Evoked directional network characteristics of epileptogenic tissue derived from single pulse electrical stimulation.

We investigated effective networks constructed from single pulse electrical stimulation (SPES) in epilepsy patients who underwent intracranial electrocorticography. Using graph analysis, we compared network characteristics of tissue within and outside the epileptogenic area. In 21 patients with subdural electrode grids (1 cm interelectrode distance), we constructed a binary, directional network derived from SPES early responses (<100 ms). We calculated in-degree, out-degree, betweenness centrality, the percentage of bidirectional, receiving and activating connections, and the percentage of connections toward the (non-)epileptogenic tissue for each node in the network. We analyzed whether these network measures were significantly different in seizure onset zone (SOZ)-electrodes compared to non-SOZ electrodes, in resected area (RA)-electrodes compared to non-RA electrodes, and in seizure free compared to not seizure-free patients. Electrodes in the SOZ/RA showed significantly higher values for in-degree and out-degree, both at group level, and at patient level, and more so in seizure-free patients. These differences were not observed for betweenness centrality. There were also more bidirectional and fewer receiving connections in the SOZ/RA in seizure-free patients. It appears that the SOZ/RA is densely connected with itself, with only little input arriving from non-SOZ/non-RA electrodes. These results suggest that meso-scale effective network measures are different in epileptogenic compared to normal brain tissue. Local connections within the SOZ/RA are increased and the SOZ/RA is relatively isolated from the surrounding cortex. This offers the prospect of enhanced prediction of epilepsy-prone brain areas using SPES.

Tailoring epilepsy surgery with fast ripples in the intraoperative electrocorticogram.

OBJECTIVE: Intraoperative electrocorticography (ECoG) can be used to delineate the resection area in epilepsy surgery. High-frequency oscillations (HFOs; 80-500 Hz) seem better biomarkers for epileptogenic tissue than spikes. We studied how HFOs and spikes in combined pre- and postresection ECoG predict surgical outcome in different tailoring approaches. METHODS: We, retrospectively, marked HFOs, divided into fast ripples (FRs; 250-500 Hz) and ripples (80-250 Hz), and spikes in pre- and postresection ECoG sampled at 2,048 Hz in people with refractory focal epilepsy. We defined four groups of electroencephalography (EEG) event occurrence: pre+post- (+/-), pre+post+ (+/+), pre-post+ (-/+) and pre-post- (-/-). We subcategorized three tailoring approaches: hippocampectomy with tailoring for neocortical involvement; lesionectomy of temporal lesions with tailoring for mesiotemporal involvement; and lesionectomy with tailoring for surrounding neocortical involvement. We compared the percentage of resected pre-EEG events, time to recurrence, and the different tailoring approaches to outcome (seizure-free vs recurrence). RESULTS: We included 54 patients (median age, 15.5 years; 25 months of follow-up; 30 seizure free). The percentage of resected FRs, ripples, or spikes in pre-ECoG did not predict outcome. The occurrence of FRs in post-ECoG, given FRs in pre-ECoG (+/-, +/+), predicted outcome (hazard ratio, 3.13; confidence interval = 1.22-6.25; p = 0.01). Seven of 8 patients without spikes in pre-ECoG were seizure free. The highest predictive value for seizure recurrence was presence of FRs in post-ECoG for all tailoring approaches. INTERPRETATION: FRs that persist before and after resection predict poor postsurgical outcome. These findings hold for different tailoring approaches. FRs can thus be used for tailoring epilepsy surgery with repeated intraoperative ECoG measurements. Ann Neurol 2017;81:664-676.

ER-detect: a pipeline for robust detection of early evoked responses in BIDS-iEEG electrical stimulation data.

Human brain connectivity can be measured in different ways. Intracranial EEG (iEEG) measurements during single pulse electrical stimulation provide a unique way to assess the spread of electrical information with millisecond precision. To provide a robust workflow to process these cortico-cortical evoked potential (CCEP) data and detect early evoked responses in a fully automated and reproducible fashion, we developed Early Response (ER)-detect. ER-detect is an open-source Python package and Docker application to preprocess BIDS structured iEEG data and detect early evoked CCEP responses. ER-detect can use three response detection methods, which were validated against 14-manually annotated CCEP datasets from two different sites by four independent raters. Results showed that ER-detect's automated detection performed on par with the inter-rater reliability (Cohen's Kappa of ~0.6). Moreover, ER-detect was optimized for processing large CCEP datasets, to be used in conjunction with other connectomic investigations. ER-detect provides a highly efficient standardized workflow such that iEEG-BIDS data can be processed in a consistent manner and enhance the reproducibility of CCEP based connectivity results.

