EEG Functional Connectivity Research Articles

Effects of acute psychosocial stress on source level EEG power and functional connectivity measures

The usage of EEG to uncover the influence of psychosocial stressors (PSSs) on neural activity has gained significant attention throughout recent years, but the results are often troubled by confounding stressor types. To investigate the effect of PSSs alone on neural activity, we employed a paradigm where participants are exposed to negative peer comparison as PSS, while other possible stressors are kept constant, and compared this with a condition where participants received neutral feedback. We analyzed commonly used sensor level EEG indices (frontal theta, alpha, and beta power) and further investigated whether source level power and functional connectivity (i.e., the temporal dependence between spatially seperated brain regions) measures, which have to our knowledge not yet been used, are more sensitive to PSSs than sensor level-derived EEG measures. Our results show that on sensor level, no significant frontal power changes are present (all p’s > 0.16), indicating that sensor level frontal power measures are not sensitive enough to be affected by only PSSs. On source level, we find increased alpha power (indicative of decreased cortical activity) in the left- and right precuneus and right posterior cingulate cortex (all p’s < 0.03) and increased functional connectivity between the left- and right precuneus (p < 0.001), indicating that acute, trial based PSSs lead to decreased precuneus/PCC activity, and possibly indicates a temporary disruption in the self-referential neural processes of an individual.

Reproducibility analysis of functional connectivity measures for application in motor imagery BCIs

Electroencephalography (EEG)-based motor imagery brain-computer interfaces (MI-BCIs) can improve motor rehabilitation processes. Nevertheless, the large variability of intra and inter-subject EEG signals has precluded translation of this technology to the clinical setting. The aim of this work was to evaluate the reproducibility of EEG functional connectivity (FC) features used to discriminate between left- and right-hand MI. Ten subjects underwent 12 EEG-MI-BCI sessions. Two frequency bands, three FC methods, four graph parameters and six electrode sites were evaluated, using statistical and classification analyses. The β band produced the largest number of statistically significantly discriminating and most stable features for the majority of subjects, and also the best classification features, suggesting that engagement of needed brain regions may be more important for stability and distinguishability among MI tasks than inhibition of unneeded cortical regions. The Cz electrode stood out in terms of largest number of statistically significantly discriminating features. The motif synchronization (MS) method produced the largest number of significantly discriminating features, the most stable, most discriminating features and best classification features for most subjects. Since this method ignores amplitude changes, this seems to indicate that signal variation patterns are more important for feature stability and class separability. Eigenvector centrality (EC) was the most stable graph parameter for most subjects while both EC and strength (S) were the most discriminating. In summary, using features from the MS method, β band, EC and/or S parameters, and central electrodes (Cz), result in the best combination for the task of distinguishing left from right-hand MI.

Pre-ictal fluctuation of EEG functional connectivity discriminates seizure phenotypes in mesial temporal lobe epilepsy

ObjectiveWe explored whether quantifiable differences between clinical seizures (CSs) and subclinical seizures (SCSs) occur in the pre-ictal state. MethodsWe analyzed pre-ictal stereo-electroencephalography (SEEG) retrospectively across mesial temporal lobe epilepsy patients with recorded CSs and SCSs. Power spectral density and functional connectivity (FC) were quantified within and between the seizure onset zone (SOZ) and the early propagation zone (PZ), respectively. To evaluate the fluctuation of neural connectivity, FC variability was computed. Measures were further verified by a logistic regression model to evaluate their classification potentiality through the area under the receiver-operating-characteristics curve (AUC). ResultsFifty-four pre-ictal SEEG epochs (27 CSs and 27 SCSs) were selected among 14 patients. Within the SOZ, pre-ictal FC variability of CSs was larger than SCSs in 1–45 Hz during 30 seconds before seizure onset. Pre-ictal FC variability between the SOZ and PZ was larger in SCSs than CSs in 55–80 Hz within 1 minute before onset. Using these two variables, the logistic regression model achieved an AUC of 0.79 when classifying CSs and SCSs. ConclusionsPre-ictal FC variability within/between epileptic zones, not signal power or FC value, distinguished SCSs from CSs. SignificancePre-ictal epileptic network stability possibly marks seizure phenotypes, contributing insights into ictogenesis and potentially helping seizure prediction.

