Kun-Hsien Chou

Diffusion Tensor Tractography Reveals Abnormal Topological Organization in Structural Cortical Networks in Alzheimer's Disease

Recent research on Alzheimers disease (AD) has shown that the decline of cognitive and memory functions is accompanied by a disrupted neuronal connectivity characterized by white matter (WM) degeneration. However, changes in the topological organization of WM structural network in AD remain largely unknown. Here, we used diffusion tensor image tractography to construct the human brain WM networks of 25 AD patients and 30 age- and sex-matched healthy controls, followed by a graph theoretical analysis. We found that both AD patients and controls had a small-world topology in WM network, suggesting an optimal balance between structurally segregated and integrative organization. More important, the AD patients exhibited increased shortest path length and decreased global efficiency in WM network compared with controls, implying abnormal topological organization. Furthermore, we showed that the WM network contained highly connected hub regions that were predominately located in the precuneus, cingulate cortex, and dorsolateral prefrontal cortex, which was consistent with the previous diffusion-MRI studies. Specifically, AD patients were found to have reduced nodal efficiency predominantly located in the frontal regions. Finally, we showed that the alterations of various network properties were significantly correlated with the behavior performances. Together, the present study demonstrated for the first time that the Alzheimers brain was associated with disrupted topological organization in the large-scale WM structural networks, thus providing the structural evidence for abnormalities of systematic integrity in this disease. This work could also have implications for understanding how the abnormalities of structural connectivity in AD underlie behavioral deficits in the patients.

Atypical Development of White Matter Microstructure in Adolescents with Autism Spectrum Disorders

Diffusion tensor imaging (DTI) studies in adolescents with autism spectrum disorders (ASD) indicate aberrant neurodevelopment of frontal white matter (WM), potentially underlying abnormal social cognition and communication in ASD. Here, we further use tract-based spatial statistics (TBSS) to examine the developmental change of WM skeleton (i.e., the most compact whole-brain WM) during adolescence in ASD. This whole-brain DTI used TBSS measures fractional anisotropy (FA) and longitudinal and radial diffusivities in fifty adolescents, 25 ASD and 25 controls. Results show that adolescents with ASD versus controls had significantly reduced FA in the right posterior limb of internal capsule (increased radial diffusivity distally and reduced longitudinal diffusivity centrally). Adolescents with ASD versus controls (covarying for age and IQ) had significantly greater FA in the frontal lobe (reduced radial diffusivity), right cingulate gyrus (reduced radial diffusivity), bilateral insula (reduced radial diffusivity and increased longitudinal diffusivity), right superior temporal gyrus (reduced radial diffusivity), and bilateral middle cerebellar peduncle (reduced radial diffusivity). Notably, a significant interaction with age by group was found in the right paracentral lobule and bilateral superior frontal gyrus as indicated by an age-related FA gain in the controls whilst an age-related FA loss in the ASD. To our knowledge, this is the first study to use TBSS to examine WM in individuals with ASD. Our findings indicate that the frontal lobe exhibits abnormal WM microstructure as well as an aberrant neurodevelopment during adolescence in ASD, which support the frontal disconnectivity theory of autism.

Does feature selection improve classification accuracy? Impact of sample size and feature selection on classification using anatomical magnetic resonance images

There are growing numbers of studies using machine learning approaches to characterize patterns of anatomical difference discernible from neuroimaging data. The high-dimensionality of image data often raises a concern that feature selection is needed to obtain optimal accuracy. Among previous studies, mostly using fixed sample sizes, some show greater predictive accuracies with feature selection, whereas others do not. In this study, we compared four common feature selection methods. 1) Pre-selected region of interests (ROIs) that are based on prior knowledge. 2) Univariate t-test filtering. 3) Recursive feature elimination (RFE), and 4) t-test filtering constrained by ROIs. The predictive accuracies achieved from different sample sizes, with and without feature selection, were compared statistically. To demonstrate the effect, we used grey matter segmented from the T1-weighted anatomical scans collected by the Alzheimers disease Neuroimaging Initiative (ADNI) as the input features to a linear support vector machine classifier. The objective was to characterize the patterns of difference between Alzheimers disease (AD) patients and cognitively normal subjects, and also to characterize the difference between mild cognitive impairment (MCI) patients and normal subjects. In addition, we also compared the classification accuracies between MCI patients who converted to AD and MCI patients who did not convert within the period of 12 months. Predictive accuracies from two data-driven feature selection methods (t-test filtering and RFE) were no better than those achieved using whole brain data. We showed that we could achieve the most accurate characterizations by using prior knowledge of where to expect neurodegeneration (hippocampus and parahippocampal gyrus). Therefore, feature selection does improve the classification accuracies, but it depends on the method adopted. In general, larger sample sizes yielded higher accuracies with less advantage obtained by using knowledge from the existing literature.

