Complex networks and deep learning for EEG signal analysis

Electroencephalogram (EEG) signals acquired from brain can provide an effective representation of the human’s physiological and pathological states. Up to now, much work has been conducted to study and analyze the EEG signals, aiming at spying the current states or the evolution characteristics of the complex brain system. Considering the complex interactions between different structural and functional brain regions, brain network has received a lot of attention and has made great progress in brain mechanism research. In addition, characterized by autonomous, multi-layer and diversified feature extraction, deep learning has provided an effective and feasible solution for solving complex classification problems in many fields, including brain state research. Both of them show strong ability in EEG signal analysis, but the combination of these two theories to solve the difficult classification problems based on EEG signals is still in its infancy. We here review the application of these two theories in EEG signal research, mainly involving brain–computer interface, neurological disorders and cognitive analysis. Furthermore, we also develop a framework combining recurrence plots and convolutional neural network to achieve fatigue driving recognition. The results demonstrate that complex networks and deep learning can effectively implement functional complementarity for better feature extraction and classification, especially in EEG signal analysis.     

Directed Network Motifs in Alzheimer’s Disease and Mild Cognitive Impairment

Directed network motifs are the building blocks of complex networks, such as human brain networks, and capture deep connectivity information that is not contained in standard network measures. In this paper we present the first application of directed network motifs in vivo to human brain networks, utilizing recently developed directed progression networks which are built upon rates of cortical thickness changes between brain regions. This is in contrast to previous studies which have relied on simulations and in vitro analysis of non-human brains. We show that frequencies of specific directed network motifs can be used to distinguish between patients with Alzheimer’s disease (AD) and normal control (NC) subjects. Especially interesting from a clinical standpoint, these motif frequencies can also distinguish between subjects with mild cognitive impairment who remained stable over three years (MCI) and those who converted to AD (CONV). Furthermore, we find that the entropy of the distribution of directed network motifs increased from MCI to CONV to AD, implying that the distribution of pathology is more structured in MCI but becomes less so as it progresses to CONV and further to AD. Thus, directed network motifs frequencies and distributional properties provide new insights into the progression of Alzheimer’s disease as well as new imaging markers for distinguishing between normal controls, stable mild cognitive impairment, MCI converters and Alzheimer’s disease.     

Graph theoretical analysis of complex networks in the brain

Since the discovery of small-world and scale-free networks the study of complex systems from a network perspective has taken an enormous flight. In recent years many important properties of complex networks have been delineated. In particular, significant progress has been made in understanding the relationship between the structural properties of networks and the nature of dynamics taking place on these networks. For instance, the 'synchronizability' of complex networks of coupled oscillators can be determined by graph spectral analysis. These developments in the theory of complex networks have inspired new applications in the field of neuroscience. Graph analysis has been used in the study of models of neural networks, anatomical connectivity, and functional connectivity based upon fMRI, EEG and MEG. These studies suggest that the human brain can be modelled as a complex network, and may have a small-world structure both at the level of anatomical as well as functional connectivity. This small-world structure is hypothesized to reflect an optimal situation associated with rapid synchronization and information transfer, minimal wiring costs, as well as a balance between local processing and global integration. The topological structure of functional networks is probably restrained by genetic and anatomical factors, but can be modified during tasks. There is also increasing evidence that various types of brain disease such as Alzheimer's disease, schizophrenia, brain tumours and epilepsy may be associated with deviations of the functional network topology from the optimal small-world pattern.     

Using Pareto optimality to explore the topology and dynamics of the human connectome

Graph theory has provided a key mathematical framework to analyse the architecture of human brain networks. This architecture embodies an inherently complex relationship between connection topology, the spatial arrangement of network elements, and the resulting network cost and functional performance. An exploration of these interacting factors and driving forces may reveal salient network features that are critically important for shaping and constraining the brain''s topological organization and its evolvability. Several studies have pointed to an economic balance between network cost and network efficiency with networks organized in an ‘economical’ small-world favouring high communication efficiency at a low wiring cost. In this study, we define and explore a network morphospace in order to characterize different aspects of communication efficiency in human brain networks. Using a multi-objective evolutionary approach that approximates a Pareto-optimal set within the morphospace, we investigate the capacity of anatomical brain networks to evolve towards topologies that exhibit optimal information processing features while preserving network cost. This approach allows us to investigate network topologies that emerge under specific selection pressures, thus providing some insight into the selectional forces that may have shaped the network architecture of existing human brains.     

