Brain Network Features Research Articles

Exploring the functional connectivity characteristics of brain networks in post-stroke patients with global aphasia: a healthy control based resting-state fMRI study.

In this study, resting-state functional magnetic resonance imaging (rs-fMRI) was used to investigate the characteristics of functional connectivity of brain networks in patients with post-stroke global aphasia (PGA). PGA patients hospitalized in Wuxi Tongren Rehabilitation Hospital during their subacute stage were selected as a case group, and healthy volunteers with matching age, sex, and education level were selected as healthy controls (HCs). rs-fMRI scans were performed to compare the differences of functional connectivity in resting-state networks (RSNs) and in the whole brain between the two groups. A total of 11 patients with PGA and 11 HCs were included in this study. PGA patients showed decreased inter-hemispheric connectivity of homologs within the sensorimotor network (SMN), salience network, and language network. In the analysis of the whole brain functional connections, PGA patients exhibited both inter-hemispheric and intra-hemispheric hypoconnectivity when compared with HCs. However, they exhibited some stronger connections than HCs between the language network and cerebellar network, which may indicate compensatory mechanisms. Using rs-fMRI to research differences in the functional connectivity of brain networks in post-stroke global aphasia will help us further understand it's neurological mechanism and provide an important basis for the accurate selection of therapeutic targets in the future to promote better recovery of language function.

Subthreshold amyloid pathology and changes in functional brain networks in patients with delayed cognitive decline after stroke

AbstractBackgroundIn clinical practice, some of the stroke survivors who have normal cognitive function until three to six months after index stroke, develop cognitive decline gradually after then. Although amyloid deposits and advanced small vessel disease are reported to be the possible predictors, pathogenesis for delayed cognitive decline has not been identified yet. Thus, we aimed to explore the brain network characteristics in those with delayed cognitive decline after stroke compared to those without.MethodUsing a prospective stroke registry, we identified patients with ischemic stroke who had undergone the cognitive evaluations including mini‐mental state examination (MMSE) twice or more after index stroke. We included patients whose baseline cognitive evaluation (between 1 and 12 months after stroke) was normal and who had one or more follow‐up evaluations after more than 6 months from the baseline evaluation. Among 222 eligible patients, decliners were defined as having a decline in 3 or more per year or an absolute decline of 5 or more between baseline and follow‐up MMSE, while non‐decliners as having a decline of 1 point per year or less and an absolute decline of less than 3 points. Brain network attributes were analyzed using resting‐state functional MRI, and superimposed amyloid positivity was assessed using 18F‐florbetaben PET.ResultEleven patients (5%, 11/222) met the criteria for decliners, and ten matched non‐decliners were enrolled. For about 5.7 years of follow‐up, the mean change of MMSE scores was ‐9.1 points in the decliners, while ‐0.2 point in the non‐decliners. Ten (10/11) decliners were amyloid‐negative on the amyloid PET, and the APOE e4 allele was identified in one decliner. However, the global SUVR value showed a negative correlation with the final MMSE scores as well as the amount of MMSE change. Normalized transitivity was lower in decliners compared to non‐decliners. Mean connectivity strengths were averaged based on the seven canonical functional networks, and the intra‐network connections within the frontoparietal network showed the considerable differences between groups. (Figure)ConclusionThis study suggests that alteration of brain networks may explain the pathogenesis of the delayed post‐stroke cognitive decline independent of amyloid PET positivity and APOE genotype.

Idle-state detection in motor imagery of articulation using early information: A functional Near-infrared spectroscopy study

A speech imagery-based brain computer interface (BCI) provides an alternative way for people to interact with the outside world intuitively. Most speech imagery BCIs based on functional near-infrared spectroscopy (fNIRS) are disadvantageous in applications outside the laboratory due to being incapable of detecting asynchronous (self-paced) actions. This work aimed to develop a two-class asynchronous BCI that detects idle state in the context of motor imagery of articulation (MIA). In this study, 19 healthy subjects were asked to rehearse the Chinese vowels/a/and/u/covertly and to accommodate a rest state. A feature selection strategy was designed to combine time domain, spatial domain, and functional connectivity features. All differentiating information was extracted from fNIRS signals in a 0–2.5 s time window. Among single-modality features, the centrality features of brain networks performed better than any other features and yielded a subject-dependent classification accuracy rate of 75.1%. The subject-dependent classification accuracy rate of the combined features at 78.9% was significantly better than that of any single feature. These results demonstrate that it is feasible to detect idle states from active MIA states in a reduced time window size. The proposed combination of information propagation pattern, spatial pattern, and activation pattern was meaningful for extracting discriminative information, hence improving classification accuracy. The high classification accuracy and fast information transfer rate of the presented BCI show promising practicality for real world applications.

