Automatic bad channel detection in intracranial electroencephalographic recordings using ensemble machine learning

Objective

Intracranial electroencephalographic (iEEG) recordings contain “bad channels”, which show non-neuronal signals. Here, we developed a new method that automatically detects iEEG bad channels using machine learning of seven signal features.

Functional organization of the human primary somatosensory cortex: A stereo-electroencephalography study

ObjectiveThe classical homunculus of the human primary somatosensory cortex (S1) established by Penfield has mainly portrayed the functional organization of convexial cortex, namely Brodmann area (BA) 1. However, little is known about the functions in fissural cortex including BA2 and BA3. We aim at drawing a refined and detailed somatosensory homunculus of the entire S1.MethodsWe recruited 20 patients with drug-resistant focal epilepsy who underwent stereo-electroencephalography for preoperative assessments. Direct electrical stimulation was performed for functional mapping. Montreal Neurological Institute coordinates of the stimulation sites lying in S1 were acquired.ResultsStimulation of 177 sites in S1 yielded 149 positive sites (84%), most of which were located in the sulcal cortex. The spatial distribution of different body-part representations across the S1 surface revealed that the gross medial-to-lateral sequence of body representations within the entire S1 was consistent with the classical “homunculus”. And we identified several unreported body-part representations from the sulcal cortex, such as forehead, deep elbow and wrist joints, and some dorsal body regions.ConclusionsOur results reveal general somatotopical characteristics of the entire S1 cortex and differences with the previous works of Penfield.SignificanceThe classical S1 homunculus was extended by providing further refinement and additional detail.     

Automated electrical source imaging with scalp EEG to define the insular irritative zone: Comparison with simultaneous intracranial EEG

Objective To evaluate the accuracy of automated interictal low-density electrical source imaging (LD-ESI) to define the insular irritative zone (IZ) by comparing the simultaneous interictal ESI localization with the SEEG interictal activity.Methods Long-term simultaneous scalp electroencephalography (EEG) and stereo-EEG (SEEG) with at least one depth electrode exploring the operculo-insular region(s) were analyzed. Automated interictal ESI was performed on the scalp EEG using standardized low-resolution brain electromagnetic tomography (sLORETA) and individual head models. A two-step analysis was performed: i) sublobar concordance between cluster-based ESI localization and SEEG-based IZ; ii) time-locked ESI-/SEEG analysis. Diagnostic accuracy values were calculated using SEEG as reference standard. Subgroup analysis was carried out, based on the involvement of insular contacts in the seizure onset and patterns of insular interictal activity.Results Thirty patients were included in the study. ESI showed an overall accuracy of 53% (C.I. 29–76%). Sensitivity and specificity were calculated as 53% (C.I. 29–76%), 55% (C.I. 23–83%) respectively. Higher accuracy was found in patients with frequent and dominant interictal insular spikes.Conclusions LD-ESI defines with good accuracy the insular implication in the IZ, which is not possible with classical interictal scalp EEG interpretation.SignificanceAutomated LD-ESI may be a valuable additional tool to characterize the epileptogenic zone in epilepsies with suspected insular involvement.     

Increased discharge threshold after an interictal spike in human focal epilepsy

It is commonly assumed that interictal spikes (ISs) in focal epilepsies set off a period of inhibition that transiently reduces tissue excitability. Post-spike inhibition was described in experimental models but was never demonstrated in the human epileptic cortex. In the present study post-spike excitability was retrospectively evaluated on intracerebral stereo-electroencephalographic recordings performed in the epileptogenic cortex of five patients suffering from drug-resistant focal epilepsy secondary to Taylor-type neocortical dysplasias. Patients typically presented with highly periodic interictal spiking activity at 2.33 +/- 0.87 Hz (mean +/- SD) in the dysplastic region. During the stereo-electroencephalographic procedure, low-frequency stimulation at 1 Hz was systematically performed for diagnostic purposes to identify the epileptogenic zone. The probability of evoking an IS during the interspike period in response to 1-Hz stimuli delivered close to the ictal-onset zone was examined. Stimuli that occurred early after a spontaneous IS (within 70% of the inter-IS period) had a very low probability of generating a further IS. On the contrary, stimuli delivered during the late inter-IS period had the highest probability of evoking a further IS. The generation of stimulus-evoked ISs is occluded for several hundred milliseconds after the occurrence of a preceding spike discharge. As previously shown in animal models, these findings suggest that, during focal, periodic interictal spiking, human neocortical excitability is phasically controlled by post-spike inhibition.     