Aberrant Functional Connectome in Neurologically Asymptomatic Patients with End-Stage Renal Disease

Purpose

This study aimed to investigate the topological organization of intrinsic functional brain networks in patients with end-stage renal disease (ESRD).

Materials and Methods

Resting-state functional MRI data were collected from 22 patients with ESRD (16 men, 18–61 years) and 29age- and gender-matched healthy controls (HCs, 19 men, 32–61 years). Whole-brain functional networks were obtained by calculating the interregional correlation of low-frequency fluctuations in spontaneous brain activity among 1,024 parcels that cover the entire cerebrum. Weighted graph-based models were then employed to topologically characterize these networks at different global, modular and nodal levels.

Results

Compared to HCs, the patients exhibited significant disruption in parallel information processing over the whole networks (P< 0.05). The disruption was present in all the functional modules (default mode, executive control, sensorimotor and visual networks) although decreased functional connectivity was observed only within the default mode network. Regional analysis showed that the disease disproportionately weakened nodal efficiency of the default mode components and tended to preferentially affect central or hub-like regions. Intriguingly, the network abnormalities correlated with biochemical hemoglobin and serum calcium levels in the patients. Finally, the functional changes were substantively unchanged after correcting for gray matter atrophy in the patients.

Conclusion

Our findings provide evidence for the disconnection nature of ESRD’s brain and therefore have important implications for understanding the neuropathologic substrate of the disease from disrupted network organization perspective.     

Large Scale Brain Functional Networks Support Sentence Comprehension: Evidence from Both Explicit and Implicit Language Tasks

Previous studies have indicated that sentences are comprehended via widespread brain regions in the fronto-temporo-parietal network in explicit language tasks (e.g., semantic congruency judgment tasks), and through restricted temporal or frontal regions in implicit language tasks (e.g., font size judgment tasks). This discrepancy has raised questions regarding a common network for sentence comprehension that acts regardless of task effect and whether different tasks modulate network properties. To this end, we constructed brain functional networks based on 27 subjects’ fMRI data that was collected while performing explicit and implicit language tasks. We found that network properties and network hubs corresponding to the implicit language task were similar to those associated with the explicit language task. We also found common hubs in occipital, temporal and frontal regions in both tasks. Compared with the implicit language task, the explicit language task resulted in greater global efficiency and increased integrated betweenness centrality of the left inferior frontal gyrus, which is a key region related to sentence comprehension. These results suggest that brain functional networks support both explicit and implicit sentence comprehension; in addition, these two types of language tasks may modulate the properties of brain functional networks.     

Resting-State Temporal Synchronization Networks Emerge from Connectivity Topology and Heterogeneity

Spatial patterns of coherent activity across different brain areas have been identified during the resting-state fluctuations of the brain. However, recent studies indicate that resting-state activity is not stationary, but shows complex temporal dynamics. We were interested in the spatiotemporal dynamics of the phase interactions among resting-state fMRI BOLD signals from human subjects. We found that the global phase synchrony of the BOLD signals evolves on a characteristic ultra-slow (<0.01Hz) time scale, and that its temporal variations reflect the transient formation and dissolution of multiple communities of synchronized brain regions. Synchronized communities reoccurred intermittently in time and across scanning sessions. We found that the synchronization communities relate to previously defined functional networks known to be engaged in sensory-motor or cognitive function, called resting-state networks (RSNs), including the default mode network, the somato-motor network, the visual network, the auditory network, the cognitive control networks, the self-referential network, and combinations of these and other RSNs. We studied the mechanism originating the observed spatiotemporal synchronization dynamics by using a network model of phase oscillators connected through the brain’s anatomical connectivity estimated using diffusion imaging human data. The model consistently approximates the temporal and spatial synchronization patterns of the empirical data, and reveals that multiple clusters that transiently synchronize and desynchronize emerge from the complex topology of anatomical connections, provided that oscillators are heterogeneous.     