Predicting Delayed Neurocognitive Recovery After Non-cardiac Surgery Using Resting-State Brain Network Patterns Combined With Machine Learning.

Delayed neurocognitive recovery (DNR) is a common subtype of postoperative neurocognitive disorders. An objective approach for identifying subjects at high risk of DNR is yet lacking. The present study aimed to predict DNR using the machine learning method based on multiple cognitive-related brain network features. A total of 74 elderly patients (≥ 60-years-old) undergoing non-cardiac surgery were subjected to resting-state functional magnetic resonance imaging (rs-fMRI) before the surgery. Seed-based whole-brain functional connectivity (FC) was analyzed with 18 regions of interest (ROIs) located in the default mode network (DMN), limbic network, salience network (SN), and central executive network (CEN). Multiple machine learning models (support vector machine, decision tree, and random forest) were constructed to recognize the DNR based on FC network features. The experiment has three parts, including performance comparison, feature screening, and parameter adjustment. Then, the model with the best predictive efficacy for DNR was identified. Finally, independent testing was conducted to validate the established predictive model. Compared to the non-DNR group, the DNR group exhibited aberrant whole-brain FC in seven ROIs, including the right posterior cingulate cortex, right medial prefrontal cortex, and left lateral parietal cortex in the DMN, the right insula in the SN, the left anterior prefrontal cortex in the CEN, and the left ventral hippocampus and left amygdala in the limbic network. The machine learning experimental results identified a random forest model combined with FC features of DMN and CEN as the best prediction model. The area under the curve was 0.958 (accuracy = 0.935, precision = 0.899, recall = 0.900, F1 = 0.890) on the test set. Thus, the current study indicated that the random forest machine learning model based on rs-FC features of DMN and CEN predicts the DNR following non-cardiac surgery, which could be beneficial to the early prevention of DNR.Clinical Trial Registration: The study was registered at the Chinese Clinical Trial Registry (Identification number: ChiCTR-DCD-15006096).

EEG-based emotion recognition in an immersive virtual reality environment: From local activity to brain network features

Emotion electroencephalography (EEG) datasets play a significant role in EEG-based emotion recognition research, providing a platform for comparisons of different emotion recognition methods. Most datasets used 2D images or videos as mood induction procedures (MIPs); however, considering the differences in EEG dynamics between 2D and 3D environments, experimental research based on 2D MIPs may have poor results being applied in a real 3D world. In this paper, we (1) developed a new emotion EEG dataset, virtual reality (VR) emotional EEG dataset (VREED), which used 3D VR videos as MIPs; and (2) Presented a baseline for the performance of negative/positive emotion classification in the new dataset. The best average accuracy of 73.77% ± 2.01% was obtained by using the combination of theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and gamma (30–49 Hz) relative power features. Additionally, we observed that occipital and frontal regions played a more critical role than other regions in emotion processing from Spearman correlation analysis and feature selection. This new VR emotion EEG dataset will be publicly available, and we encourage other researchers to evaluate their emotion classification methods on the VREED.

Brain functional networks analysis of five fingers grasping in virtual reality environment.

This study investigated the effects of different center of mass (COM) of the grasping device and visual time-delay on the information interaction between brain regions during five-finger grasping process. Nine healthy right-handed subjects used five fingers to grasp a special device in a virtual reality (VR) environment. Two independent variables were set in the experiment: the COM of the grasping device and the visual delay time. Place a 50 g mass randomly at five different directions of the grasping device base. The three levels of visual delay time appear randomly. The kinematics and dynamics and electroencephalogram (EEG) signals were recorded during the experiment. The brain network was constructed based on multiplex horizontal visibility graph (MHVG). Interlayer mutual information (MI) and phase locking value (PLV) were calculated to quantify the network, while clustering coefficient (C), shortest path length (L) and overall network efficiency (E) are selected to quantify the network characteristic. Statistical results show that when the mass is located in the radial side, during the load phase of grasping, the C and E is significantly higher than that in the proximal, ulnar and medial side, and L was significantly lower than that in the proximal and radial side. This shows that when grasping an object with a COM bias on the radial side, the process of brain feedforward control has higher level of information interaction and ability and it can build stronger sensorimotor memory. It is also found that the brain network features of theta, beta and gamma bands of EEG are positively correlated, especially between beta and gamma bands, which suggests there is a coupling relationship between different bands in information processing and transmission.Clinical Relevance- This study explains the neural mechanism of grasping control from the topological structure of the whole brain network level and the informatics.

