Graph Metrics Research Articles

Classification of multiple sclerosis clinical profiles using machine learning and grey matter connectome.

Purpose: The main goal of this study is to investigate the discrimination power of Grey Matter (GM) thickness connectome data between Multiple Sclerosis (MS) clinical profiles using statistical and Machine Learning (ML) methods. Materials and Methods: A dataset composed of 90 MS patients acquired at the MS clinic of Lyon Neurological Hospital was used for the analysis. Four MS profiles were considered, corresponding to Clinical Isolated Syndrome (CIS), Relapsing-Remitting MS (RRMS), Secondary Progressive MS (SPMS), and Primary Progressive MS (PPMS). Each patient was classified in one of these profiles by our neurologist and underwent longitudinal MRI examinations including T1-weighted image acquisition at each examination, from which the GM tissue was segmented and the cortical GM thickness measured. Following the GM parcellation using two different atlases (FSAverage and Glasser 2016), the morphological connectome was built and six global metrics (Betweenness Centrality (BC), Assortativity (r), Transitivity (T), Efficiency (E g ), Modularity (Q) and Density (D)) were extracted. Based on their connectivity metrics, MS profiles were first statistically compared and second, classified using four different learning machines (Logistic Regression, Random Forest, Support Vector Machine and AdaBoost), combined in a higher level ensemble model by majority voting. Finally, the impact of the GM spatial resolution on the MS clinical profiles classification was analyzed. Results: Using binary comparisons between the four MS clinical profiles, statistical differences and classification performances higher than 0.7 were observed. Good performances were obtained when comparing the two early clinical forms, RRMS and PPMS (F1 score of 0.86), and the two neurodegenerative profiles, PPMS and SPMS (F1 score of 0.72). When comparing the two atlases, slightly better performances were obtained with the Glasser 2016 atlas, especially between RRMS with PPMS (F1 score of 0.83), compared to the FSAverage atlas (F1 score of 0.69). Also, the thresholding value for graph binarization was investigated suggesting more informative graph properties in the percentile range between 0.6 and 0.8. Conclusion: An automated pipeline was proposed for the classification of MS clinical profiles using six global graph metrics extracted from the GM morphological connectome of MS patients. This work demonstrated that GM morphological connectivity data could provide good classification performances by combining four simple ML models, without the cost of long and complex MR techniques, such as MR diffusion, and/or deep learning architectures.

Functional Connectivity in Patients With Parkinson’s Disease and Freezing of Gait Using Resting-State EEG and Graph Theory

Background Although many studies have shown abnormalities in brain structure and function in people with Parkinson’s disease (PD), we still have a poor understanding of how brain structure and function relates to freezing of gait (FOG). Graph theory analysis of electroencephalography (EEG) can explore the relationship between brain network structure and gait function in PD. Methods Scalp EEG signals of 83 PD (42 PDFOG+ and 41 PDFOG−) and 42 healthy controls were recorded in an eyes-opened resting-state. The phase lag index was calculated for each electrode pair in different frequency bands, but we focused our analysis on the theta-band and performed global analyses along with nodal analyses over a midfrontal channel. The resulting connectivity matrices were converted to weighted graphs, whose structure was characterized using strength and clustering coefficient measurements, our main outcomes. Results We observed increased global strength and increased global clustering coefficient in people with PD compared to healthy controls in the theta-band, though no differences were observed in midfrontal nodal strength and midfrontal clustering coefficient. Furthermore, no differences in global nor midfrontal nodal strength nor global clustering coefficients were observed between PDFOG+ and PDFOG− in the theta-band. However, PDFOG+ exhibited a significantly diminished midfrontal nodal clustering coefficient in the theta-band compared to PDFOG−. Furthermore, FOG scores were negatively correlated with midfrontal nodal clustering coefficient in the theta-band. Conclusion The present findings support the involvement of midfrontal theta oscillations in FOG symptoms in PD and the sensitivity of graph metrics to characterize functional networks in PDFOG+.