N°297 – Vagus Nerve Stimulation modulates EEG functional connectivity in people with drug-resistant epilepsy

N°297 – Vagus Nerve Stimulation modulates EEG functional connectivity in people with drug-resistant epilepsy

Testing EEG functional connectivity between sensorimotor and face processing visual regions in individuals with congenital facial palsy.

Moebius syndrome (MBS) is characterized by the congenital absence or underdevelopment of cranial nerves VII and VI, leading to facial palsy and impaired lateral eye movements. As a result, MBS individuals cannot produce facial expressions and did not develop motor programs for facial expressions. In the latest model of sensorimotor simulation, an iterative communication between somatosensory, motor/premotor cortices, and visual regions has been proposed, which should allow more efficient discriminations among subtle facial expressions. Accordingly, individuals with congenital facial motor disability, specifically with MBS, should exhibit atypical communication within this network. Here, we aimed to test this facet of the sensorimotor simulation models. We estimated the functional connectivity between the visual cortices for face processing and the sensorimotor cortices in healthy and MBS individuals. To this aim, we studied the strength of beta band functional connectivity between these two systems using high-density EEG, combined with a change detection task with facial expressions (and a control condition involving non-face stimuli). The results supported our hypothesis such that when discriminating subtle facial expressions, participants affected by congenital facial palsy (compared to healthy controls) showed reduced connectivity strength between sensorimotor regions and visual regions for face processing. This effect was absent for the condition with non-face stimuli. These findings support sensorimotor simulation models and the communication between sensorimotor and visual areas during subtle facial expression processing.

Correlation between electroencephalographic markers in the healthy brain

Mental disorders have an increasing tendency and represent the main burden of disease to society today. A wide variety of electroencephalographic (EEG) markers have been successfully used to assess different symptoms of mental disorders. Different EEG markers have demonstrated similar classification accuracy, raising a question of their independence. The current study is aimed to investigate the hypotheses that different EEG markers reveal partly the same EEG features reflecting brain functioning and therefore provide overlapping information. The assessment of the correlations between EEG signal frequency band power, dynamics, and functional connectivity markers demonstrates that a statistically significant correlation is evident in 37 of 66 (56%) comparisons performed between 12 markers of different natures. A significant correlation between the majority of the markers supports the similarity of information in the markers. The results of the performed study confirm the hypotheses that different EEG markers reflect partly the same features in brain functioning. Higuchi’s fractal dimension has demonstrated a significant correlation with the 82% of other markers and is suggested to reveal a wide spectrum of various brain disorders. This marker is preferable in the early detection of symptoms of mental disorders.

Resting-State EEG Connectivity at High-Frequency Bands and Attentional Performance Dysfunction in Stabilized Schizophrenia Patients.