Self-paced learning and privileged information based KRR classification algorithm for diagnosis of Parkinson’s disease

Computer aided diagnosis (CAD) methods for Parkinson's disease (PD) can assist clinicians in diagnosis and treatment. Magnetic resonance imaging (MRI) based CAD methods can help reveal structural changes in brain. Classifier is a key component in CAD system, which directly affects the classification performance. Privileged information (PI) can assist to train the classifier by providing additional information, which makes test samples have less error and improves the classification accuracy. In this paper, we proposed a PI based kernel ridge regression plus (KRR+) in diagnosis of PD. Specifically, morphometric features and brain network features are extracted from MRI. Then, empirical kernel mapping feature expression method is used to make the data separable in high-dimensional space. Besides, we introduce self-paced learning that can adaptively select the sample in training of the model, which can further improve the classification performance. The experimental results show that the proposed method is effective for PD diagnosis, its performance is superior to existing classification model. This method is helpful to assist clinicians to find out possible neuroimaging biomarkers in the diagnosis of PD.

Cross-scanner reproducibility and harmonization of a diffusion MRI structural brain network: A traveling subject study of multi-b acquisition

Characterization of brain networks by diffusion MRI (dMRI) has rapidly evolved, and there are ongoing movements toward data sharing and multi-center studies. To extract meaningful information from multi-center data, methods to correct for the bias caused by scanner differences, that is, harmonization, are urgently needed. In this work, we report the cross-scanner differences in structural network analyses using data from nine traveling subjects (four males and five females, 21–49 years-old) who underwent scanning using four 3T scanners (public database available from the Brain/MINDS Beyond Human Brain MRI project (http://mriportal.umin.jp/)). The reliability and reproducibility were compared to those of data from another set of four subjects (all males, 29–42 years-old) who underwent scan-rescan (interval, 105–147 days) with the same scanner as well as scan-rescan data from the Human Connectome Project database. The results demonstrated that the reliability of the edge weights and graph theory metrics was lower for data including different scanners, compared to the scan-rescan with the same scanner. Besides, systematic differences between scanners were observed, indicating the risk of bias in comparing networks obtained from different scanners directly. We further demonstrate that it is feasible to reduce inter-scanner variabilities while preserving the inter-subject differences among healthy individuals by modeling the scanner effects at the level of network matrices, when traveling-subject data are available for calibration between scanners. The present data and results are expected to serve as a basis for developing and evaluating novel harmonization methods.

Investigation of functional brain network reconfiguration during exposure to naturalistic stimuli using graph-theoretical analysis

Objective. One of the most significant features of the human brain is that it can dynamically reconfigure itself to adapt to a changing environment. However, dynamic interaction characteristics of the brain networks in naturalistic scenes remain unclear. Approach. We used open-source functional magnetic resonance imaging (fMRI) data from 15 participants who underwent fMRI scans while watching an audio–visual movie ‘Forrest Gump’. The community detection algorithm based on inter-subject functional correlation was used to study the time-varying functional networks only induced by the movie stimuli. The whole brain reconfiguration patterns were quantified by the temporal co-occurrence matrix that describes the probability of two brain regions engage in the same community (or putative functional module) across time and the time-varying brain modularity. Four graph metrics of integration, recruitment, spatio-temporal diversity and within-community normalised centrality were further calculated to summarise the brain network dynamic roles and hub features in their spatio-temporal topology. Main results. Our results suggest that the networks that were involved in attention and audio–visual information processing, such as the visual network, auditory network, and dorsal attention network, were considered to play a role of ‘stable loners’. By contrast, ‘unstable loner’ networks such as the default mode network (DMN) and fronto-parietal network tended to interact more flexibly with the other networks. In addition, global brain network showed significant fluctuations in modularity. The ‘stable loner’ networks always maintained high functional connectivity (FC) strength while ‘unstable loner’ networks, especially the DMN, exhibited high intra- and inter-network FC only during a low modularity period. Finally, changes in brain modularity were significantly associated with variations in emotions induced by the movie. Significance. Our findings provide new insight for understanding the dynamic interaction characteristics of functional brain networks during naturalistic stimuli.

