 mm each (9). Cortical thickness analysis
The automatic cortical thickness analysis of the MRI was conducted using the pipelining method developed at Montreal Neurological Institute (10). First, intensity-based inuniformities, which resulted from the inhomogeneities in the magnetic field, were corrected with the N3 algorithm and the brain tissue was classified into gray matter, white matter, and cerebrospinal fluid, using a 3D stereotaxic brain mask and the intensity-normalized stereotaxic environment for classification of tissues algorithm (11). Next, the brain was automatically divided into two separate hemispheres
The correlation between PWMH and cognitive impairments
Figure 1. Topographic illustration of ROIs measurements in fractional anisotropy maps. Twelve spots for ROIs measurements were selected. Brain regions are indicated as: 1 and 2 = Right and left temporal deep white matter, 3 and 4 = genu and splenium of corpus callosum, 5 and 6 = right and left anterior medial deep white matter, 9 and 10 = right and left posterior deep white matter, 7 and 8 = right and left anterior periventricular white matter, and 11 and 12 = right and left posterior periventricular white matter.
and the inner and outer surfaces of the cortex were automatically extracted, using the Constrained Laplacian-Based Automated Segmentation With proximities (CLASP) algorithm. The CLASP algorithm extracts the boundaries of the cortex by using a dense polygon mesh that is first expanded inside the white matter and fitted into the surface between the white matter and the inner surface of cortex. The outer surface of the cortex is detected by expanding the polygon mesh further into the surface between the gray matter and cerebrospinal fluid. Cortical thickness was defined as the Euclidean distance between the linked vertices of the outer and inner surfaces. Finally, the cortical thickness maps were smoothened using a surface-based 20 mm WHM diffusion kernel that has been demonstrated to increase the signal-to-noise ratio and the statistical power (12). The cortical thickness was calculated in the native space through applying an inverse transformation matrix to cortical surfaces and reconstructing in the native space (13). Neuropsychological tests
All participants completed an extensive neuropsychological test battery. Cognitive measurements were focused on five cognitive domains: executive function, attention, verbal memory, visuospatial skills and verbal fluency. Executive function was assessed with Trail-Making Test, including A and B parts. Both part of the test consists of 25 circles distributed over a sheet of paper. The circles are numbered from 1 to 25 in part A, whereas the circles include both numbers (from 1 to 13) and letters (from A to L) in part B. Participants were required to link all the numbers or letters in digit sequence (part A) or alternating digit
sequence and alphabetical order (part B) as fast as possible. The final score for analysis in this study was calculated by part B score subtracting part A score (14, 15). Attention abilities were examined with Digit Span Forward from the Wechsler Adult Intelligence Scale 3rd edition (WAIS-III) (16, 17). Verbal memory was assessed with Huashan Auditory Verbal Learning Test (HAVLT) in Chinese (18), which was modified from Rey Auditory Verbal Learning Test (19) and California Verbal Learning Test (20) to be suited to Chinese characters. The total learning score for each subject, that is, the mean score of three learning-free recall trials of 12 words, was collected and statistcally analyzed in this study. Visuospatial skills were evaluated with Clock Drawing Test (0–5 clock drawing test rating scale) (19). In the verbal fluency task, participants were instructed to name as many vegetables, animals, and fruits as possible in 1 min for each category (20). Besides, Montreal Cognitive Assessment (MOCA) was adopted in this study to investigate other aspects of cognitive functions and evaluate the global cognitive abilities (21). Statistical analysis
Data analyses were performed using the statistical software package SPSS 13.0 for Windows (SPSS, Chicago, IL, USA). Descriptive statistics were run for neuropsychological tests, DTI measurements, clinical and demographic data. The mean (standard deviation) or the median (interquartile range) was used to express the participants’ characteristics for continuous variables. Categorical variables were expressed as frequencies and/or percentages, and were evaluated with the Chi-squared test or the Fisher’s exact test as appropriate. Continuous 3
Zi et al. Table 1 Demographic and clinical data
Age (years)1 Sex (n (%) female)2 Education (years)3 Vascular risk factors Hyperlipidemia (n (%))2 Hypertension (n (%))2 Diabetes mellitus (n (%))4 Current smoker (n (%))4
Table 2 Performance of cognitive function tests
PWMHS (n = 16)
Controls (n = 16)
P
62.00 (4.89) 9 (56.3) 8 (6.25–10.25)
61.46 (3.68) 9 (56.2) 9 (6–12)
0.12 0.24 0.23
8 10 3 3
(50.0) (62.5) (18.8) (18.8)
8 12 4 4
(50.1) (60) (20.2) (20)
0.32 0.68 0.28 0.28
PWMHS, periventricular white matter hyperintensities; Values are means (standard deviations) in Student’s t-test or medians (interquartile range) in Mann-Whitney test for continuous variables. Values are n (%) for categorical variables in chi-square test and Fisher’s exact test. P value is the result of statistical comparison between participants with and without periventricular white matter lesions. 1 Student’s t-test. 2 Chi-square test. 3 Mann–Whitney U-test. 4 Fisher’s exact test.