Group’n Route: An Edge Learning-Based Clustering and Efficient Routing Scheme Leveraging Social Strength for the Internet of Vehicles

The Internet of Vehicles (IoV) is undoubtedly at the core of the future of intelligent transportation. It will prevail over the road ecosystem, and it will have a huge impact on our lives throughout the provision of seamless connectivity among diverse transportation means. For the network to operate efficiently, the data needs to be quickly spread throughout the network, which requires low computational and bandwidth overheads. However, the dynamics of vehicular environments due to frequent node mobility poses many challenges to realize efficient data dissemination. This work addresses this type of problem by proposing a novel clustering algorithm at the edge of the network and an efficient message routing approach, which is known as Group’n Route (GnR). Both mechanisms resort to machine learning and graph metrics that reflect the social relationships between the nodes. Our performance evaluation reveals that the clustering algorithm yields stable results with varying road scenarios, which are becoming an advisable approach in the presence of mobile IoV nodes. Also, the designed routing protocol achieves two orders of magnitude smaller overhead and almost double the delivery rate when it is compared to traditional routing protocols, which thereby justify that the combination of our two proposed clustering and routing methods are a plausible alternative to support IoV communications in real-world setups.

Brain mechanisms underlying the influence of emotions on spatial decision-making: An EEG study.

It is common for people to make bad decisions because of their emotions in life. When these decisions are important, such as aeronautical decisions and driving decisions, the mistakes of decisions can cause irreversible damage. Therefore, it is important to explore how emotions influence decision-making, so as to avoid the negative influence of emotions on decision-making as much as possible. Although existing researchers have found some mechanisms of emotion's influence on decision-making, only a few studies focused on the influence of emotions on decision-making based on electroencephalography (EEG). In addition, most of them were focused on risky and uncertain decision-making. We designed a novel experimental task to explore the influence of emotion on spatial decision-making and recorded subjective data, decision-making behavioral data, and EEG data. By analyzing these data, we came to three conclusions. Firstly, we observed three similar event-related potentials (ERP) microstates in the decision-making process under different emotions by microstate analysis. Additionally, the prefrontal, parietal and occipital lobes played key roles in decision-making. Secondly, we found that the P2 component of the prefrontal lobe presented the influence of different emotions on decision-making by ERP analysis. Among them, positive emotion evoked the largest P2 amplitude compared to negative emotions and no stimuli. Thirdly, we found some graph metrics that were significantly associated with decision accuracy by effective connectivity analysis combined with graph theoretic analysis. In consequence, the finding of our study may shed more light on the brain mechanisms underlying the influence of emotions on spatial decision-making, thereby providing a basis for avoiding decision-making accidents caused by emotions and realizing better decision-making.

Early Childhood Development of Node Centrality in the White Matter Connectome and Its Relationship to IQ at Age 6 Years

BackgroundThe white matter (WM) connectome is important for cognitive development and intelligence and is altered in neuropsychiatric illnesses. Little is known about how the WM connectome develops or its relationship to IQ in early childhood. MethodsThe development of node centrality in the WM connectome was studied in a longitudinal cohort of 226 (123 female) children from the University of North Carolina Early Brain Development Study. Structural and diffusion-weighted images were acquired after birth and at 1, 2, 4, and 6 years, and IQ was assessed at 6 years. Eigenvector centrality, betweenness centrality, and the global graph metrics of global efficiency, small worldness, and modularity were determined at each age. ResultsThe greatest developmental change in eigenvector centrality and betweenness centrality occurred during the first year of life, with relative stability between ages 1 and 6 years. Most of the high-centrality hubs at age 6 were also high-centrality hubs at 1 year, and many were already high-centrality hubs at birth. There were generally small but significant changes in global efficiency and modularity from birth to 6 years, while small worldness increased between 2 and 4 years. Individual node centrality was not significantly correlated with IQ at 6 years. ConclusionsNode centrality in the WM connectome is established very early in childhood and is relatively stable from age 1 to 6 years. Many high-centrality hubs are established before birth, and most are present by age 1.

Global Alterations of Whole Brain Structural Connectome in Parkinson's Disease: A Meta-analysis.

