@article {907, title = {Multimodal fusion analysis of structural connectivity and gray matter morphology in migraine}, journal = {Human Brain Mapping}, volume = {42}, year = {2021}, pages = {908-921}, abstract = {

No specific migraine biomarkers have been found in single-modality MRI studies. We aimed at establishing biomarkers for episodic and chronic migraine using diverse MRI modalities. We employed canonical correlation analysis and joint independent component analysis to find structural connectivity abnormalities that are related to gray matter morphometric alterations. The number of streamlines (trajectories of estimated fiber-tracts from tractography) was employed as structural connectivity measure, while cortical curvature, thickness, surface area, and volume were used as gray matter parameters. These parameters were compared between 56 chronic and 54 episodic migraine patients, and 50 healthy controls. Cortical curvature alterations were associated with abnormalities in the streamline count in episodic migraine patients compared to controls, with higher curvature values in the frontal and temporal poles being related to a higher streamline count. Lower streamline count was found in migraine compared to controls in connections between cortical regions within each of the four lobes. Higher streamline count was found in migraine in connections between subcortical regions, the insula, and the cingulate and orbitofrontal cortex, and between the insula and the temporal region. The connections between the caudate nucleus and the orbitofrontal cortex presented worse connectivity in chronic compared to episodic migraine. The hippocampus was involved in connections with higher and lower number of streamlines in chronic migraine. Strengthening of structural networks involving pain processing and subcortical regions coexists in migraine with weakening of cortical networks within each lobe. The multimodal analysis offers a new insight about the association between brain structure and connectivity.

}, keywords = {Brain, Magnetic Resonance Imaging, connectome, diffusion magnetic resonance imaging, migraine disorders}, doi = {https://doi.org/10.1002/hbm.25267}, url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/hbm.25267}, author = {{\'A}lvaro Planchuelo-G{\'o}mez and Garc{\'\i}a-Azor{\'\i}n, David and {\'A}ngel L. Guerrero and Santiago Aja-Fern{\'a}ndez and Rodr{\'\i}guez, Margarita and Rodrigo de Luis-Garc{\'\i}a} } @article {423, title = {Automatic bayesian classification of healthy controls, bipolar disorder, and schizophrenia using intrinsic connectivity maps from fMRI data}, journal = {IEEE Transactions on Biomedical Engineering}, volume = {57}, year = {2010}, pages = {2850-2860}, abstract = {

We present a method for supervised, automatic, and reliable classification of healthy controls, patients with bipolar disorder, and patients with schizophrenia using brain imaging data. The method uses four supervised classification learning machines trained with a stochastic gradient learning rule based on the minimization of KullbackLeibler divergence and an optimal model complexity search through posterior probability estimation. Prior to classification, given the high dimensionality of functional MRI (fMRI) data, a dimension reduction stage comprising two steps is performed: first, a one-sample univariate t-test mean-difference Tscore approach is used to reduce the number of significant discriminative functional activated voxels, and then singular value decomposition is performed to further reduce the dimension of the input patterns to a number comparable to the limited number of subjects available for each of the three classes. Experimental results using functional brain imaging (fMRI) data include receiver operation characteristic curves for the three-way classifier with area under curve values around 0.82, 0.89, and 0.90 for healthy control versus nonhealthy, bipolar disorder versus nonbipolar, and schizophrenia patients versus nonschizophrenia binary problems, respectively. The average three-way correct classification rate (CCR) is in the range of 70\%-72\%, for the test set, remaining close to the estimated Bayesian optimal CCR theoretical upper bound of about 80\%, estimated from the one nearest-neighbor classifier over the same data. {\^A}{\textcopyright} 2010 IEEE.

}, keywords = {Algorithms, Artificial Intelligence, Bayes Theorem, Bayesian learning, Bayesian networks, Biological, Brain, Case-Control Studies, Classifiers, Computer-Assisted, Diseases, Functional MRI (fMRI), Humans, Learning machines, Learning systems, Magnetic Resonance Imaging, Models, Operation characteristic, Optimization, ROC Curve, Reproducibility of Results, Signal Processing, Singular value decomposition, Statistical tests, Stochastic models, Student t test, area under the curve, article, bipolar disorder, classification, controlled study, functional magnetic resonance imaging, human, machine learning, major clinical study, neuroimaging, patient coding, receiver operating characteristic, reliability, schizophrenia}, issn = {00189294}, doi = {10.1109/TBME.2010.2080679}, url = {http://www.scopus.com/inward/record.url?eid=2-s2.0-78649311169\&partnerID=40\&md5=d3b90f1a3ee4ef209d131ef986e142db}, author = {J I Arribas and V D Calhoun and T Adali} }