Functional Neural Correlates of Attentional Deficits in Amnestic Mild Cognitive Impairment

Although amnestic mild cognitive impairment (aMCI; often considered a prodromal phase of Alzheimer’s disease, AD) is most recognized by its implications for decline in memory function, research suggests that deficits in attention are present early in aMCI and may be predictive of progression to AD. The present study used functional magnetic resonance imaging to examine differences in the brain during the attention network test between 8 individuals with aMCI and 8 neurologically healthy, demographically matched controls. While there were no significant behavioral differences between groups for the alerting and orienting functions, patients with aMCI showed more activity in neural regions typically associated with the networks subserving these functions (e.g., temporoparietal junction and posterior parietal regions, respectively). More importantly, there were both behavioral (i.e., greater conflict effect) and corresponding neural deficits in executive control (e.g., less activation in the prefrontal and anterior cingulate cortices). Although based on a small number of patients, our findings suggest that deficits of attention, especially the executive control of attention, may significantly contribute to the behavioral and cognitive deficits of aMCI.     

Mapping Molecular Association Networks of Nervous System Diseases via Large-Scale Analysis of Published Research

Network medicine has been applied successfully to elicit the structure of large-scale molecular interaction networks. Its main proponents have claimed that this approach to integrative medical investigation should make it possible to identify functional modules of interacting molecular biological units as well as interactions themselves. This paper takes a significant step in this direction. Based on a large-scale analysis of the nervous system molecular medicine literature, this study analyzes and visualizes the complex structure of associations between diseases on the one hand and all types of molecular substances on the other. From this analysis it then identifies functional co-association groups consisting of several types of molecular substances, each consisting of substances that exhibit a pattern of frequent co-association with similar diseases. These groups in turn exhibit interlinking in a complex pattern, suggesting that such complex interactions between functional molecular modules may play a role in disease etiology. We find that the patterns exhibited by the networks of disease – molecular substance associations studied here correspond well to a number of recently published research results, and that the groups of molecular substances identified by statistical analysis of these networks do appear to be interesting groups of molecular substances that are interconnected in identifiable and interpretable ways. Our results not only demonstrate that networks are a convenient framework to analyze and visualize large-scale, complex relationships among molecular networks and diseases, but may also provide a conceptual basis for bridging gaps in experimental and theoretical knowledge.     

Community Size Effects on Epidemic Spreading in Multiplex Social Networks

The dynamical process of epidemic spreading has drawn much attention of the complex network community. In the network paradigm, diseases spread from one person to another through the social ties amongst the population. There are a variety of factors that govern the processes of disease spreading on the networks. A common but not negligible factor is people’s reaction to the outbreak of epidemics. Such reaction can be related information dissemination or self-protection. In this work, we explore the interactions between disease spreading and population response in terms of information diffusion and individuals’ alertness. We model the system by mapping multiplex networks into two-layer networks and incorporating individuals’ risk awareness, on the assumption that their response to the disease spreading depends on the size of the community they belong to. By comparing the final incidence of diseases in multiplex networks, we find that there is considerable mitigation of diseases spreading for full phase of spreading speed when individuals’ protection responses are introduced. Interestingly, the degree of community overlap between the two layers is found to be critical factor that affects the final incidence. We also analyze the consequences of the epidemic incidence in communities with different sizes and the impacts of community overlap between two layers. Specifically, as the diseases information makes individuals alert and take measures to prevent the diseases, the effective protection is more striking in small community. These phenomena can be explained by the multiplexity of the networked system and the competition between two spreading processes.     

Group Independent Component Analysis and Functional MRI Examination of Changes in Language Areas Associated with Brain Tumors at Different Locations

Object

This study investigates the effect of tumor location on alterations of language network by brain tumors at different locations using blood oxygenation level dependent (BOLD) fMRI and group independent component analysis (ICA).

Subjects and Methods

BOLD fMRI data were obtained from 43 right handed brain tumor patients. Presurgical mapping of language areas was performed on all 43 patients with a picture naming task. All data were retrospectively analyzed using group ICA. Patents were divided into three groups based on tumor locations, i.e., left frontal region, left temporal region or right hemisphere. Laterality index (LI) was used to assess language lateralization in each group.

Results

The results from BOLD fMRI and ICA revealed the different language activation patterns in patients with brain tumors located in different brain regions. Language areas, such as Broca’s and Wernicke’s areas, were intact in patients with tumors in the right hemisphere. Significant functional changes were observed in patients with tumor in the left frontal and temporal areas. More specifically, the tumors in the left frontal region affect both Broca’s and Wernicke’s areas, while tumors in the left temporal lobe affect mainly Wernicke’s area. The compensated activation increase was observed in the right frontal areas in patients with left hemisphere tumors.