Recent graph-theoretical studies of Parkinson's disease (PD) have examined alterations in the global properties of the brain structural connectome; however, reported alterations are not consistent. The present study aimed to identify the most robust global metric alterations in PD via a meta-analysis. A comprehensive literature search was conducted for all available diffusion MRI structural connectome studies that compared global graph metrics between PD patients and healthy controls (HC). Hedges' g effect sizes were calculated for each study and then pooled using a random-effects model in Comprehensive Meta-Analysis software, and the effects of potential moderator variables were tested. A total of 22 studies met the inclusion criteria for review. Of these, 16 studies reporting 10 global graph metrics (916 PD patients; 560 HC) were included in the meta-analysis. In the structural connectome of PD patients compared with HC, we found a significant decrease in clustering coefficient (g = -0.357, P = 0.005) and global efficiency (g = -0.359, P < 0.001), and a significant increase in characteristic path length (g = 0.250, P = 0.006). Dopaminergic medication, sex and age of patients were potential moderators of global brain network changes in PD. These findings provide evidence of decreased global segregation and integration of the structural connectome in PD, indicating a shift from a balanced small-world network to 'weaker small-worldization', which may provide useful markers of the pathophysiological mechanisms underlying PD.

Reorganized Hubs of Brain Functional Networks after Acute Mild Traumatic Brain Injury.

Mild traumatic brain injury (mTBI)-associated damage to hub regions can lead to disrupted modular structures of functional brain networks and may result in widespread cognitive and behavioral deficits. The spatial layout of brain connections and modules is essential for understanding the reorganization of brain networks to trauma. We investigated the roles of hubs in inter-subnetwork information coordination and integration using participation coefficients (PCs) in 74 patients with acute mTBI and 51 matched healthy controls. In some brain networks, such as default mode network (DMN) and frontoparietal network (FPN), mild TBI patients had decreased PC levels, while this measure was saliently increased in patients in other networks, such as the visual network. The hub disruption index was defined as the gradient of a straight line fitted to scatterplots of individual mTBI in participation coefficient versus mean participation coefficient of healthy groups. There was a trend of radical reorganization of some efficient "hub" nodes in patients (κ = -0.15), compared with controls (κ close to 0). The PC of brain hubs can also differentiate mTBI patients from controls with an 88% accuracy, and decreased PC levels in FPN can predict patient' s worse cognitive information processing speed (r = 0.36, p < 0.002) and working memory performance (r = 0.35, p < 0.002). Reduced PC within the DMN was associated with patients' complaints of post-concussion symptoms (r = -0.35, p < 0.002). This evidence suggests a trend of spatial transition of hub profiles in acute mTBI, and graph metrics of PC measures can be used as potential diagnostic biomarkers.

Global Network Analysis of Alzheimer's Disease with Minimum Spanning Trees.

A minimum spanning tree (MST) is a unique efficient network comprising the necessary connections needed to connect all regions in a network while retaining the lowest possible cost of connection weight. This study aimed to utilize functional near-infrared spectroscopy (fNIRS) to analyze brain activity in different regions and then construct MST-based regions to characterize the brain topologies of participants with Alzheimer's disease (AD), mild cognitive impairment (MCI), and normal controls (NC). A 46 channel fNIRS setup was used on all participants, with correlation being calculated for each channel pair. An MST was constructed from the resulting correlation matrix, from which graph theory measures were calculated. The average number of connections within a lobe in the left versus right hemisphere was calculated to identify which lobes displayed and abnormal amount of connectivity. Compared to those in the MCI group, the AD group showed a less integrated network structure, with a higher characteristic path length, but lower leaf fraction, maximum degree, and degree divergence. The AD group also showed a higher number of connections in the frontal lobe within the left hemisphere and a lower number between hemispheric frontal lobes as compared to MCI. These results indicate a deviation in network structure and connectivity within patient groups that is consistent with the theory of dysconnectivity for AD. Additionally, the AD group showed strong correlations between the Hamilton depression rating scale and different graph metrics, suggesting a link between network organization and the recurrence of depression in AD.

Default mode network-associated intrinsic connectivity relates to individual learnability differences in errorless and trial-and-error learning

The intrinsic functional network architecture accounts for task-evoked brain activity changes and variabilities in cognitive performance. Relationships between the intrinsic functional network architecture and task performance or learning ability have been previously reported. However, the relationships between learning benefits and the characteristics of intrinsic functional network architecture for different types of learning methods remain unclear. In this study, we used graph theoretical analysis to examine the relationships between intrinsic functional network connectivity and learning benefits in two well-known learning methods in the field of cognitive rehabilitation—errorless learning (EL learning) and trial-and-error learning (T&E learning). We focused on the default mode network (DMN) as a task-relevant network, which can differentiate between EL and T&E learning and was found to be more important for T&E learning in a previous study. Participants performed a color–name association task with both learning methods. The graph metrics used were within-network connectivity and efficiency for the DMN. Within-DMN connectivity and DMN efficiency showed a significantly weak positive correlation with T&E scores but not with EL scores. These findings show that the intrinsic integration strength within the DMN relates to individuals’ learnability through the T&E method.

