Valproate treatment normalizes EEG functional connectivity in successfully treated idiopathic generalized epilepsy patients.

AIM: To investigate the effect of chronic VPA treatment of EEG functional connectivity in successfully treated idiopathic generalized epilepsy (IGE) patients. PATIENTS AND METHODS: 19-channel waking, resting-state EEG records of 26 IGE patients were analyzed before treatment (IGE) and after the 90th day of treatment (VPA), in seizure-free condition. Three minutes of artifact-free EEG background activity (without epileptiform potentials) was analyzed for each patient in both conditions. A group of 26 age-matched healthy normative control persons (NC) was analyzed in the same way. All the EEG samples were processed to LORETA (Low Resolution Electromagnetic Tomography) to localize multiple distributed sources of EEG activity. Current source density time series were generated for 33 regions of interest (ROI) in each hemisphere for four frequency bands. Pearson correlation coefficients (R) were computed between all ROIs in each hemisphere, for four bands across the investigated samples. R values corresponded to intrahemispheric, cortico-cortical functional EEG connectivity (EEGfC). Group and condition differences were analyzed by statistical parametric network method. MAIN RESULTS: p<0.05, corrected for multiple comparisons: (1) The untreated IGE group showed increased EEGfC in the delta and theta bands, and decreased EEGfC in the alpha band (as compared to the NC group); (2) VPA treatment normalized EEGfC in the delta, theta and alpha bands; and (3) degree of normalization depended on frequency band and cortical region. CONCLUSIONS: VPA treatment normalizes EEGfC in IGE patients.

Neurophysiology of juvenile myoclonic epilepsy: EEG-based network and graph analysis of the interictal and immediate preictal states.

INTRODUCTION: The neuronal mechanisms of enduring seizure propensity and seizure precipitation in juvenile myoclonic epilepsy (JME) are not known. We investigated these issues, within the framework of the "network concept" of epilepsy. METHODS: Design1: 19, unmedicated JME patients were compared with nineteen, age-, and sex-matched normal control persons (NC). A total of 120s, artifact-free, paroxysm-free, eyes-closed, resting state EEG background activity was analyzed for each person. Design2: interictal and immediate preictal periods of the JME patients were compared in order to explore interictal-preictal network differences. For both comparison designs, statistically significant differences of EEG functional connectivity (EEGfC), nodal and global graph parameters were evaluated. MAIN RESULTS: Design1: maximum abnormalities were: increased delta, theta, alpha1 EEGfC and decreased alpha2 and beta EEGfC in the JME group as compared to the NC group, mainly among cortical areas that are involved in sensory-motor integration. Nodal degree and efficiency of three, medial, basal frontal nodes were greater in JME than in NC, in the alpha1 band. Design2: preictal delta EEGfC showed further increase in the above-mentioned areas, as compared to the interictal state. DISCUSSION: Increased EEGfC indicates a hypercoupled state among the specified cortical areas. This interictal abnormality further increases in the preictal state. Nodal graph statistics indicates abnormal neuronal dynamics in the cortical area that is the ictal onset zone in JME. SIGNIFICANCE: Interictal and preictal neuronal dysfunction has been described in terms of network dynamics and topography in JME patients. Forthcoming investigations of seizure precipitation and therapeutic drug effects are encouraged on this basis.

Remission of benign epilepsy with rolandic spikes: an EEG-based connectivity study at the onset of the disease and at remission.

PURPOSE: The neuronal mechanisms of remission of epilepsy are not known. Based on the principles of the "network theory of epilepsy" we postulated the existence of abnormal cortico-cortical interactions at the onset of epilepsy (Hypothesis-1), and postulated that remission is associated with the decrease or disappearance of the abnormal quantitative EEG findings (Hypothesis-2). METHODS: Four children with benign epilepsy with rolandic sharp waves (BERS) were investigated. 21-channel EEG was recorded at the onset of the disease (Setting No. 1) and in remission (Setting No. 2). Local EEG synchronization was estimated by LORETA (low resolution electromagnetic tomography). Remote EEG synchronization (intra-hemispheric, cortico-cortical EEG functional connectivity, EEGfC) was computed by the LSC (LORETA Source Correlation) method, among 23 regions of interest (ROI) in both hemispheres. Both local and remote EEG synchronization were evaluated in very narrow frequency bands of 1Hz bandwidth (VNB), from 1 to 25Hz. RESULTS: Individual results were presented. Abnormal but topographically very dissimilar LORETA and LSC findings were found at the onset of the disease. The disappearance of the initial abnormalities was found in Setting No. 2. An unforeseen finding was the presence of abnormal EEGfC results in Setting No. 2. DISCUSSION: The authors confirmed both hypotheses. The dissimilarity of the initial abnormalities is in accord with the network concept of epilepsy and the etiology of BERS. The disappearance of the initial abnormalities reflects "normalization" of network dynamics while the emergence of new EEGfC abnormalities is interpreted as "compensation". CONCLUSION: EEG-based local and remote connectivity (EEGfC) are appropriate tools to describe network dynamics in the active state of BERS and in remission.

EEG functional connectivity of the intrahemispheric cortico-cortical network of idiopathic generalized epilepsy.

AIMS: Intrahemispheric, cortico-cortical EEG functional connectivity (fC) was investigated in untreated patients with idiopathic generalized epilepsy (IGE) in this explorative study. PATIENTS AND METHODS: Group comparison was carried out between 19, drug-naive IGE patients and 19, matched healthy persons. 90×2s of 19 channels waking, interictal background EEG signal (without epileptiform potentials) were processed to the LORETA (low resolution electromagnetic tomography) software to compute current source density for 2394 voxels representing parcels of the cerebral cortex for 25 very narrow bands of 1Hz bandwidth (VNBs) from 1 to 25Hz. EEG fC was investigated among the already localized sources. Pearson correlation coefficients (R) were computed among the 33 regions of interest (ROI) within the left and within the right hemisphere, separately. Group differences were computed by means of t-statistics. Corrected p<0.05 differences were accepted as statistically significant. MAIN RESULTS: (1) The anatomical patterns of the fC differences showed great frequency-dependency. (2) Hemispheric asymmetry was prominent within most VNBs. (3) Decreased fC in the IGE group was found across all VNBs in the 1-6Hz frequency range as compared to mixed patterns comprising both increased and decreased fC at >6Hz frequencies. (4) In the 5-25Hz range, decreased fC dominated in the anterior, increased fC in the posterior parts of the cortex. (5) The results delineated an anterior and a posterior network. DISCUSSION: (1) Decreased fC in the 1-6Hz band might indicate some relationship to yet hidden structure network abnormalities. (2) The anatomical patterns of fC indicate frequency-dependent, pathological coupling and decoupling processes in the interictal state. (3) The two networks might help to understand seizure liability and seizure precipitation in IGE. SIGNIFICANCE: This is the first study to explore EEG fC in the interictal condition of IGE patients. The importance of EEG frequencies in evaluating fC in IGE was demonstrated and starting points for further research were given.