Background and Objectives: Attentional dysfunction has long been viewed as one of the fundamental underlying cognitive deficits in schizophrenia. There is an urgent need to understand its neural underpinning and develop effective treatments. In the process of attention, neural oscillation has a central role in filtering information and allocating resources to either stimulus-driven or goal-relevant objects. Here, we asked if resting-state EEG connectivity correlated with attentional performance in schizophrenia patients. Materials and Methods: Resting-state EEG recordings were obtained from 72 stabilized patients with schizophrenia. Lagged phase synchronization (LPS) was used to measure whole-brain source-based functional connectivity between 84 intra-cortical current sources determined by eLORETA (exact low-resolution brain electromagnetic tomography) for five frequencies. The Conners' Continuous Performance Test-II (CPT-II) was administered for evaluating attentional performance. Linear regression with a non-parametric permutation randomization procedure was used to examine the correlations between the whole-brain functional connectivity and the CPT-II measures. Results: Greater beta-band right hemispheric fusiform gyrus (FG)-lingual gyrus (LG) functional connectivity predicted higher CPT-II variability scores (r = 0.44, p < 0.05, corrected), accounting for 19.5% of variance in the CPT-II VAR score. Greater gamma-band right hemispheric functional connectivity between the cuneus (Cu) and transverse temporal gyrus (TTG) and between Cu and the superior temporal gyrus (STG) predicted higher CPT-II hit reaction time (HRT) scores (both r = 0.50, p < 0.05, corrected), accounting for 24.6% and 25.1% of variance in the CPT-II HRT score, respectively. Greater gamma-band right hemispheric Cu-TTG functional connectivity predicted higher CPT-II HRT standard error (HRTSE) scores (r = 0.54, p < 0.05, corrected), accounting for 28.7% of variance in the CPT-II HRTSE score. Conclusions: Our study indicated that increased right hemispheric resting-state EEG functional connectivity at high frequencies was correlated with poorer focused attention in schizophrenia patients. If replicated, novel approaches to modulate these networks may yield selective, potent interventions for improving attention deficits in schizophrenia.

TMS-Induced Modulation of EEG Functional Connectivity Is Affected by the E-Field Orientation.

Coregistration of transcranial magnetic stimulation (TMS) and electroencephalography (EEG) allows non-invasive probing of brain circuits: TMS induces brain activation due to the generation of a properly oriented focused electric field (E-field) using a coil placed on a selected position over the scalp, while EEG captures the effects of the stimulation on brain electrical activity. Moreover, the combination of these techniques allows the investigation of several brain properties, including brain functional connectivity. The choice of E-field parameters, such as intensity, orientation, and position, is crucial for eliciting cortex-specific effects. Here, we evaluated whether and how the spatial pattern, i.e., topography and strength of functional connectivity, is modulated by the stimulus orientation. We systematically altered the E-field orientation when stimulating the left pre-supplementary motor area and showed an increase of functional connectivity in areas associated with the primary motor cortex and an E-field orientation-specific modulation of functional connectivity intensity.

Targeted rhythmic visual stimulation at individual participants’ intrinsic alpha frequency causes selective increase of occipitoparietal BOLD-fMRI and EEG functional connectivity

Neural oscillations in distinct frequency bands are ubiquitous in the brain and play a role in many cognitive processes. The "communication by coherence" hypothesis, poses that the synchronization through phase coupling of frequency-specific neural oscillations regulate information flow across distribute brain regions. Specifically, the posterior alpha frequency band (7-12Hz) is thought to gate bottom-up visual information flow by inhibition during visual processing. Evidence shows that increased alpha phase coherency positively correlates with functional connectivity in resting state connectivity networks, supporting alpha mediates neural communication through coherency. However, these findings have mainly been derived from spontaneous changes in the ongoing alpha rhythm. In this study, we experimentally modulate the alpha rhythm by targeting individuals' intrinsic alpha frequency with sustained rhythmic light to investigate alpha-mediated synchronous cortical activity in both EEG and fMRI. We hypothesize increased alpha coherency and fMRI connectivity should arise from modulation of the intrinsic alpha frequency (IAF) as opposed to control frequencies in the alpha range. Sustained rhythmic and arrhythmic stimulation at the IAF and at neighboring frequencies within the alpha band range (7-12Hz) was implemented and assessed in a separate EEG and fMRI study. We observed increased cortical alpha phase coherency in the visual cortex during rhythmic stimulation at the IAF as in comparison to rhythmic stimulation of control frequencies. In the fMRI, we found increased functional connectivity for stimulation at the IAF in visual and parietal areas as compared to other rhythmic control frequencies by correlating time courses from a set of regions of interest for the different stimulation conditions and applying network-based statistics. This suggests that rhythmic stimulation at the IAF frequency induces a higher degree of synchronicity of neural activity across the occipital and parietal cortex, which supports the role of the alpha oscillation in gating information flow during visual processing.