Spontaneous brain activity, graph metrics, and head motion related to prospective post-traumatic stress disorder trauma-focused therapy response.

IntroductionTrauma-focused psychotherapy for post-traumatic stress disorder (PTSD) is effective in about half of all patients. Investigating biological systems related to prospective treatment response is important to gain insight in mechanisms predisposing patients for successful intervention. We studied if spontaneous brain activity, brain network characteristics and head motion during the resting state are associated with future treatment success.MethodsFunctional magnetic resonance imaging scans were acquired from 46 veterans with PTSD around the start of treatment. Psychotherapy consisted of trauma-focused cognitive behavioral therapy (tf-CBT), eye movement desensitization and reprocessing (EMDR), or a combination thereof. After intervention, 24 patients were classified as treatment responders and 22 as treatment resistant. Differences between groups in spontaneous brain activity were evaluated using amplitude of low-frequency fluctuations (ALFF), while global and regional brain network characteristics were assessed using a minimum spanning tree (MST) approach. In addition, in-scanner head motion was assessed.ResultsNo differences in spontaneous brain activity and global network characteristics were observed between the responder and non-responder group. The right inferior parietal lobule, right putamen and left superior parietal lobule had a more central position in the network in the responder group compared to the non-responder group, while the right dorsolateral prefrontal cortex (DLPFC), right inferior frontal gyrus and left inferior temporal gyrus had a less central position. In addition, responders showed less head motion.DiscussionThese results show that areas involved in executive functioning, attentional and action processes, learning, and visual-object processing, are related to prospective PTSD treatment response in veterans. In addition, these findings suggest that involuntary micromovements may be related to future treatment success.

Altered pattern analysis and identification of subjective cognitive decline based on morphological brain network.

Subjective cognitive decline (SCD) is considered the first stage of Alzheimer’s disease (AD). Accurate diagnosis and the exploration of the pathological mechanism of SCD are extremely valuable for targeted AD prevention. However, there is little knowledge of the specific altered morphological network patterns in SCD individuals. In this present study, 36 SCD cases and 34 paired-matched normal controls (NCs) were recruited. The Jensen-Shannon distance-based similarity (JSS) method was implemented to construct and derive the attributes of multiple brain connectomes (i.e., morphological brain connections and global and nodal graph metrics) of individual morphological brain networks. A t-test was used to discriminate between the selected nodal graph metrics, while the leave-one-out cross-validation (LOOCV) was used to obtain consensus connections. Comparisons were performed to explore the altered patterns of connectome features. Further, the multiple kernel support vector machine (MK-SVM) was used for combining brain connectomes and differentiating SCD from NCs. We showed that the consensus connections and nodal graph metrics with the most discriminative ability were mostly found in the frontal, limbic, and parietal lobes, corresponding to the default mode network (DMN) and frontoparietal task control (FTC) network. Altered pattern analysis demonstrated that SCD cases had a tendency for modularity and local efficiency enhancement. Additionally, using the MK-SVM to combine the features of multiple brain connectomes was associated with optimal classification performance [area under the curve (AUC): 0.9510, sensitivity: 97.22%, specificity: 85.29%, and accuracy: 91.43%]. Therefore, our study highlighted the combination of multiple connectome attributes based on morphological brain networks and offered a valuable method for distinguishing SCD individuals from NCs. Moreover, the altered patterns of multidimensional connectome attributes provided a promising insight into the neuroimaging mechanism and early intervention in SCD subjects.

An efficient supply management in water flow network using graph spectral techniques.