Conclusion

Group ICA provides a model free alternative approach for mapping functional networks in brain tumor patients. Altered language activation by different tumor locations suggested reorganization of language functions in brain tumor patients and may help better understanding of the language plasticity.     

Altered Spontaneous Brain Activity in Patients with Parkinson’s Disease Accompanied by Depressive Symptoms,as Revealed by Regional Homogeneity and Functional Connectivity in the Prefrontal-Limbic System

As patients with Parkinson’s disease (PD) are at high risk for comorbid depression, it is hypothesized that these two diseases are sharing common pathogenic pathways. Using regional homogeneity (ReHo) and functional connectivity approaches, we characterized human regional brain activity at resting state to examine specific brain networks in patients with PD and those with PD and depression (PDD). This study comprised 41 PD human patients and 25 normal human subjects. The patients completed the Hamilton Depression Rating Scale and were further divided into two groups: patients with depressive symptoms and non-depressed PD patients (nD-PD). Compared with the non-depressed patients, those with depressive symptoms exhibited significantly increased regional activity in the left middle frontal gyrus and right inferior frontal gyrus, and decreased ReHo in the left amygdala and bilateral lingual gyrus. Brain network connectivity analysis revealed decreased functional connectivity within the prefrontal-limbic system and increased functional connectivity in the prefrontal cortex and lingual gyrus in PDD compared with the nD-PD group. In summary, the findings showed regional brain activity alterations and disruption of the mood regulation network in PDD patients. The pathogenesis of PDD may be attributed to abnormal neural activity in multiple brain regions.     

Analysing Local Sparseness in the Macaque Brain Network

Understanding the network structure of long distance pathways in the brain is a necessary step towards developing an insight into the brain’s function, organization and evolution. Dense global subnetworks of these pathways have often been studied, primarily due to their functional implications. Instead we study sparse local subnetworks of the pathways to establish the role of a brain area in enabling shortest path communication between its non-adjacent topological neighbours. We propose a novel metric to measure the topological communication load on a vertex due to its immediate neighbourhood, and show that in terms of distribution of this local communication load, a network of Macaque long distance pathways is substantially different from other real world networks and random graph models. Macaque network contains the entire range of local subnetworks, from star-like networks to clique-like networks, while other networks tend to contain a relatively small range of subnetworks. Further, sparse local subnetworks in the Macaque network are not only found across topographical super-areas, e.g., lobes, but also within a super-area, arguing that there is conservation of even relatively short-distance pathways. To establish the communication role of a vertex we borrow the concept of brokerage from social science, and present the different types of brokerage roles that brain areas play, highlighting that not only the thalamus, but also cingulate gyrus and insula often act as “relays” for areas in the neocortex. These and other analysis of communication load and roles of the sparse subnetworks of the Macaque brain provide new insights into the organisation of its pathways.     

Serum Interleukin (IL)-15 as a Biomarker of Alzheimer's Disease

Interleukin (IL-15), a pro-inflammatory cytokine has been studied as a possible marker of Alzheimer’s disease (AD); however its exact role in neuro-inflammation or the pathogenesis AD is not well understood yet. A Multiple Indicators Multiple Causes (MIMIC) approach was used to examine the relationship between serum IL-15 levels and AD in a well characterized AD cohort, the Texas Alzheimer''s Research and Care Consortium (TARCC). Instead of categorical diagnoses, we used two latent construct d (for dementia) and g’ (for cognitive impairments not contributing to functional impairments) in our analysis. The results showed that the serum IL-15 level has significant effects on cognition, exclusively mediated by latent construct d and g’. Contrasting directions of association lead us to speculate that IL-15’s effects in AD are mediated through functional networks as d scores have been previously found to be specifically related to default mode network (DMN). Our finding warrants the need for further research to determine the changes in structural and functional networks corresponding to serum based biomarkers levels.     