Prediction of Tinnitus Treatment Outcomes Based on EEG Sensors and TFI Score Using Deep Learning

Tinnitus is a hearing disorder that is characterized by the perception of sounds in the absence of an external source. Currently, there is no pharmaceutical cure for tinnitus, however, multiple therapies and interventions have been developed that improve or control associated distress and anxiety. We propose a new Artificial Intelligence (AI) algorithm as a digital prognostic health system that models electroencephalographic (EEG) data in order to predict patients' responses to tinnitus therapies. The EEG data was collected from patients prior to treatment and 3-months following a sound-based therapy. Feature selection techniques were utilised to identify predictive EEG variables with the best accuracy. The patients' EEG features from both the frequency and functional connectivity domains were entered as inputs that carry knowledge extracted from EEG into AI algorithms for training and predicting therapy outcomes. The AI models differentiated the patients' outcomes into either therapy responder or non-responder, as defined by their Tinnitus Functional Index (TFI) scores, with accuracies ranging from 98%-100%. Our findings demonstrate the potential use of AI, including deep learning, for predicting therapy outcomes in tinnitus. The research suggests an optimal configuration of the EEG sensors that are involved in measuring brain functional changes in response to tinnitus treatments. It identified which EEG electrodes are the most informative sensors and how the EEG frequency and functional connectivity can better classify patients into the responder and non-responder groups. This has potential for real-time monitoring of patient therapy outcomes at home.

Abnormalities of EEG Functional Connectivity and Effective Connectivity in Children with Autism Spectrum Disorder.

Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental disorder that interferes with normal brain development. Brain connectivity may serve as a biomarker for ASD in this respect. This study enrolled a total of 179 children aged 3-10 years (90 typically developed (TD) and 89 with ASD). We used a weighted phase lag index and a directed transfer function to investigate the functional and effective connectivity in children with ASD and TD. Our findings indicated that patients with ASD had local hyper-connectivity of brain regions in functional connectivity and simultaneous significant decrease in effective connectivity across hemispheres. These connectivity abnormalities may help to find biomarkers of ASD.

Evaluating brain spectral and connectivity differences between silent mind-wandering and trance states.

Trance is an altered state of consciousness characterized by alterations in cognition. In general, trance states induce mental silence (i.e., cognitive thought reduction), and mental silence can induce trance states. Conversely, mind-wandering is the mind's propensity to stray its attention away from the task at hand and toward content irrelevant to the current moment, and its main component is inner speech. Building on the previous literature on mental silence and trance states and incorporating inverse source reconstruction advances, the study's objectives were to evaluate differences between trance and mind-wandering states using: (1) electroencephalography (EEG) power spectra at the electrode level, (2) power spectra at the area level (source reconstructed signal), and (3) EEG functional connectivity between these areas (i.e., how they interact). The relationship between subjective trance depths ratings and whole-brain connectivity during trance was also evaluated. Spectral analyses revealed increased delta and theta power in the frontal region and increased gamma in the centro-parietal region during mind-wandering, whereas trance showed increased beta and gamma power in the frontal region. Power spectra at the area level and pairwise comparisons of the connectivity between these areas demonstrated no significant difference between the two states. However, subjective trance depth ratings were inversely correlated with whole-brain connectivity in all frequency bands (i.e., deeper trance is associated with less large-scale connectivity). Trance allows one to enter mentally silent states and explore their neurophenomenological processes. Limitations and future directions are discussed.

Spontaneity matters! Network alterations before and after spontaneous and active facial self-touches: An EEG functional connectivity study

