A FMRI Study Of Functional Connectivity Of Brain Default Mode Network And Sensorimotor Network In Patients With Chronic Cerebral Infarction

Objective:To investigate intrinsic mechanism of cognitive impairment and motor recovery after stroke, we explored the changes within and between default mode network(DMN) and sensorimotor mode network(SMN) in patients with chronic basal ganglia infarction using functional connectivity(FC) of resting state functional magnetic resonance imaging(rs-f MRI).Subjects and Methods:Seventeen well-recovered patients(13 males and 4 females) with unilateral chronic basal ganglia infarction and twenty two sex, age and level of education-matched healthy controls(11 males and 11 females) were scanned with rs-f MRI. Montreal Cognitive Assessment(Mo CA) and Mini-Mental State Examination(MMSE) scores were recorded.1. All patients and health controls were performed GE 3.0T HD-X MR scan with a 8-channel head coil to obtain whole brain high resolution T1 WI anatomy images and rs-f MRI data.2. Before the rs-f MRI data preprocessing, the lesions located in the left basal ganglia area were flipped into the right corresponding area to create a mirror image using the Matlab software. We defined the right hemisphere as ipsilesional hemisphere, the left hemisphere as contralesional hemisphere. The SPM8 and DPARSF softwares which are based on the Matlab platform were used to data preprocessing. The data preprocessing included discarded the first 10 volumes, slice timing, realign, normalization, resample 3×3×3 mm3 voxels size, detrend, remove the covariates, band-pass filtering(bandwidth:0.01~0.08Hz) and smooth.3. Firstly, the three typical brain regions in DMN included ventromedial prefrontal cortex(VMPFC), right posterior cingulate cortex(R-PCC) and left posterior cingulate cortex(L-PCC) were selected as three diameter 10 mm spherical ROIs(region of interest) using the REST software. Each ROI of all participants was performed to calculate FC within the whole brain and Fisher-Z transformation. The three positive FC distribution maps were obtained using one-sample t-test(P< 0.05, FWE corrected)and get the ROI of DMN with their intersection. Secondly, voxel-based FC analysis was performed between the ROI of DMN and the voxels of DMN to acquire FC map of DMN. Lastly, one-sample t-test was performed for the FC of DMN within patients and health controls respectively. We also made binarization and union of the two FC maps to acquire DMN mask. Once again the two typical brain regions in SMN included right precentral gyrus and left precentral gyrus were selected as two diameter 10 mm spherical ROIs. The ROI of SMN was acquired using the above same methods and steps. Finally, voxel-based FC analysis was performed to achieve FC map of SMN and SMN mask.4. Altered FC of brain areas were displayed between patients and healthy controls. The group comparison between the patients and healthy controls of FC within DMN and SMN mask was performed by two-sample t-test in SPM8 with age, gender and the level of education as covariates. Multiple comparisons were corrected by Monte Carlo simulation. Then the corrected results were overlapped onto the three-dimension MNI standard template respectively. The location, voxel sizes, MNI coordinate and t value of each statistically significant cluster were recorded.5. For the inter-network analysis, the data which came from step 2 preprocessing procedures were performed to calculate FC analysis between DMN and SMN at the level of ROI-wise. The general linear model(GLM) in SPSS17 statistical software was used for statistic analysis. The group comparison between the patients and healthy controls of FC strengths of inter-network was performed by independent sample t-test with age, gender and the level of education as covariates(P < 0.05).6. The infarction lesions were manually sketched in patients on T2 WI images step by step using MRI cro N software. According to the lesion area and thickness in each layer, volume of infarction lesion was calculated by the software automatically.7. Pearson correlation analysis was performed between FC strengths of patients and Mo CA scores, MMSE scores and volume of infarction lesions.Results:1. Compared with the healthy controls, chronic basal ganglia infarction patients presented significantly decreased FC in the medial prefrontal cortex(MPFC) and posterior cingulate cortex/ precuneus(PCC/PCu)(P < 0.05, Alphasim corrected).2. Compared with the healthy controls, chronic basal ganglia infarction patients presented significantly increased FC in the bilateral supplementary motor areas(SMA)(P < 0.05, Alphasim corrected).3. Compared with the healthy controls, chronic basal ganglia infarction patients presented significantly decreased FC between DMN and SMN(P < 0.05).4. Correlation analysis showed that FC strengths of MPFC within DMN in patients had significantly positive correlation with Mo CA scores(r = 0.619, P = 0.008).Conclusions:1. The damaged DMN after cerebral infarction can happen prior to cognitive dysfunction, which may be the initial factor of cognitive impairment.2. The increased FC of bilateral supplementary motor areas in patients with chronic basal ganglia infarction played a role in functional compensation which promoted the recovery of motor function.3. The decreased FC between DMN and SMN may reflect abnormal internetwork functional interaction.