EEG-fMRI correlation patterns in the presurgical evaluation of focal epilepsy: a comparison with electrocorticographic data and surgical outcome measures.

EEG-correlated functional MRI (EEG-fMRI) visualizes brain regions associated with interictal epileptiform discharges (IEDs). This technique images the epileptiform network, including multifocal, superficial and deeply situated cortical areas. To understand the role of EEG-fMRI in presurgical evaluation, its results should be validated relative to a gold standard. For that purpose, EEG-fMRI data were acquired for a heterogeneous group of surgical candidates (n=16) who were later implanted with subdural grids and strips (ECoG). The EEG-fMRI correlation patterns were systematically compared with brain areas involved in IEDs ECoG, using a semi-automatic analysis method, as well as to the seizure onset zone, resected area, and degree of seizure freedom. In each patient at least one of the EEG-fMRI areas was concordant with an interictally active ECoG area, always including the early onset area of IEDs in the ECoG data. This confirms that EEG-fMRI reflects a pattern of onset and propagation of epileptic activity. At group level, 76% of the BOLD regions that were covered with subdural grids, were concordant with interictally active ECoG electrodes. Due to limited spatial sampling, 51% of the BOLD regions were not covered with electrodes and could, therefore, not be validated. From an ECoG perspective it appeared that 29% of the interictally active ECoG regions were missed by EEG-fMRI and that 68% of the brain regions were correctly identified as inactive with EEG-fMRI. Furthermore, EEG-fMRI areas included the complete seizure onset zone in 83% and resected area in 93% of the data sets. No clear distinction was found between patients with a good or poor surgical outcome: in both patient groups, EEG-fMRI correlation patterns were found that were either focal or widespread. In conclusion, by comparison of EEG-fMRI with interictal invasive EEG over a relatively large patient population we were able to show that the EEG-fMRI correlation patterns are spatially accurate at the level of neurosurgical units (i.e. anatomical brain regions) and reflect the underlying network of IEDs. Therefore, we expect that EEG-fMRI can play an important role for the determination of the implantation strategy.

Inverse modeling in magnetic source imaging: Comparison of MUSIC, SAM(g2), and sLORETA to interictal intracranial EEG.

Magnetoencephalography (MEG) is used in the presurgical work-up of patients with focal epilepsy. In particular, localization of MEG interictal spikes may guide or replace invasive electroencephalography monitoring that is required in difficult cases. From literature, it is not clear which MEG source localization method performs best in this clinical setting. Therefore, we applied three source localization methods to the same data from a large patient group for which a gold standard, interictal spikes as identified in electrocorticography (ECoG), was available. The methods used were multiple signal classification (MUSIC), Synthetic Aperture Magnetometry kurtosis [SAM(g2)], and standardized low-resolution electromagnetic tomography. MEG and ECoG data from 38 patients with refractory focal epilepsy were obtained. Results of the three source localization methods applied to the interictal MEG data were assigned to predefined anatomical regions. Interictal spikes as identified in ECoG were also assigned to these regions. Identified regions by each MEG method were compared to ECoG. Sensitivity and positive predictive value (PPV) of each MEG method were calculated. All three MEG methods showed a similar overall correlate with ECoG spikes, but the methods differ in which regions they detect. The choice of the inverse model thus has an unexpected influence on the results of magnetic source imaging. Combining inverse methods and seeking consensus can be used to improve specificity at the cost of some sensitivity. Combining MUSIC with SAM(g2) gives the best results (sensitivity = 38% and PPV = 82%).

Time-frequency analysis of single pulse electrical stimulation to assist delineation of epileptogenic cortex.