BackgroundDespite humans frequently performing spontaneous facial self-touches (sFST), the function of this behavior remains speculative. sFST have been discussed in the context of self-regulation, emotional homeostasis, working memory processes, and attention focus. First evidence indicates that sFST and active facial self-touches (aFST) are neurobiologically different phenomena. The aim of the present analysis was to examine EEG-based connectivity in the course of sFST and aFST to test the hypotheses that sFST affect brain network interactions relevant for other than sensorimotor processes. MethodsTo trigger spontaneous FST a previously successful setting was used: 60 healthy participants manually explored two haptic stimuli and held the shapes of the stimuli in memory for a 14 min retention interval. Afterwards the shapes were drawn on a sheet of paper. During the retention interval, artifact-free EEG-data of 97 sFST by 32 participants were recorded. At the end of the experiment, the participants performed aFST with both hands successively. For the EEG-data, connectivity was computed and compared between the phases before and after sFST and aFST and between the respective before-and the after-phases. ResultsFor the before-after comparison, brainwide distributed significant connectivity differences (p < .00079) were observed for sFST, but not for aFST. Additionally, comparing the before- and after-phases of sFST and aFST, respectively, revealed increased similarity between the after-phases than between the before-phases. ConclusionThe results support the assumption that sFST and aFST are neurobiologically different phenomena. Furthermore, the aligned network properties of the after-phases compared to the before-phases indicate that sFST serve self-regulatory functions that aFST do not serve.

Pre vs post‐task modulation of resting state EEG functional connectivity and network topology

AbstractBackgroundRecently, we have shown a significant decrease in resting‐state EEG (rsEEG) spectral power at post‐task as compared to pre‐task cognitive engagement in older adults. Further, older adults with mild cognitive impairment (MCI) showed greater decreases in rsEEG in temporal and central cortical regions at post‐task than healthy controls (HC). Here, we present functional connectivity (FC) and network topology derived from the same pre‐ to post‐rsEEG changes.MethodWe studied a sample of community dwelling African American older participants over age 60 years, consisting of consensus‐based diagnosis for 58 HC and 41 MCI. rsEEGs were recorded before and after approximately 20 minutes of a visual motion direction discrimination task. FC between channels was assessed by the phase lag index, while minimum spanning tree (MST) was used to characterize network topology.ResultFor both HC and MCI groups, when compared to pre‐task measurements, the post‐task FC was significantly increased within the middle frontal, middle central, and left temporal regions in the lower alpha frequency band. An opposite pattern with decreased FC was found within the right parietal region in the delta, theta, lower alpha, upper alpha, and beta frequency bands. In contrast, compared to HC, the MCI group did show significantly decreased FC within the middle frontal region in the delta band, following the cognitive task. All results from network topology failed to demonstrate differences between HC and MCI. On average, both groups' post‐task functional network topology showed significant decreased eccentricity and diameter in the delta and beta bands, indicating a higher global network integration.ConclusionOur findings suggest that visual‐based cognitive engagement affects both post‐task FC and network topology similarly in both HC and MCI. For the most part, increased FC and functional topology in both groups are likely associated with after‐task effects, resulting from additional time to return to baseline, though decreased FC in MCI as compared to HC was seen within the middle frontal region in the delta. Our results demonstrate that both FC and topoloty are responsive to cognitive challenge, though frontal/executive‐driven differences in FC may be an area of future promise in distinguishing HC and MCI.

The effect of low-frequency rTMS on auditory hallucinations, EEG source localization and functional connectivity in schizophrenia

BackgroundLow-frequency repetitive transcranial magnetic stimulation (LF-rTMS) diminishes auditory hallucinations (AHs). The aims of our study were a) to assess the efficacy of LF-rTMS in a randomized, sham-controlled double-blind alignment, b) to identify the electrophysiological changes accompanying the LF-rTMS, and c) to identify the influence of LF-rTMS on brain functional connectivity (FC). MethodsNineteen schizophrenia patients with antipsychotic-resistant AHs were randomized to either active (n = 10) or sham (n = 9) LF-rTMS administered over the left temporo-parietal region for ten days. The clinical effect was assessed by the Auditory Hallucination Rating Scale (AHRS). The localization of the differences in electrical activity was identified by standardized low resolution brain electromagnetic tomography (sLORETA) and FC was measured by lagged phase synchronization. ResultsAHRS scores were significantly improved for patients receiving active rTMS compared to the sham (median reduction: 40 % vs 12 %; p = 0.01). sLORETA revealed a decrease of alpha-2, beta-1,-2 bands in the left hemisphere in the active group. Active rTMS led to a decrease of the lagged phase connectivity in beta bands originating in areas close to the site of stimulation, and to a prevailing increase of alpha-2 FC. No significant differences in current density or FC were observed in the sham group. LimitationsLimitations to our study included the small group sizes, and the disability of LORETA to assess subcortical neuronal activity. ConclusionsLF-rTMS attenuated AHs and induced a decrease of higher frequency bands on the left hemisphere. The FC changes support the assumption that LF-rTMS is linked to the modulation of cortico-cortical coupling.

