Data-driven electrophysiological feature based on deep learning to detect epileptic seizures.

Objective. To identify a new electrophysiological feature characterising the epileptic seizures, which is commonly observed in different types of epilepsy.Methods. We recorded the intracranial electroencephalogram (iEEG) of 21 patients (12 women and 9 men) with multiple types of refractory epilepsy. The raw iEEG signals of the early phase of epileptic seizures and interictal states were classified by a convolutional neural network (Epi-Net). For comparison, the same signals were classified by a support vector machine (SVM) using the spectral power and phase-amplitude coupling. The features learned by Epi-Net were derived by a modified integrated gradients method. We considered the product of powers multiplied by the relative contribution of each frequency amplitude as a data-driven epileptogenicity index (d-EI). We compared the d-EI and other conventional features in terms of accuracy to detect the epileptic seizures. Finally, we compared the d-EI among the electrodes to evaluate its relationship with the resected area and the Engel classification.Results. Epi-Net successfully identified the epileptic seizures, with an area under the receiver operating characteristic curve of 0.944 ± 0.067, which was significantly larger than that of the SVM (0.808 ± 0.253,n =21;p =0.025). The learned iEEG signals were characterised by increased powers of 17-92 Hz and >180 Hz in addition to decreased powers of other frequencies. The proposed d-EI detected them with better accuracy than the other iEEG features. Moreover, the surgical resection of areas with a larger increase in d-EI was observed for all nine patients with Engel class ⩽1, but not for the 4 of 12 patients with Engel class >1, demonstrating the significant association with seizure outcomes.Significance.We derived an iEEG feature from the trained Epi-Net, which identified the epileptic seizures with improved accuracy and might contribute to identification of the epileptogenic zone.

Case Report: Multimodal Functional and Structural Evaluation Combining Pre-operative nTMS Mapping and Neuroimaging With Intraoperative CT-Scan and Brain Shift Correction for Brain Tumor Surgical Resection.

Background: Maximum safe resection of infiltrative brain tumors in eloquent area is the primary objective in surgical neuro-oncology. This goal can be achieved with direct electrical stimulation (DES) to perform a functional mapping of the brain in patients awake intraoperatively. When awake surgery is not possible, we propose a pipeline procedure that combines advanced techniques aiming at performing a dissection that respects the anatomo-functional connectivity of the peritumoral region. This procedure can benefit from intraoperative monitoring with computerized tomography scan (iCT-scan) and brain shift correction. Associated with this intraoperative monitoring, the additional value of preoperative investigation combining brain mapping by navigated transcranial magnetic stimulation (nTMS) with various neuroimaging modalities (tractography and resting state functional MRI) has not yet been reported. Case Report: A 42-year-old left-handed man had increased intracranial pressure (IICP), left hand muscle deficit, and dysarthria, related to an infiltrative tumor of the right frontal lobe with large mass effect and circumscribed contrast enhancement in motor and premotor cortical areas. Spectroscopy profile and intratumoral calcifications on CT-scan suggested an WHO grade III glioma, later confirmed by histology. The aforementioned surgical procedure was considered, since standard awake surgery was not appropriate for this patient. In preoperative time, nTMS mapping of motor function (deltoid, first interosseous, and tibialis anterior muscles) was performed, combined with magnetic resonance imaging (MRI)-based tractography reconstruction of 6 neural tracts (arcuate, corticospinal, inferior fronto-occipital, uncinate and superior and inferior longitudinal fasciculi) and resting-state functional MRI connectivity (rs-fMRI) of sensorimotor and language networks. In intraoperative time, DES mapping was performed with motor evoked response recording and tumor resection was optimized using non-rigid image transformation of the preoperative data (nTMS, tractography, and rs-fMRI) to iCT data. Image guidance was updated with correction for brain shift and tissue deformation using biomechanical modeling taking into account brain elastic properties. This correction was done at crucial surgical steps, i.e., when tumor bulged through the craniotomy after dura mater opening and when approaching the presumed eloquent brain regions. This procedure allowed a total resection of the tumor region with contrast enhancement as well as a complete regression of IICP and dysarthria. Hand paresis remained stable with no additional deficit. Postoperative nTMS mapping confirmed the good functional outcome. Conclusion: This case report and technical note highlights the value of preoperative functional evaluation by nTMS updated intraoperatively with correction of brain deformation by iCT. This multimodal approach may become the optimized technique of reference for patients with brain tumors in eloquent areas that are unsuitable for awake brain surgery.

Somatosensation Evoked by Cortical Surface Stimulation of the Human Primary Somatosensory Cortex.

Electrical stimulation of the primary somatosensory cortex using intracranial electrodes is crucial for the evocation of artificial somatosensations, typically tactile sensations associated with specific regions of the body, in brain-machine interface (BMI) applications. The qualitative characteristics of these artificially evoked somatosensations has been well documented. As of yet, however, the quantitative aspects of these evoked somatosensations, that is to say the quantitative relationship between intensity of electrical stimulation and perceived intensity of the resultant somatosensation remains obscure. This study aimed to explore this quantitative relationship by surface electrical stimulation of the primary somatosensory cortex in two human participants undergoing electrocorticographic monitoring prior to surgical treatment of intractable epilepsy. Electrocorticogram electrodes on the primary somatosensory cortical surface were stimulated with varying current intensities, and a visual analogue scale was employed to provide a quantitative measure of intensity of the evoked sensations. Evoked sensations included those of the thumb, tongue, and hand. A clear linear relationship between current intensity and perceived intensity of sensation was observed. These findings provide novel insight into the quantitative nature of primary somatosensory cortex electrical stimulation-evoked sensation for development of somatosensory neuroprosthetics for clinical use.

Detection of Epileptic Seizures Using Phase-Amplitude Coupling in Intracranial Electroencephalography.

Seizure detection using intracranial electroencephalography (iEEG) contributes to improved treatment of patients with refractory epilepsy. For that purpose, a feature of iEEG to characterize the ictal state with high specificity and sensitivity is necessary. We evaluated the use of phase-amplitude coupling (PAC) of iEEG signals over a period of 24 h to detect the ictal and interictal states. PAC was estimated by using a synchronisation index (SI) for iEEG signals from seven patients with refractory temporal lobe epilepsy. iEEG signals of the ictal state was characterised by a strong PAC between the phase of ß and the amplitude of high γ. Furthermore, using SI values, the ictal state was successfully detected with significantly higher accuracy than by using the amplitude of high γ alone. In conclusion, PAC accurately distinguished the ictal state from the interictal state.