Epilepsy surgery depends on reliable pre-surgical markers of epileptogenic tissue. The current gold standard is the seizure onset zone in ictal, i.e. chronic, electrocorticography recordings. Single pulse electrical stimulation can evoke epileptic, spike-like responses in areas of seizure onset also recorded by electrocorticography. Recently, spontaneous pathological high-frequency oscillations (80-520 Hz) have been observed in the electrocorticogram that are related to epileptic spikes, but seem more specific for epileptogenic cortex. We wanted to see whether a quantitative electroencephalography analysis using time-frequency information including the higher frequency range could be applied to evoked responses by single pulse electrical stimulation, to enhance its specificity and clinical use. Electrocorticography data were recorded at a 2048-Hz sampling rate from 13 patients. Single pulse electrical stimulation (10 stimuli, 1 ms, 8 mA, 0.2 Hz) was performed stimulating pairs of adjacent electrodes. A time-frequency analysis based on Morlet wavelet transformation was performed in a [-1 s : 1 s] time interval around the stimulus and a frequency range of 10-520 Hz. Significant (P = 0.05) changes in power spectra averaged for 10 epochs were computed, resulting in event-related spectral perturbation images. In these images, time-frequency analysis of single pulse-evoked responses, in the range of 10-80 Hz for spikes, 80-250 Hz for ripples and 250-520 Hz for fast ripples, were scored by two observers independently. Sensitivity, specificity and predictive value of time-frequency single pulse-evoked responses in the three frequency ranges were compared with seizure onset zone and post-surgical outcome. In all patients, evoked responses included spikes, ripples and fast ripples. For the seizure onset zone, the median sensitivity of time-frequency single pulse-evoked responses decreased from 100% for spikes to 67% for fast ripples and the median specificity increased from 17% for spikes to 79% for fast ripples. A median positive predictive value for the evoked responses in the seizure onset zone of 17% was found for spikes, 26% for ripples and 37% for fast ripples. Five out of seven patients with <50% of fast ripples removed by resection had a poor outcome. A wavelet transform-based time-frequency analysis of single pulse electrical stimulation reveals evoked responses in the frequency range of spikes, ripples and fast ripples. We demonstrate that time-frequency analysis of single pulse electrical stimulation can assist in delineation of the epileptogenic cortex using time-frequency single pulse-evoked fast ripples as a potential new marker.

Accurate differentiation between physiological and pathological ripples recorded with scalp-EEG.

OBJECTIVE: To compare scalp-EEG recorded physiological ripples co-occurring with vertex waves to pathological ripples co-occurring with interictal epileptiform discharges (IEDs). METHODS: We marked ripples in sleep EEGs of children. We compared the start of ripples to vertex wave- or IED-start, and duration, frequency, and root mean square (RMS) amplitude of physiological and pathological ripples using multilevel modeling. Ripples were classified as physiological or pathological using linear discriminant analysis. RESULTS: We included 40 children with and without epilepsy. Ripples started (χ2(1) = 38.59, p < 0.001) later if they co-occurred with vertex waves (108.2 ms after vertex wave-start) than if they co-occurred with IEDs (4.3 ms after IED-start). Physiological ripples had longer durations (75.7 ms vs 53.0 ms), lower frequencies (98.3 Hz vs 130.6 Hz), and lower RMS amplitudes (0.9 µV vs 1.8 µV, all p < 0.001) than pathological ripples. Ripples could be classified as physiological or pathological with 98 % accuracy. Ripples recorded in children with idiopathic or symptomatic epilepsy seemed to form two subgroups of pathological ripples. CONCLUSIONS: Ripples co-occurring with vertex waves or IEDs have different characteristics and can be differentiated as physiological or pathological with high accuracy. SIGNIFICANCE: This is the first study that compares physiological and pathological ripples recorded with scalp EEG.

A skew-based method for identifying intracranial EEG channels with epileptic activity without detecting spikes, ripples, or fast ripples.

OBJECTIVE: To develop a method for identifying intracranial EEG (iEEG) channels with epileptic activity without the need to detect spikes, ripples, or fast ripples. METHODS: We compared the skew of the distribution of power values from five minutes non-rapid eye movement stage N3 sleep for the 5-80 Hz, 80-250 Hz (ripple), and 250-500 Hz (fast ripple) bands of epileptic (located in seizure-onset or irritative zone) and non-epileptic iEEG channels recorded in patients with drug-resistant focal epilepsy. We optimized settings in 120 bipolar channels from 10 patients, compared the results to 120 channels from another 10 patients, and applied the method to channels of 12 individual patients. RESULTS: The distribution of power values was more skewed in epileptic than in non-epileptic channels in all three frequency bands. The differences in skew were correlated with the presence of spikes, ripples, and fast ripples. When classifying epileptic and non-epileptic channels, the mean accuracy over 12 patients was 0.82 (sensitivity: 0.76, specificity: 0.91). CONCLUSIONS: The 'skew method' can distinguish epileptic from non-epileptic channels with good accuracy and, in particular, high specificity. SIGNIFICANCE: This is an easy-to-apply method that circumvents the need to visually mark or automatically detect interictal epileptic events.