Estimated EEG functional connectivity and aperiodic component induced by vagal nerve stimulation in patients with drug-resistant epilepsy.

Vagal nerve stimulation (VNS) improves seizure frequency and quality of life in patients with drug-resistant epilepsy (DRE), although the exact mechanism is not fully understood. Previous studies have evaluated the effect of VNS on functional connectivity using the phase lag index (PLI), but none has analyzed its effect on EEG aperiodic parameters (offset and exponent), which are highly conserved and related to physiological functions. This study aimed to evaluate the effect of VNS on PLI and aperiodic parameters and infer whether these changes correlate with clinical responses in subjects with DRE. PLI, exponent, and offset were derived for each epoch (and each frequency band for PLI), on scalp-derived 64-channel EEG traces of 10 subjects with DRE, recorded before and 1 year after VNS. PLI, exponent, and offset were compared before and after VNS for each patient on a global basis, individual scalp regions, and channels and separately in responders and non-responders. A correlation analysis was performed between global changes in PLI and aperiodic parameters and clinical response. PLI (global and regional) decreased after VNS for gamma and delta bands and increased for an alpha band in responders, but it was not modified in non-responders. Aperiodic parameters after VNS showed an opposite trend in responders vs. non-responders: both were reduced in responders after VNS, but they were increased in non-responders. Changes in aperiodic parameters correlated with the clinical response. This study explored the action of VNS therapy from a new perspective and identified EEG aperiodic parameters as a new and promising method to analyze the efficacy of neuromodulation.

Deep brain stimulation improves electroencephalogram functional connectivity of patients with minimally conscious state.

Deep brain stimulation (DBS) is a potential neuromodulatory therapy that enhances recovery from disorders of consciousness, especially minimally conscious state (MCS). This study measured the effects of DBS on the brain and explored the underlying mechanisms of DBS on MCS. Nine patients with MCS were recruited for this study. The neuromodulation effects of 100 Hz DBS were explored via cross-control experiments. Coma Recovery Scale-Revised (CRS-R) and EEG were recorded, and corresponding functional connectivity and network parameters were calculated. Our results showed that 100 Hz DBS could improve the functional connectivity of the whole, local and local-local brain regions, while no significant change in EEG functional connectivity was observed in sham DBS. The whole brain's network parameters (clustering coefficient, path length, and small world characteristic) were significantly improved. In addition, a significant increase in the CRS-R and functional connectivity of three MCS patients who received 100 Hz DBS for 6months were observed. This study showed that DBS improved EEG functional connectivity and brain networks, indicating that the long-term use of DBS could improve the level of consciousness of MCS patients.

Resting-state EEG functional connectivity predicts post-traumatic stress disorder subtypes in veterans