Structural and cognitive deficits in remitting and non-remitting recurrent depression: A voxel-based morphometric study

Remission is the optimal outcome for major depressive disorder (MDD), but many patients do not improve appreciably despite treatment with medication. Treatment-resistant patients may experience deterioration in cognitive functions. Research has reported structural abnormalities in certain brain areas that may contribute to a poor clinical response. We hypothesize that there will be structural differences between patients able to achieve remission and those responding poorly to antidepressants. In the first voxel-based morphometric (VBM) study comparing remitting with non-remitting MDD, we investigated gray matter volume (GMV) differences between depressives to determine which structural abnormalities existed, and correlated these with diminished cognitive functioning. Of 44 adults with recurrent MDD, 19 had full remissions and 25 were non-remitters after a 6-week trial with antidepressant treatment. Remission was defined by 17-item Hamilton Depression Rating Scale scores of </=7 for at least 2 weeks. VBM and neuropsychological studies were conducted on all patients and 25 healthy controls. The patients who remitted revealed milder visual attention deficits than did controls. This correlated with reduced GMV in the left postcentral gyrus (Brodmann area, or BA, 3) and the bilateral medial/superior frontal gyrus (BA 6). The non-remitting patients had reduced GMV in the left dorsolateral prefrontal cortex (DLPFC, BA 9), and impaired acoustic and visual attention associated with GMV differences in several cortical regions, thalamus and amygdala/parahippocampal gyrus. These findings indicated that patients whose MDD remitted were cognitively and morphologically different from non-remitters. Voxel-based structural deficits in the left DLPFC may characterize a subgroup of people with recurrent MDD who respond poorly to antidepressants.

Probabilistic topography of human corpus callosum using cytoarchitectural parcellation and high angular resolution diffusion imaging tractography

The function of the corpus callosum (CC) is to distribute perceptual, motor, cognitive, learned, and voluntary information between the two hemispheres of the brain. Accurate parcellation of the CC according to fiber composition and fiber connection is of upmost important. In this work, population‐based probabilistic connection topographies of the CC, in the standard Montreal Neurological Institute (MNI) space, are estimated by incorporating anatomical cytoarchitectural parcellation with high angular resolution diffusion imaging (HARDI) tractography. First, callosal fibers are extracted using multiple fiber assignment by continuous tracking algorithm based on q‐ball imaging (QBI), on 12 healthy and young subjects. Then, the fiber tracts are aligned in the standard MNI coordinate system based on a tract‐based transformation scheme. Next, twenty‐eight Brodmanns areas on the surface of cortical cortex are registered to the MNI space to parcellate the aligned callosal fibers. Finally, the population‐based topological subdivisions of the midsagittal CC to each cortical target are then mapped. And the resulting subdivisions of the CC that connect to the frontal and somatosensory associated cortex are also showed. To our knowledge, it is the first topographic subdivisions of the CC done using HARDI tractography and cytoarchitectonic information. In conclusion, this sophisticated topography of the CC may serve as a landmark to further understand the correlations between the CC, brain intercommunication, and functional cytoarchitectures. Hum Brain Mapp 2009.

Abnormal topological organization of structural brain networks in schizophrenia

Schizophrenia is a debilitating mental disorder characterized by disturbances of thought and emotion as well as neurocognitive deficits. It is hypothesized that the core symptoms of schizophrenia arise from the inability to integrate neural processes segregated across distributed brain regions. Graph theory allows us to verify this hypothesis at large-scale structural network level. In this study, a sample of 101 schizophrenic patients and 101 healthy controls was included. We sought to investigate the abnormality of network topological organization in patients with schizophrenia by using the cortical thickness measurement from magnetic resonance imaging. Brain networks were constructed by thresholding cortical thickness correlation matrices of 78 regions and analyzed using graph theoretical approaches. Compared to healthy controls, patients showed increased characteristic path length and clustering coefficient in the structural cortical networks. Moreover, schizophrenia patients were associated with reduced nodal centrality in several regions of the default network and increased nodal centrality mainly in primary cortex and paralimbic cortex regions. These findings suggest that the structural networks of schizophrenic patients have a less optimal topological organization, resulting in reduced capacity to integrate information across brain regions.

Sex-linked white matter microstructure of the social and analytic brain

Sexual dimorphism in the brain is known to underpin sex differences in neuropsychological behaviors. The white matter (WM) microstructure appears to be coupled with cognitive performances. However, the issues concerning sex differences in WM remains to be determined. This study used the tract-based spatial statistics on diffusion tensor imaging concurrently with the assessments of Empathizing Quotient (EQ) and Systemizing Quotient (SQ) in forty healthy female and forty male adults. Females exhibited greater fractional anisotropy (FA) in the fronto-occipital fasciculus, body of the corpus callosum, and WM underlying the parahippocampal gyrus. Males exhibited larger FA in the bilateral internal capsule, WM underlying the medial frontal gyrus, fusiform gyrus, hippocampus, insula, postcentral gyrus, frontal and temporal lobe. Interestingly, the interaction analysis of dispositional measures by sex showed that females had a positive correlation between FA of the WM underlying the inferior parietal lobule and superior temporal gyrus and EQ but a negative correlation between FA of the occipital and postcentral gyrus and SQ. Males displayed the opposite effect. The findings indicate a sexual dimorphism of WM microstructure. Divergent correlations of WM microstructure and neuropsychological behaviors between sexes may account for the higher prevalence of autism spectrum disorders in males.