立体定位脑电图长程监测在顽固性癫(痫)致(痫)灶定位中的应用

目的:探讨立体定向技术植入的深部电极构成的立体定位脑电图对药物难治性癫(痫)致(痫)灶的定位阳性率和近期手术效果.方法:对于头皮脑电、CT、MRI难以确定致(痫)灶位置的药物难治性癫(痫)的10例患者,采用立体定向方法植入深部电极2根6例、3根4例,长程记录发作3次以上,结合发作间期和发作期深部电极脑电确定致(痫)灶,外科治疗后随访术后发作改善情况.结果:外科术前监测106例,实行手术33例,其中10例患者需要植入深部电极辅助定位.此10例中无植入电极并发症,均有发作间期局灶(痫)样放电和发作期局灶性起源位置,其中局灶性起源7例、区域性起源3例.参照深部电极脑电图结果采取外科手术治疗,术后9～20个月,癫(痫)缓解4例,2例好转90％以上,4例好转50％以上,无手术并发症.结论:立体定位脑电图可对头皮脑电难以确定致(痫)灶的患者进行定位,为外科提供定位性诊断信息,提高外科手术效果.     

Four-dimensional tractography animates propagations of neural activation via distinct interhemispheric pathways

ObjectiveTo visualize and validate the dynamics of interhemispheric neural propagations induced by single-pulse electrical stimulation (SPES).MethodsThis methodological study included three patients with drug-resistant focal epilepsy who underwent measurement of cortico-cortical spectral responses (CCSRs) during bilateral stereo-electroencephalography recording. We delivered SPES to 83 electrode pairs and analyzed CCSRs recorded at 268 nonepileptic electrode sites. Diffusion-weighted imaging (DWI) tractography localized the interhemispheric white matter pathways as streamlines directly connecting two electrode sites. We localized and visualized the putative SPES-related fiber activation, at each 1-ms time window, based on the propagation velocity defined as the DWI-based streamline length divided by the early CCSR peak latency.ResultsThe resulting movie, herein referred to as four-dimensional tractography, delineated the spatiotemporal dynamics of fiber activation via the corpus callosum and anterior commissure. Longer streamline length was associated with delayed peak latency and smaller amplitude of CCSRs. The cortical regions adjacent to each fiber activation site indeed exhibited CCSRs at the same time window.ConclusionsOur four-dimensional tractography successfully animated neural propagations via distinct interhemispheric pathways.SignificanceOur novel animation method has the potential to help investigators in addressing the mechanistic significance of the interhemispheric network dynamics supporting physiological function.     

Effects of depth electrode montage and single-pulse electrical stimulation sites on neuronal responses and effective connectivity

ObjectiveTo determine the optimal depth electrode montages for the assessment of effective connectivity based on single-pulse electrical stimulation (SPES). To determine the effect of SPES locations on the extent of resulting neuronal propagations.MethodsWe studied 14 epilepsy patients who underwent invasive monitoring with depth electrodes and measurement of cortico-cortical evoked potentials (CCEPs) and cortico-cortical spectral responses (CCSRs). We determined the effects of electrode montage and stimulus sites on the CCEP/CCSR amplitudes.ResultsBipolar and Laplacian montages effectively reduced the degree of SPES-related signal deflections at extra-cortical levels, including outside the brain, while maintaining those at the cortical level. SPES of structures more proximal to the deep white matter, compared to the cortical surface, elicited greater CCEPs and CCSRs.ConclusionsOn depth electrode recording, bipolar and Laplacian montages are suitable for measurement of near-field CCEPs and CCSRs. SPES of the white matter axons may induce neuronal propagations to extensive regions of the cerebral cortex.SignificanceThis study helps to establish the practical guidelines on the diagnostic use of CCEPs/CCSRs.     

难治性局灶性癫痫的多学科团队模式诊疗策略

目的:探讨难治性局灶性癫痫的多学科团队模式(multidisciplinary team,MDT)诊疗策略.方法:介绍1例药物难治性颞叶癫痫,头皮脑电图监测示患者双侧颞区不同步放电,常规磁共振序列扫描未见明显异常,初步评估为患者病灶位置和侧向性不明.进行MDT讨论和后续进一步术前评估,明确患者致痫灶和治疗方法.结果:患...     

Modeling of intracerebral interictal epileptic discharges: Evidence for network interactions

Objective

The interictal epileptic discharges (IEDs) occurring in stereotactic EEG (SEEG) recordings are in general abundant compared to ictal discharges, but difficult to interpret due to complex underlying network interactions. A framework is developed to model these network interactions.