Neurodegenerative diseases-Caps: a capsule network based early screening system for the classification of neurodegenerative diseases

The two most generally diagnosed Neurodegenerative diseases are the Alzheimer and Parkinson diseases. So this paper presents a fully automated early screening system based on the Capsule network for the classification of these two Neurodegenerative diseases. In this study, we hypothesized that the Neurodegenerative diseases-Caps system based on the Capsule network architecture accurately performs the multiclass i.e. three class classification into either the Alzheimer class or Parkinson class or Healthy control and delivers better results in comparison other deep transfer learning models. The real motivation behind choosing the capsule network architecture is its more resilient nature towards the affine transformations as well as rotational & translational invariance, which commonly persists in the medical image datasets. Apart from this, the capsule networks overcomes the pooling layers related deficiencies from which conventional CNNs are mostly affected and unable to delivers accurate results especially in the tasks related to image classification. The various Computer aided systems based on machine learning for the classification of brain tumors and other types of cancers are already available. Whereas for the classification of Neurodegenerative diseases, the amount of research done is very limited and the number of persons suffering from this type of diseases are increasing especially in developing countries like India, China etc. So there is a need to develop an early screening system for the correct multiclass classification into Alzheimer’s, Parkinson’s and Normal or Healthy control cases. The Alzheimer disease and Parkinson progression (ADPP) dataset is used in this research study for the training of the proposed Neurodegenerative diseases-Caps system. This ADPP dataset is developed with the aid of both the Parkinson''s Progression Markers Initiative (PPMI) and Alzheimer’s disease Neuroimaging Initiative (ADNI) databases. There is no such early screening system exist yet, which can perform the accurate classification of these two Neurodegenerative diseases. For the sake of genuine comparison, other popular deep transfer learning models like VGG19, VGG16, ResNet50 and InceptionV3 are implemented and also trained over the same ADPP dataset. The proposed Neurodegenerative diseases-Caps system deliver accuracies of 97.81, 98, 96.81% for the Alzheimer, Parkinson and Healthy control or Normal cases with 70/30 (training/validation split) and performs way better as compare to the other popular Deep transfer learning models.Supplementary InformationThe online version contains supplementary material available at 10.1007/s11571-022-09787-1.     

Complex Disease Interventions from a Network Model for Type 2 Diabetes

There is accumulating evidence that the proteins encoded by the genes associated with a common disorder interact with each other, participate in similar pathways and share GO terms. It has been anticipated that the functional modules in a disease related functional linkage network are informative to reveal significant metabolic processes and disease’s associations with other complex disorders. In the current study, Type 2 diabetes associated functional linkage network (T2DFN) containing 2770 proteins and 15041 linkages was constructed. The functional modules in this network were scored and evaluated in terms of shared pathways, co-localization, co-expression and associations with similar diseases. The assembly of top scoring overlapping members in the functional modules revealed that, along with the well known biological pathways, circadian rhythm, diverse actions of nuclear receptors in steroid and retinoic acid metabolisms have significant occurrence in the pathophysiology of the disease. The disease’s association with other metabolic and neuromuscular disorders was established through shared proteins. Nuclear receptor NRIP1 has a pivotal role in lipid and carbohydrate metabolism, indicating the need to investigate subsequent effects of NRIP1 on Type 2 diabetes. Our study also revealed that CREB binding protein (CREBBP) and cardiotrophin-1 (CTF1) have suggestive roles in linking Type 2 diabetes and neuromuscular diseases.     

Network Effects on Scientific Collaborations

Background

The analysis of co-authorship network aims at exploring the impact of network structure on the outcome of scientific collaborations and research publications. However, little is known about what network properties are associated with authors who have increased number of joint publications and are being cited highly.

Methodology/Principal Findings

Measures of social network analysis, for example network centrality and tie strength, have been utilized extensively in current co-authorship literature to explore different behavioural patterns of co-authorship networks. Using three SNA measures (i.e., degree centrality, closeness centrality and betweenness centrality), we explore scientific collaboration networks to understand factors influencing performance (i.e., citation count) and formation (tie strength between authors) of such networks. A citation count is the number of times an article is cited by other articles. We use co-authorship dataset of the research field of ‘steel structure’ for the year 2005 to 2009. To measure the strength of scientific collaboration between two authors, we consider the number of articles co-authored by them. In this study, we examine how citation count of a scientific publication is influenced by different centrality measures of its co-author(s) in a co-authorship network. We further analyze the impact of the network positions of authors on the strength of their scientific collaborations. We use both correlation and regression methods for data analysis leading to statistical validation. We identify that citation count of a research article is positively correlated with the degree centrality and betweenness centrality values of its co-author(s). Also, we reveal that degree centrality and betweenness centrality values of authors in a co-authorship network are positively correlated with the strength of their scientific collaborations.

Conclusions/Significance

Authors’ network positions in co-authorship networks influence the performance (i.e., citation count) and formation (i.e., tie strength) of scientific collaborations.