Objective:We use resting state functional magnetic resonance imagingto evaluate whether the globus pallidus based functional connectivity changedin patients with carbon monoxide poisoning. And the characteristic of thechanged functional connectivity map. Three months later, the patients werescanned twice to see the functional connectivity changes.Methods:We recruited8patients with carbon monoxide poisoning,10age-gender-and education-matched healthy controls and a large sample groupof43college students for resting state fMRI study. The globus pallidus ROIswere defined by Harvard-Oxford Structural Atlas. In order to make sure thatthe ROIs were located within the globus pallidus and therefore to reduce theconfounding effect from adjacent non-globus pallidus signals, the thresholdwas set as80%for each side of globus pallidus. Two smaller globus pallidusROIs were generated. The Globus pallidus ROIs were resampled to3×3×3mm3for further study. Since the Harvard-Oxford Structural Atlas was notsymmetric in the two hemispheres, the size of the two ROIs in our study wasalso different, with a cluster of32voxels (864mm3) in left globus pallidusROI and a cluster of33voxels (891mm3) in right globus pallidus ROI. UsingDPARSFA2.3, REST1.8and SPM8to process the raw data for each group toget the globus pallidus based functional connectivity map. Then one sample ttest was performed to achieve the significant functional connectivity map anda combined binary mask. Finally, two sample t test between patients withcarbon monoxide poisoning and healthy controls was performed within themask to reveal the abnormal functional connectivity. Three patients haveeffective follow up. Statistical analysis was not taken due to the small samplesize. In the individual level, the mean signal of each sphere ROI with4mmradius were extracted, which centered the most significant different voxel in each cluster between patients and healthy controls. Then use a line chart toexhibit the trend of the two scans.Results:One sample t test within group analysis showed that some brain areasshowed positive functional connectivity with bilateral globus pallidus,including bilateral putman, globus pallidus, caudate, thalamus, insula,precentral cortex, anterior cingulated cortex and some areas in the cerebellum.Some brain regions showed negative functional connectivity with bilateralglobus pallidus, which mainly in the default model network, includingbilateral posterior cingulated, precunes, inferior parietal lobe, and medialprefrontal cortex. The within group results in patients and healthy controlswere less than that in large group, mainly in basial ganglia and visual cortex.Two sample t test between groups analysis showed that, compared to healthycontrols, bilateral putman, precentral cortex, insula showed increased positivefunctional connectivity, bilateral visual cortex showed reduced negativefunctional connectivity to left globus pallidus in patients. Right globuspallidus showed increased positive functional connectivity with bilateralputman, insula; increased negative functional connectivity with left visualcortex; reduced positive functional connectivity with superior frontal lobe;reduced negative functional connectivity with posterior cingulate cortex. Forthe three patients who scaned twice, the trend of functional connectivitychanges went toward normal. Some regions which showed higher positivefunctional connectivity with left globus pallidus than healthy controlsdecreased, including bilateral putman, bilateral precentral cortex. In the firstscan, left visual cortex of patients showed stronger negative functionalconnectivity with left globus pallidus than healthy controls, decreased in theall of the three patients in the second scan. Genially speaking, regions whichshowed significant difference of functional connectivity with right globuspallidus with healthy controls went toward nomarl, but less consistent in threepatients than left globud pallidus. Conclusion:1The globus pallidus based functional connectivity pattern were mainlycontained the positive connectivity with basal ganglia, thalamus, frontal lobeand precentral cortex, and negative connectivity with default model network.2After carbon monoxide poisoning, the basal ganglia circuits wascompensatory increased in patients, the change of compensatory may indicatethe prognosis and guide the treatment.3After carbon monoxide poisoning, the frontal-striatal circuits wasdecreased in patients, which indicate that high level brain function likecognition, mental and behavior injury.4Increased task related network deduce decreased connectivity withdefault model network. The high level brain function like self attention focus,internal thoughts, and episodic memory retrieval may decreased in patients.5Resting-state functional MRI has an important clinical value inevaluating the subclinical high-level brain function disorders and guiding thetreatment and rehabilitation of patients with carbon monoxide poisoning. |