Altered gray matter volume in the frontal pain modulation network in patients with cluster headache

&NA; Reduced gray matter volume was found in frontal pain‐modulation areas in patients with cluster headache, suggesting the involvement of insufficient pain‐modulating capacity in this disorder. &NA; Previous functional imaging studies in episodic cluster headache (CH) patients revealed altered brain metabolism concentrated on the central descending pain control system. However, it remains unclear whether changes in brain metabolism during the “in bout” period are due to structural changes and whether these structural changes vary between the “in bout” and “out of bout” periods. To quantify brain structural changes in CH patients, the regional gray matter volume (GMV) was compared among 49 episodic CH patients during the “in bout” period and 49 age‐ and sex‐matched controls. Twelve patients were rescanned during the “out of bout” period to evaluate the changes, if any, between these 2 periods. Compared with healthy controls, CH patients showed significant “in bout” GMV reductions in the bilateral middle frontal, left superior, and medial frontal gyri. Compared to “out of bout” scans, the “in bout” scans revealed significant GMV increases in the left anterior cingulate, insula, and fusiform gyrus. Additionally, compared to healthy controls, the “out of bout” scans revealed a trend of GMV reduction in the left middle frontal gyrus. These affected regions primarily belong to frontal pain modulation areas, and thus these GMV changes may reflect insufficient pain‐modulating capacity in the frontal areas of CH patients.

Impairments in Cognitive Function and Brain Connectivity in Severe Asymptomatic Carotid Stenosis

Background and Purpose— Severe asymptomatic carotid stenosis has been associated with cognitive impairment, but it is unknown whether this association is attributable to effects on brain connectivity. We present cognitive network abnormalities in a group of patients at a presymptomatic stage. Methods— Seventeen patients with ≥70% asymptomatic stenosis of unilateral internal carotid artery were compared with 26 healthy controls utilizing a comprehensive neuropsychological battery, the dizziness handicap inventory, and multimodality neuroimaging including diffusion tensor imaging and resting-state functional connectivity magnetic resonance imaging. Longitudinally, assessments were completed in a subgroup of 10 patients at 3 months after carotid artery stenting. Results— Compared with the healthy controls, the patients had worse dizziness scores, poorer memory, complex visuo-spatial performances, and lower whole-brain mean fractional anisotropy. The Scheltens scores of leukoaraiosis/infarction were not different between groups. Their seed-based functional connectivity magnetic resonance imaging showed marked decrements of interhemispheric and intrahemispheric, ipsilaterally to carotid stenosis, functional connectivity in the frontoparietal network. In the default mode network, the intrahemispheric functional connectivity was bilaterally impaired. Importantly, the disrupted mean fractional anisotropy in the patients significantly correlated with the attention and verbal memory functions. After successful carotid artery stenting, small but measurable increments of the mean fractional anisotropy and little functional connectivity in the default mode network ipsilateral-to-carotid artery stenting were noted. Conclusions— We identified for the first time distinct patterns of network disruption that correlate with cognitive fragility in patients with asymptomatic carotid stenosis. Brain connectivity may provide early and useful biomarkers for brain ischemia and reperfusion.

White matter abnormalities in schizophrenia patients with tardive dyskinesia: a diffusion tensor image study.

OBJECTIVE Tardive dyskinesia (TD) is a severe side effect of antipsychotics. While increasing evidence suggests that damaged brain microcircuitry of white matter (WM) is responsible for the clinical symptoms in schizophrenia, no reports of WM abnormality associated with TD were noted. METHOD Brain white matter abnormalities were investigated among 20 schizophrenia patients with TD (Schizophrenia with TD group), 20 age-, gender-, and handedness-matched schizophrenic patients without TD (Schizophrenia without TD group), and 20 matched healthy subjects with magnetic resonance imaging and diffusion tensor imaging analysis. Voxel-wise analysis was used to compare fractional anisotropy (FA) maps of the white matter following intersubject registration to Talairach space. Clinical ratings included the Positive and Negative Symptoms Scale (PANSS), Abnormal Involuntary Movement Scale (AIMS), and the Simpson-Angus Scale (SAS). RESULTS The study subjects were 75% female with average of 40.1+/-9. 8 years. The Schizophrenia with TD group had significantly higher PANSS total scores (p=0.024), PANSS negative score (p=0.001), SAS (p<0.001) and AIMS (p<0.001) scores; and demonstrated more widespread FA decreases than the Schizophrenia without TD group, especially over the inferior frontal gyrus, temporal sublobar extranuclear WM (around the basal ganglion), parietal precuneus gyrus WM (around somatosensory cortex), and medial frontal gyrus WM (around dorsolateral prefrontal cortex). The AIMS (p<0.01) and SAS (p<0.01) score positively correlated with decreased FA over these areas, and PANSS negative score positively correlated with FA decrease over medial frontal gyrus WM (p<0.01). CONCLUSIONS More widespread abnormality of white matter was noted among schizophrenia patients than those without, especially involved cortico-basal ganglion circuits with clinical symptom correlation of involuntary movements and negative symptoms. Further studies with larger sample size are required to validate the findings.