An efficient way of distributing water in urban cities is one of the major global challenges of our time. A Water Distribution Network (WDN) is a vital infrastructure, intended to provide fresh water to households within a city or designated boundary. Given the WDN's complexity, effective numerical techniques are required to aid in the development of an ideal monitoring system. Flow meters and gate valves should be placed in low-connected areas of the WDN with water flow reaching several regions of the network. This study proposes a general strategy for assisting water utilities in making decisions for effective water supply. The aim of the research is to use weighted spectral clustering algorithms to outline water districts while addressing hydraulic restrictions via weighted adjacency and Laplacian matrices of the weighted network. This work aims to identify influenced nodes in the network based on Eigen centrality and effectively distribute water across those nodes. This project also looks at how to measure the network's connection strength to avoid water leaks. The best clusters are found using Eigenvalues and Eigenvectors of weighted adjacency matrices and Laplacian matrices of the water network in the proposed graph spectral framework. In order to establish the optimum water network division approach, topological and graph metrics were used to compare multiple spectral clustering techniques. The proposed graph spectral approach is tested using a genuine water distribution network serving an urban area of Coimbatore city in India and offering a method for partitioning complex networks that employ the spectral graph partitioning algorithm.

Isolated Rupture in Composite Networks

Computer networks are prone to targeted attacks and random failures. Robustness is a measure of an ability of a network to continue functioning when part of the network is either naturally damaged or targeted for attack. The study of network robustness is a critical tool in the characterization and understanding of complex interconnected systems. There are several proposed graph metrics that predicates network resilience against such attacks. Isolated rupture degree is a novel graph-theoretic concept defined as a measure of network vulnerability. Isolated rupture degree is argued as an appropriate measure for modelling the robustness of network topologies in the face of possible node destruction. In this paper, the relationships between isolated rupture degree and some other graph parameters such as connectivity, covering number, minimum vertex degree are established. The isolated rupture degrees of [Formula: see text]-free graphs, middle graphs, corona graphs of a middle graph and a complete graph [Formula: see text] on two vertices are evaluated, then compared and the more stable graph types are reported. A sharp upper bound for the isolated rupture degree of middle graphs is established.

Brain Functional Connectivity in Low- and High-Grade Gliomas: Differences in Network Dynamics Associated with Tumor Grade and Location.

Simple SummaryThis study investigates brain network modifications related to tumor grade and location using resting-state functional magnetic resonance imaging and graph theory. We demonstrated that low-grade gliomas (LGG) lead to increased efficiency of the surrounding functional network, while high-grade gliomas (HGG) seem to disrupt brain connectivity in remote areas. Tumor location appears to influence the pattern of reorganization, including the recruitment of the contralateral hemisphere. Overall, LGG may show more favorable connectivity changes than HGG. If confirmed by future studies, the ability to discriminate between ‘maladaptive’ (detrimental) and ‘adaptive’ (beneficial) functional reorganization based on graph theory metrics may provide biomarkers to select patients for surgery and monitor recovery. Brain tumors lead to modifications of brain networks. Graph theory plays an important role in clarifying the principles of brain connectivity. Our objective was to investigate network modifications related to tumor grade and location using resting-state functional magnetic resonance imaging (fMRI) and graph theory. We retrospectively studied 30 low-grade (LGG), 30 high-grade (HGG) left-hemispheric glioma patients and 20 healthy controls (HC) with rs-fMRI. Tumor location was labeled as: frontal, temporal, parietal, insular or occipital. We collected patients’ clinical data from records. We analyzed whole-brain and hemispheric networks in all patients and HC. Subsequently, we studied lobar networks in subgroups of patients divided by tumor location. Seven graph-theoretical metrics were calculated (FDR p < 0.05). Connectograms were computed for significant nodes. The two-tailed Student t-test or Mann–Whitney U-test (p < 0.05) were used to compare graph metrics and clinical data. The hemispheric network analysis showed increased ipsilateral connectivity for LGG (global efficiency p = 0.03) and decreased contralateral connectivity for HGG (degree/cost p = 0.028). Frontal and temporal tumors showed bilateral modifications; parietal and insular tumors showed only local effects. Temporal tumors led to a bilateral decrease in all graph metrics. Tumor grade and location influence the pattern of network reorganization. LGG may show more favorable network changes than HGG, reflecting fewer clinical deficits.