Ripples in scalp EEGs of children: co-occurrence with sleep-specific transients and occurrence across sleep stages.

Study Objectives: A dialogue between hippocampal ripples (80-250 Hz) and neocortical sleep-specific transients is important for memory consolidation. Physiological neocortical ripples can be recognized in scalp EEGs of children. We investigated how often scalp-EEG recorded ripples co-occur with different types of sleep-specific transients, the distribution and spatial extent of ripples with and without co-occurring sleep-specific transients, and the occurrence of ripples across sleep stages. Methods: We marked ripples in daytime sleep-EEGs of 19 children and determined for each ripple if it co-occurred with a sleep-specific transient. We compared the distribution of ripples without co-occurring sleep-specific transients to the distribution of all ripples. We estimated the spatial extent of simultaneously occurring ripples by counting how many EEG regions they comprised. We compared ripple rate per sleep stage using Friedman's analysis of variance and Wilcoxon signed-rank test. Results: 74.4 % of ripples co-occurred with sleep-specific transients: 27.8 % with vertex waves, 14.7 % with hypnagogic hypersynchrony, 13.7 % with slow waves, 12.2 % with spindles, and 6.0 % with K-complexes. Ripples without co-occurring sleep-specific transients showed the same central dominance but a significantly less pronounced midline dominance than the overall distribution pattern. Spatial extent was larger when ripples co-occurred with sleep-specific transients. Ripple rates during nonrapid eye movement (N) sleep stages N1 and N2 were higher than during N3 (T = 22.00, p = 0.02 and T = 23.00, p = 0.01). Conclusions: Scalp-EEG recorded physiological ripples co-occur with various sleep-specific EEG-transients, especially with vertex waves. These ripples occur most frequently during light sleep.

Electrocorticographic language mapping with a listening task consisting of alternating speech and music phrases.

OBJECTIVE: Electrocorticographic (ECoG) mapping of high gamma activity induced by language tasks has been proposed as a more patient friendly alternative for electrocortical stimulation mapping (ESM), the gold standard in pre-surgical language mapping of epilepsy patients. However, ECoG mapping often reveals more language areas than considered critical with ESM. We investigated if critical language areas can be identified with a listening task consisting of speech and music phrases. METHODS: Nine patients with implanted subdural grid electrodes listened to an audio fragment in which music and speech alternated. We analysed ECoG power in the 65-95 Hz band and obtained task-related activity patterns in electrodes over language areas. We compared the spatial distribution of sites that discriminated between listening to speech and music to ESM results using sensitivity and specificity calculations. RESULTS: Our listening task of alternating speech and music phrases had a low sensitivity (0.32) but a high specificity (0.95). CONCLUSIONS: The high specificity indicates that this test does indeed point to areas that are critical to language processing. SIGNIFICANCE: Our test cannot replace ESM, but this short and simple task can give a reliable indication where to find critical language areas, better than ECoG mapping using language tasks alone.

Residual fast ripples in the intraoperative corticogram predict epilepsy surgery outcome.

OBJECTIVE: We studied whether residual high-frequency oscillations (80-500 Hz; ripples, 80-250 Hz), especially fast ripples (FRs, 250-500 Hz), in post-resection intraoperative electrocorticography (ECoG) predicted seizure recurrence in comparison to residual interictal spikes and ictiform spike patterns. METHODS: We studied, retrospectively, ECoG recorded at 2,048 Hz after resection in a cohort of patients with refractory focal epilepsy. We analyzed occurrence and number of residual FRs, ripples, interictal spikes, and ictiform spike patterns within the last minute of each recording and compared these to seizure recurrence. RESULTS: We included 54 patients (median age 15.5 years) with 25 months median follow-up. Twenty-four patients had recurrent seizures. We found residual FRs, ripples, spikes, and ictiform spike patterns in 12, 51, 38, and 9 patients. Nine out of 12 patients with residual FRs had recurrent seizures (p = 0.016, positive predictive value 75%). Other ECoG events did not predict seizure recurrence. Patients with seizures had higher FR rates than seizure-free patients (p = 0.022). FRs near the resection and in distant pathologic areas could have changed the resection in 8 patients without harming functionally eloquent areas. One seizure-free patient had FRs in distant functionally eloquent areas. CONCLUSIONS: Residual FRs in post-resection ECoG are prognostic markers for seizure recurrence, especially if their number is high. Tailoring could rely on FRs, but requires careful assessment of the ECoG, as FRs in functionally eloquent areas might not be pathologic.