Social brain network predicts real-world social network in individuals with social anhedonia

Social anhedonia (SA) impairs social functioning in schizophrenia. Previous evidence suggested that certain brain regions predict longitudinal change of real-world social outcomes, yet previous study designs have failed to capture the corresponding functional connectivity among the brain regions involved. This study measured the real-world social network in 22 pairs of individuals with high and low levels of SA, and followed up them for 21 months. We further explored whether resting-state social brain network characteristics could predict the longitudinal variations of real-world social network. Our results showed that social brain network characteristics could predict the change of real-world social networks in both the high SA and low SA groups. However, the results differed between the two groups, i.e., the topological characteristics of the social brain network predicted real-world social network change in the high SA group; whereas the functional connectivity within the social brain network predicted real-world social network change in the low SA group. Principal component analysis and linear regression analysis on the entire sample showed that the functional connectivity component centered at the right orbital inferior frontal gyrus could best predict social network change. Our findings support the notion that social brain network characteristics could predict social network development.

Abnormal Local Spontaneous Nerve Activity and Brain Network Characteristics in Schizophrenia: Potential Biomarkers Based on fMRI

Abnormal Local Spontaneous Nerve Activity and Brain Network Characteristics in Schizophrenia: Potential Biomarkers Based on fMRI

On the Stability of Brain Network Features Through Ageing and Pathologies

On the Stability of Brain Network Features Through Ageing and Pathologies

Network Analysis of Human Brain Connectivity Reveals Neural Fingerprints of a Compositionality Bias in Signaling Systems.

Compositionality is a hallmark of human language and other symbolic systems: a finite set of meaningful elements can be systematically combined to convey an open-ended array of ideas. Compositionality is not uniformly distributed over expressions in a language or over individuals' communicative behavior: at both levels, variation is observed. Here, we investigate the neural bases of interindividual variability by probing the relationship between intrinsic characteristics of brain networks and compositional behavior. We first collected functional resting-state and diffusion magnetic resonance imaging data from a large participant sample (N = 51). Subsequently, participants took part in two signaling games. They were instructed to learn and reproduce an auditory symbolic system of signals (tone sequences) associated with affective meanings (human faces expressing emotions). Signal-meaning mappings were artificial and had to be learned via repeated signaling interactions. We identified a temporoparietal network in which connection length was related to the degree of compositionality introduced in a signaling system by each player. Graph-theoretic analysis of resting-state functional connectivity revealed that, within that network, compositional behavior was associated with integration measures in 2 semantic hubs: the left posterior cingulate cortex and the left angular gyrus. Our findings link individual variability in compositional biases to variation in the anatomy of semantic networks and in the functional topology of their constituent units.

Deep Learning-Based Assessment of Brain Connectivity Related to Obstructive Sleep Apnea and Daytime Sleepiness.

PurposeObstructive sleep apnea (OSA) is associated with altered pairwise connections between brain regions, which might explain cognitive impairment and daytime sleepiness. By adopting a deep learning method, we investigated brain connectivity related to the severity of OSA and daytime sleepiness.Patients and MethodsA cross-sectional design applied a deep learning model on structural brain networks obtained from 553 subjects (age, 59.2 ± 7.4 years; men, 35.6%). The model performance was evaluated with the Pearson’s correlation coefficient (R) and probability of absolute error less than standard deviation (PAE<SD) between the estimated and the actual scores. In addition, we investigated sex effects on deep learning outputs for OSA and daytime sleepiness and examined the differences in brain connectivity related to daytime sleepiness between OSA and non-OSA groups.ResultsWe achieved a meaningful R (up to 0.74) and PAE<SD (up to 0.92) in a test dataset of whole group and subgroups. Motor, frontal and limbic areas, and default mode network were the prominent hubs of important connectivity to predict OSA severity and daytime sleepiness. Sex affected brain connectivity relevant to OSA severity as well as daytime sleepiness. Brain connectivity associated with daytime sleepiness also differed by the presence vs absence of OSA.ConclusionA deep learning method can assess the association of brain network characteristics with OSA severity and daytime sleepiness and specify the relevant brain connectivity.