Beta-band network modularity in resting-state EEG negatively correlates with level of intelligence

IntroductionRecent studies mostly focus on the links between measures of alpha-band EEG networks and intelligence. However, associations between wide frequency range EEG networks and general intelligence level remain underresearched.ObjectivesIn this study in a student sample we aimed to correlate the intelligence level and graph metrics of the sensors/sources-level networks constructed in different frequency EEG bands.MethodsWe recorded eyes-closed resting-state EEG in 28 healthy participants (21.4±2.1 y.o., 18 females, 1 left-handed). The Raven’s Standard Progressive Matrices Plus (‘SPM Plus’, 60 figures) was used as an intelligence measure. We constructed networks for all possible combinations of sensors/sources-level and 4-8, 8-13, 13-30, or 4-30 Hz frequency bands using Weighted Phase-Lag Index (wPLI), and calculated four graph metrics (Characteristic Path Length, Clustering Coefficient, Modularity, and Small World Index) for each network. Spearman correlation (with Holm-Sidak correction) was applied to characterize the relations between the SPM Plus scores and all the network metrics.ResultsSPM Plus scores varied from 35 to 57 (mean 45.3±4.2), and the intelligence level negatively correlated with Modularity in beta-band (r = -0.63, pcorr = 0.0253).ConclusionsHigh modularity may reflect relatively high segregation, but not integration, of networks (Girn, Mills, Christoff, 2019). Accordingly, our findings may shed light on the neural mechanisms of the general inefficiency of global cognitive processing in the case of intellectual decline related to different mental disorders. Funding: This research has been supported by the Interdisciplinary Scientific and Educational School of Lomonosov Moscow State University ‘Brain, Cognitive Systems, Artificial Intelligence’.DisclosureNo significant relationships.

Resting-state EEG networks characterized by intramodular and global hyperconnectivity in depressive sample

IntroductionDepression is characterized by a pattern of specific changes in the network organization of brain functioning.ObjectivesWe researched a graph structure specificity in a depressive student sample by analyzing resting-state EEG. All possible combinations of graph metrics, frequency bands, and sensors/sources levels of networks were examined.MethodsWe recorded resting-state EEG in fourteen participants with high Beck Depression Inventory score (24.4 ± 9.7; 20.4 ± 1.5 y.o.; 14 females; 1 left-handed) and fourteen participants with a low score (6.8 ± 3.7; 21.3 ± 2.0 y.o.; 8 females; 1 left-handed). We applied weighted phase-lag index (wPLI) to construct functional networks at sensors and sources levels and computed characteristic path length (CPL), clustering coefficient (CC), index of modularity (Q), small-world index (SWI) in 4-8, 8-13, 13-30, and 4-30 Hz frequency bands. We used Mann-Whitney U-test (p &lt; 0.05) to investigate between-group differences in the graph metrics.ResultsThe depressive sample was characterized by increased CC and Q in the 4-30 Hz band networks and decreased CPL in the beta-band network (sensors-level for CPL and CC, and sources-level for Q).ConclusionsElevated CC and Q may relate to an increase of intramodular connectivity, and CPL reduction reflects the global connectivity increasing. We hypothesize that intramodular hyperconnectivity could explain the rise of global functional connectivity in participants with depressive symptoms. Funding: This research has been supported by the Interdisciplinary Scientific and Educational School of Lomonosov Moscow State University ‘Brain, Cognitive Systems, Artificial Intelligence’.DisclosureNo significant relationships.

Network classification-based structural analysis of real networks and their model-generated counterparts

AbstractData-driven analysis of complex networks has been in the focus of research for decades. An important area of research is to study how well real networks can be described with a small selection of metrics, furthermore how well network models can capture the relations between graph metrics observed in real networks. In this paper, we apply machine-learning techniques to investigate the aforementioned problems. We study 500 real-world networks along with 2000 synthetic networks generated by four frequently used network models with previously calibrated parameters to make the generated graphs as similar to the real networks as possible. This paper unifies several branches of data-driven complex network analysis, such as the study of graph metrics and their pair-wise relationships, network similarity estimation, model calibration, and graph classification. We find that the correlation profiles of the structural measures significantly differ across network domains and the domain can be efficiently determined using a small selection of graph metrics. The structural properties of the network models with fixed parameters are robust enough to perform parameter calibration. The goodness-of-fit of the network models highly depends on the network domain. By solving classification problems, we find that the models lack the capability of generating a graph with a high clustering coefficient and relatively large diameter simultaneously. On the other hand, models are able to capture exactly the degree-distribution-related metrics.