Objective. Post-traumatic stress disorder (PTSD) is highly heterogeneous, and identification of quantifiable biomarkers that could pave the way for targeted treatment remains a challenge. Most previous electroencephalography (EEG) studies on PTSD have been limited to specific handpicked features, and their findings have been highly variable and inconsistent. Therefore, to disentangle the role of promising EEG biomarkers, we developed a machine learning framework to investigate a wide range of commonly used EEG biomarkers in order to identify which features or combinations of features are capable of characterizing PTSD and potential subtypes. Approach. We recorded 5 min of eyes-closed and 5 min of eyes-open resting-state EEG from 202 combat-exposed veterans (53% with probable PTSD and 47% combat-exposed controls). Multiple spectral, temporal, and connectivity features were computed and logistic regression, random forest, and support vector machines with feature selection methods were employed to classify PTSD. To obtain robust results, we performed repeated two-layer cross-validation to test on an entirely unseen test set. Main results. Our classifiers obtained a balanced test accuracy of up to 62.9% for predicting PTSD patients. In addition, we identified two subtypes within PTSD: one where EEG patterns were similar to those of the combat-exposed controls, and another that were characterized by increased global functional connectivity. Our classifier obtained a balanced test accuracy of 79.4% when classifying this PTSD subtype from controls, a clear improvement compared to predicting the whole PTSD group. Interestingly, alpha connectivity in the dorsal and ventral attention network was particularly important for the prediction, and these connections were positively correlated with arousal symptom scores, a central symptom cluster of PTSD. Significance. Taken together, the novel framework presented here demonstrates how unsupervised subtyping can delineate heterogeneity and improve machine learning prediction of PTSD, and may pave the way for better identification of quantifiable biomarkers.

Interictal Gamma Event Connectivity Differentiates the Seizure Network and Outcome in Patients after Temporal Lobe Epilepsy Surgery.

Studies of interictal EEG functional connectivity in the epileptic brain seek to identify abnormal interactions between brain regions involved in generating seizures, which clinically often is defined by the seizure onset zone (SOZ). However, there is evidence for abnormal connectivity outside the SOZ (NSOZ), and removal of the SOZ does not always result in seizure control, suggesting, in some cases, that the extent of abnormal connectivity indicates a larger seizure network than the SOZ. To better understand the potential differences in interictal functional connectivity in relation to the seizure network and outcome, we computed event connectivity in the theta (4-8 Hz, ThEC), low-gamma (30-55 Hz, LGEC), and high-gamma (65-95 Hz, HGEC) bands from interictal depth EEG recorded in surgical patients with medication-resistant seizures suspected to begin in the temporal lobe. Analysis finds stronger LGEC and HGEC in SOZ than NSOZ of seizure-free (SF) patients (p = 1.10e-9, 0.0217), but no difference in not seizure-free (NSF) patients. There were stronger LGEC and HGEC between mesial and lateral temporal SOZ of SF than NSF patients (p = 0.00114, 0.00205), and stronger LGEC and ThEC in NSOZ of NSF than SF patients (p = 0.0089, 0.0111). These results show that event connectivity is sensitive to differences in the interactions between regions in SOZ and NSOZ and SF and NSF patients. Patients with differential strengths in event connectivity could represent a well-localized seizure network, whereas an absence of differences could indicate a larger seizure network than the one localized by the SOZ and higher likelihood for seizure recurrence.

Graph Theoretical Analysis of EEG Functional Connectivity Patterns and Fusion with Physiological Signals for Emotion Recognition.

Emotion recognition is a key attribute for realizing advances in human-computer interaction, especially when using non-intrusive physiological sensors, such as electroencephalograph (EEG) and electrocardiograph. Although functional connectivity of EEG has been utilized for emotion recognition, the graph theory analysis of EEG connectivity patterns has not been adequately explored. The exploitation of brain network characteristics could provide valuable information regarding emotions, while the combination of EEG and peripheral physiological signals can reveal correlation patterns of human internal state. In this work, a graph theoretical analysis of EEG functional connectivity patterns along with fusion between EEG and peripheral physiological signals for emotion recognition has been proposed. After extracting functional connectivity from EEG signals, both global and local graph theory features are extracted. Those features are concatenated with statistical features from peripheral physiological signals and fed to different classifiers and a Convolutional Neural Network (CNN) for emotion recognition. The average accuracy on the DEAP dataset using CNN was 55.62% and 57.38% for subject-independent valence and arousal classification, respectively, and 83.94% and 83.87% for subject-dependent classification. Those scores went up to 75.44% and 78.77% for subject-independent classification and 88.27% and 90.84% for subject-dependent classification using a feature selection algorithm, exceeding the current state-of-the-art results.