Locked-in Intact Functional Networks in Children with Autism Spectrum Disorder: A Case-Control Study.

Resting-state functional magnetic resonance imaging (rs-fMRI) has the potential to investigate abnormalities in brain network structure and connectivity on an individual level in neurodevelopmental disorders, such as autism spectrum disorder (ASD), paving the way toward using this technology for a personalized, precision medicine approach to diagnosis and treatment. Using a case-control design, we compared five patients with severe regressive-type ASD to five patients with temporal lobe epilepsy (TLE) to examine the association between brain network characteristics and diagnosis. All children with ASD and TLE demonstrated intact motor, language, and frontoparietal (FP) networks. However, aberrant networks not usually seen in the typical brain were also found. These aberrant networks were located in the motor (40%), language (80%), and FP (100%) regions in children with ASD, while children with TLE only presented with aberrant networks in the motor (40%) and language (20%) regions, in addition to identified seizure onset zones. Fisher’s exact test indicated a significant relationship between aberrant FP networks and diagnosis (p = 0.008), with ASD and atypical FP networks co-occurring more frequently than expected by chance. Despite severe cognitive delays, children with regressive-type ASD may demonstrate intact typical cortical network activation despite an inability to use these cognitive facilities. The functions of these intact cognitive networks may not be fully expressed, potentially because aberrant networks interfere with their long-range signaling, thus creating a unique “locked-in network” syndrome.

Time-varying measures of cerebral network centrality correlate with visual saliency during movie watching.

The extensive development of graph‐theoretic analysis for functional connectivity has revealed the multifaceted characteristics of brain networks. Network centralities identify the principal functional regions, individual differences, and hub structure in brain networks. Neuroimaging studies using movie‐watching have investigated brain function under naturalistic stimuli. Visual saliency is one of the promising measures for revealing cognition and emotions driven by naturalistic stimuli. This study investigated whether the visual saliency in movies was associated with network centrality. The study examined eigenvector centrality (EC), which is a measure of a region's influence in the brain network, and the participation coefficient (PC), which reflects the hub structure in the brain, was used for comparison. Static and time‐varying EC and PC were analyzed by a parcel‐based technique. While EC was correlated with brain activity in parcels in the visual and auditory areas during movie‐watching, it was only correlated with parcels in the visual areas in the retinotopy task. In addition, high PC was consistently observed in parcels in the putative hub both during the tasks and the resting‐state condition. Time‐varying EC in the parietal parcels and time‐varying PC in the primary sensory parcels significantly correlated with visual saliency in the movies. These results suggest that time‐varying centralities in brain networks are distinctively associated with perceptual processing and subsequent higher processing of visual saliency.

Different Modes of Low-Frequency Focused Ultrasound-Mediated Attenuation of Epilepsy Based on the Topological Theory.

Epilepsy is common brain dysfunction, where abnormal synchronized activities can be observed across multiple brain regions. Low-frequency focused pulsed ultrasound has been proven to modulate the epileptic brain network. In this study, we used two modes of low-intensity focused ultrasound (pulsed-wave and continuous-wave) to sonicate the brains of KA-induced epileptic rats, analyzed the EEG functional brain connections to explore their respective effect on the epileptic brain network, and discuss the mechanism of ultrasound neuromodulation. By comparing the brain network characteristics before and after sonication, we found that two modes of ultrasound both significantly affected the functional brain network, especially in the low-frequency band below 12 Hz. After two modes of sonication, the power spectral density of the EEG signals and the connection strength of the brain network were significantly reduced, but there was no significant difference between the two modes. Our results indicated that the ultrasound neuromodulation could effectively regulate the epileptic brain connections. The ultrasound-mediated attenuation of epilepsy was independent of modes of ultrasound.