Computing traffic accident high-risk locations using graph analytics

Analysis of the dynamic relationship between traffic accident events and road network topology based on connectivity and graph analytics offers a new approach to identifying, ranking and profiling traffic accident high risk-locations at different levels of space and time granularities. Previous studies on traffic accident hot spots have mostly adopted spatial statistics and Geographic Information Systems (GIS) where spatial point patterns are discovered based only on spatial dependence with no recognition of the temporal dependence of the events. A limitation arises from the fact that the results are either under or over-estimated because of the temporal aggregation of the events to an absolute time point. Furthermore, the existing methods apart from the Network Kernel Density Estimation (NETKDE), consider traffic accident events as events randomly on a 2-D geographic space. However, traffic accident events are network constrained events that happens majorly on the road network space. Therefore, in this paper, we adopt the connectivity of graph on a network space approach that identifies accident high risk-locations based on space-time-varying connectivity between traffic accident events and the road network geometry. A simple but extensible traffic accident space time-varying graph (STVG) model is developed and implemented for this study. Traffic accident high risk-locations are identified and ranked in space and time using time-dependent degree centrality and PageRank centrality graph metrics respectively through time-incremental graph queries. This study offers urban traffic accident analysts with a new and efficient approach to identify, rank and profile accident-prone areas in space and time at different scales.

Brain network connectivity feature extraction using deep learning for Alzheimer's disease classification

Early diagnosis and therapeutic intervention for Alzheimer's disease (AD) is currently the only viable option for improving clinical outcomes. Combining structural magnetic resonance imaging (sMRI) and resting-state functional magnetic resonance imaging (rs-fMRI) to diagnose AD has yielded promising results. Most studies assume fixed time lags when constructing functional networks. Since the propagation delays between brain signals are constantly changing, these methods cannot reflect more detailed relationships between brain regions. In this work, we use a deep learning-based Granger causality estimator for brain connectivity construction. It exploits the strength of long short-term memory in ever-changing time series processing. This research involves data analysis from sMRI and rs-fMRI. We use sMRI to analyze the cerebral cortex properties and use rs-fMRI to analyze the graph metrics of functional networks. We extract a small subset of optimal features from both types of data. A support vector machine (SVM) is trained and tested to classify AD (n = 27) from healthy controls (n = 20) using rs-fMRI and sMRI features. Using a subset of optimal features in SVM, we achieve a classification accuracy of 87.23% for sMRI, 78.72% for rs-fMRI, and 91.49% for combined sMRI with rs-fMRI. The results show the potential to identify AD from healthy controls by integrating rs-fMRI and sMRI. The integration of sMRI and rs-fMRI modalities can provide supplemental information to improve the diagnosis of AD relative to either the sMRI or fMRI modalities alone.

Hybrid machine learning method for a connectivity-based epilepsy diagnosis with resting-state EEG.

This study investigates the performance of a convolutional neural network (CNN) algorithm on epilepsy diagnosis. Without pathology, diagnosis involves long and costly electroencephalographic (EEG) monitoring. Novel approaches may overcome this by comparing brain connectivity using graph metrics. This study, however, uses deep learning to learn connectivity patterns directly from easily acquired EEG data. A CNN algorithm was applied on directed Granger causality (GC) connectivity measures, derived from 50s of resting-state surface EEG recordings from 30 subjects with epilepsy and a 30 subject control group. The trained CNN filters reflected reduced delta band connectivity in frontal regions and increased left lateralized frontal-posterior gamma band connectivity. A diagnosis accuracy of 85% (F1 score 85%) was achieved by an ensemble of CNN models, each trained on differently prepared data from different electrode combinations. Appropriate preparation of connectivity data enables generic CNN algorithms to be used for detection of multiple discriminative epileptic features. Differential patterns revealed in this study may help to shed light on underlying altered cognitive abilities in epilepsy patients. The accuracy achieved in this study shows that, in combination with other methods, this approach could prove a valuable clinical decision support system for epilepsy diagnosis. 1: EEG measurements and subsequent connectivity calculation, 2: training of a neural network on resulting connectivity matrices, 3: extraction of most efficient CNN filters, which are neuromarker for epilepsy.