variables were analyzed with Student’s t-test for independent samples or the Mann–Whitney U-test as appropriate. Associations of clinical and demographic variables with cognitive function were measured using Student’s t-test, Pearson or Spearman rank correlation coefficients to clarify potential confounding variables that might influence the association of the DTI measurements with cognitive function. Pearson correlations were calculated to examine the associations of the DTI measurements or cortical thickness with cognitive function, respectively. To explore the relationships among PWMHS, cortical thickness, and cognitive performance, we performed a multiple linear regression analysis, after controlling for age, gender, years of education and treatable vascular risk factors.
Executive function TMT (B–A) Attention Digit span forward test Verbal fluency Category fluency test Visuospatial skills Clock drawing test Verbal memory Total learning score of HAVLT
PWMHS (n = 16)
Controls (n = 16)
P
88.0 (10.1)
51.2 (12.5)
0.001*
5.0 (0.9)
5.5 (0.5)
0.124
23.6 (3.5)
37.8 (6.7)
0.003*
2.9 (0.8)
3.0 (0.7)
0.226
16.68 (4.2)
17.64 (4.6)
0.112
TMT (B-A): Score difference of Part B subtracting Part A in Trail-Making Test; HAVLT, Huashan Auditory Verbal Learning Test; Data are presented as mean (standard deviation); P value is the result of statistical comparison by Student’s t-test between participants with and without periventricular white matter lesions. *P < 0.05. Table 3 Measurements of fraction anisotropy of region-of-interest PWMHS (n = 16)
Controls (n = 16)
Diffusion tensor measurements in periventricular white matter Anterior right 0.31 (0.04) 0.39 (0.07) Anterior left 0.31 (0.03) 0.40 (0.06) Posterior right 0.45 (0.05) 0.48 (0.07) Posterior left 0.44 (0.05) 0.48 (0.08) Diffusion tensor measurements in deep white matter Anterior right 0.43 (0.04) 0.44 (0.06) Anterior left 0.48 (0.04) 0.47 (0.08) Posterior right 0.43 (0.04) 0.44 (0.05) Posterior left 0.42 (0.04) 0.43 (0.04) Temporal right 0.45 (0.03) 0.46 (0.07) Temporal left 0.49 (0.05) 0.48 (0.05) Diffusion tensor measurements in corpus callosum Genu 0.69 (0.04) 0.79 (0.06) Splenium 0.74 (0.06) 0.74 (0.06)
P 0.01* 0.01* 0.04* 0.03* 0.24 0.32 0.48 0.38 0.56 0.61 0.03* 0.26
FA, fraction anisotropy (standard deviation); Values are means (standard deviations); P value is the result of statistical comparison by Student’s t-test between participants with and without periventricular white matter lesions. *P < 0.05.
Results Demographic and clinical data
The demographics and relevant clinical information, including mean age, gender, years of education, and vascular risk factors, were listed in Table 1. There were no differences in demographic and clinical variables between groups (Table 1). Cognitive function
Comparing with the controls, the cognitive tests of the subjects with PWMHs showed significant decline in the domains of verbal fluency and executive function (P < 0.05) (Table 2). In addition, the PWMHs subjects had significantly lower MOCA scores for global cognitive function 4
compared with the controls, which were mainly contributed by the domain of verbal fluency and executive function. DTI measurement
The DTI measurements showed a significant intergroup difference with decreased FA in the genu of the corpus callosum and in the bilateral anterior and posterior PWM in the subjects with PWMHS (P < 0.05) (Table 3). Cortical thickness measurement
Comparison of the cortical mantle between the two groups showed significant cortical thinning
The correlation between PWMH and cognitive impairments Table 4 Brain regions demonstrating statistically significant difference in cortical thickness between subjects with PWMHS and controls PWMHS