Brain function and network features in patients with chronic epilepsy before and after antiseizure medication withdrawal

ObjectivesA considerable proportion of epilepsy patients who achieved long-term seizure freedom with standardized treatment of antiseizure medication will attempt to withdraw medications. Epilepsy is currently considered as a network disease, however, the characteristics of brain function and neural network before and after medication withdrawal remain to be discovered. MethodsResting-state functional magnetic resonance imaging was obtained for 32 healthy controls, 32 seizure-free patients initiating medication tapering (PG1 group), and 16 seizure-free patients that had completely discontinued medications (PG2 group). Amplitude of low-frequency fluctuation and regional homogeneity were calculated to measure local functional activity. Global and nodal metrics of small-world network were calculated based on Graph theory. One-way analysis of variance was applied to analyze intergroup difference, withpost hoc analysis being conducted for each pair of groups. ResultsSex, age at scanning and other clinical variables showed no significant difference between groups. As compared to control, the amplitude of low-frequency fluctuation, regional homogeneity or nodal metrics of neural network in some brain areas were abnormal in the PG1 or PG2 group; when compared between patient groups, significant between-group differences were also found in the amplitude of low-frequency fluctuation, regional homogeneity or nodal metrics. But, the global metrics of neural network showed no differences among groups. ConclusionsThe global metrics of patients with long-term seizure freedom were normal either before or after antiseizure medication withdrawal, while the local functional activity and nodal metrics in some brain areas were abnormal and differed between before and after antiseizure medication withdrawal.

Predicting MCI to AD Conversation Using Integrated sMRI and rs-fMRI: Machine Learning and Graph Theory Approach.

BackgroundGraph theory and machine learning have been shown to be effective ways of classifying different stages of Alzheimer’s disease (AD). Most previous studies have only focused on inter-subject classification with single-mode neuroimaging data. However, whether this classification can truly reflect the changes in the structure and function of the brain region in disease progression remains unverified. In the current study, we aimed to evaluate the classification framework, which combines structural Magnetic Resonance Imaging (sMRI) and resting-state functional Magnetic Resonance Imaging (rs-fMRI) metrics, to distinguish mild cognitive impairment non-converters (MCInc)/AD from MCI converters (MCIc) by using graph theory and machine learning.MethodsWith the intra-subject (MCInc vs. MCIc) and inter-subject (MCIc vs. AD) design, we employed cortical thickness features, structural brain network features, and sub-frequency (full-band, slow-4, slow-5) functional brain network features for classification. Three feature selection methods [random subset feature selection algorithm (RSFS), minimal redundancy maximal relevance (mRMR), and sparse linear regression feature selection algorithm based on stationary selection (SS-LR)] were used respectively to select discriminative features in the iterative combinations of MRI and network measures. Then support vector machine (SVM) classifier with nested cross-validation was employed for classification. We also compared the performance of multiple classifiers (Random Forest, K-nearest neighbor, Adaboost, SVM) and verified the reliability of our results by upsampling.ResultsWe found that in the classifications of MCIc vs. MCInc, and MCIc vs. AD, the proposed RSFS algorithm achieved the best accuracies (84.71, 89.80%) than the other algorithms. And the high-sensitivity brain regions found with the two classification groups were inconsistent. Specifically, in MCIc vs. MCInc, the high-sensitivity brain regions associated with both structural and functional features included frontal, temporal, caudate, entorhinal, parahippocampal, and calcarine fissure and surrounding cortex. While in MCIc vs. AD, the high-sensitivity brain regions associated only with functional features included frontal, temporal, thalamus, olfactory, and angular.ConclusionsThese results suggest that our proposed method could effectively predict the conversion of MCI to AD, and the inconsistency of specific brain regions provides a novel insight for clinical AD diagnosis.

Functional Brain Connectivity Hyper-Network Embedded with Structural Information for Epilepsy Diagnosis

Automatic diagnosis of brain diseases based on brain connectivity network (BCN) classification is one of the hot research fields in medical image analysis. The functional brain network reflects the brain functional activities and structural brain network reflects the neural connections of the main brain regions. It is of great significance to explore and explain the inner mechanism of the brain and to understand and treat brain diseases. In this paper, based on the graph structure characteristics of brain network, the fusion model of functional brain network and structural brain network is designed to classify the diagnosis of brain mental diseases. Specifically, the main work of this paper is to use the Laplacian graph embed the information of diffusion tensor imaging, which contains the characteristics of structural brain networks, into the functional brain network with hyper-order functional connectivity information built based on functional magnetic resonance data using the sparse representation method, to obtain brain network with both functional and structural characteristics. Projection of the brain network and the two original modes data to the kernel space respectively and then classified by the multi-task learning method. Experiments on the epilepsy dataset show that our method has better performance than several state-of-the-art methods. In addition, brain regions and connections that are highly correlated with disease revealed by our method are discussed.