Frequency band coupling with high-frequency activities during seizures shifts from θ band in tonic phase to δ band in clonic phase

Objective: To clarify variations in the relationship between high-frequency activities (HFAs) and low frequency bands from the tonic to the clonic phase in focal to bilateral tonic-clonic seizures (FBTCS), using phase-amplitude coupling. Methods: This retrospective study enrolled six patients with drug-resistant focal epilepsy who underwent intracranial electrode placement for presurgical invasive electroencephalography at Osaka University Hospital (July 2018–July 2019). We used intracranial electrodes to record seizures in focal epilepsy (11 FBTCS). The magnitude of synchronization index (SIm) and receiver operating characteristic (ROC) analysis were used to analyze the coupling between HFA amplitude (80–250 Hz) and lower frequencies phase. Results: The θ (4–8 Hz)-HFA SIm peaked in the tonic phase, whereas the δ (2–4 Hz)-HFA SIm peaked in the clonic phase. ROC analysis indicated that the δ-HFA SIm discriminated well the clonic from the tonic phase. Conclusions: The main low–frequency band modulating the HFA shifted from the θ band in the tonic phase to the δ band in the clonic phase. Significance: In FBTCS, low-frequency band coupling with HFA amplitude varies temporally. Especially, the δ band is specific to the clonic phase. These results suggest dynamically neurophysiological changes in the thalamus or basal ganglia throughout FBTCS.

Minimal phrase composition revealed by intracranial recordings

The ability to comprehend meaningful phrases is an essential component of language. Here we evaluate a minimal compositional scheme - the 'red-boat' paradigm - using intracranial recordings to map the process of semantic composition in phrase structure comprehension. 18 human participants, implanted with penetrating depth or surface subdural intracranial electrode for the evaluation of medically refractory epilepsy, were presented with auditory recordings of adjective-noun, pseudoword-noun and adjective-pseudoword phrases before being presented with a colored drawing, and were asked to judge whether the phrase matched the object presented. Significantly greater broadband gamma activity (70-150Hz) occurred in temporo-occipital junction (TOJ) and posterior middle temporal gyrus (pMTG) for pseudowords over words (300-700ms post-onset) in both first- and second-word positions. Greater inter-trial phase coherence (8-12Hz) was found for words than for pseudowords in posterior superior temporal gyrus (pSTG). Isolating phrase structure sensitivity, we identified a portion of TOJ and posterior superior temporal sulcus (pSTS) that showed increased gamma activity for phrase composition than for non-composition, while left anterior temporal lobe (ATL) showed greater low frequency (2-15Hz) activity for phrase composition, likely coordinating distributed semantic representations. Greater functional connectivity between pSTS-TOJ and pars triangularis, and between pSTS-TOJ and ATL, was also found for phrase composition. STG, ATL and pars triangularis were found to encode anticipation of composition in the beta band (15-30Hz), and alpha (8-12Hz) power increases in ATL were also linked to anticipation. These results indicate that pSTS-TOJ appears to be crucial hub in the network responsible for the retrieval and computation of minimal phrases, and that anticipation of such composition is encoded in fronto-temporal regions.

Local and Distant responses to single pulse electrical stimulation reflect different forms of connectivity

BackgroundMeasuring connectivity in the human brain can involve innumerable approaches using both noninvasive (fMRI, EEG) and invasive (intracranial EEG or iEEG) recording modalities, including the use of external probing stimuli, such as direct electrical stimulation.Objective/HypothesisTo examine how different measures of connectivity correlate with one another, we compared ‘passive’ measures of connectivity during resting state conditions map to the more ‘active’ probing measures of connectivity with single pulse electrical stimulation (SPES).MethodsWe measured the network engagement and spread of the cortico-cortico evoked potential (CCEP) induced by SPES at 53 total sites across the brain, including cortical and subcortical regions, in patients with intractable epilepsy (N=11) who were undergoing intracranial recordings as a part of their clinical care for identifying seizure onset zones. We compared the CCEP network to functional, effective, and structural measures of connectivity during a resting state in each patient. Functional and effective connectivity measures included correlation or Granger causality measures applied to stereoEEG (sEEGs) recordings. Structural connectivity was derived from diffusion tensor imaging (DTI) acquired before intracranial electrode implant and monitoring (N=8).ResultsThe CCEP network was most similar to the resting state voltage correlation network in channels near to the stimulation location. In contrast, the distant CCEP network was most similar to the DTI network. Other connectivity measures were not as similar to the CCEP network.ConclusionsThese results demonstrate that different connectivity measures, including those derived from active stimulation-based probing, measure different, complementary aspects of regional interrelationships in the brain.

A Swallowing Decoder Based on Deep Transfer Learning: AlexNet Classification of the Intracranial Electrocorticogram

To realize a brain–machine interface to assist swallowing, neural signal decoding is indispensable. Eight participants with temporal-lobe intracranial electrode implants for epilepsy were asked to swallow during electrocorticogram (ECoG) recording. Raw ECoG signals or certain frequency bands of the ECoG power were converted into images whose vertical axis was electrode number and whose horizontal axis was time in milliseconds, which were used as training data. These data were classified with four labels (Rest, Mouth open, Water injection, and Swallowing). Deep transfer learning was carried out using AlexNet, and power in the high-[Formula: see text] band (75–150[Formula: see text]Hz) was the training set. Accuracy reached 74.01%, sensitivity reached 82.51%, and specificity reached 95.38%. However, using the raw ECoG signals, the accuracy obtained was 76.95%, comparable to that of the high-[Formula: see text] power. We demonstrated that a version of AlexNet pre-trained with visually meaningful images can be used for transfer learning of visually meaningless images made up of ECoG signals. Moreover, we could achieve high decoding accuracy using the raw ECoG signals, allowing us to dispense with the conventional extraction of high-[Formula: see text] power. Thus, the images derived from the raw ECoG signals were equivalent to those derived from the high-[Formula: see text] band for transfer deep learning.

Visualization of Phase-Amplitude Coupling Using Rhythmic High-Frequency Activity

ObjectiveHigh-frequency activities (HFAs) and phase-amplitude coupling (PAC) are gaining attention as key neurophysiological biomarkers for studying human epilepsy. We aimed to clarify and visualize how HFAs are modulated by the phase of low-frequency bands during seizures.MethodsWe used intracranial electrodes to record seizures of symptomatic focal epilepsy (15 seizures in seven patients). Ripples (80–250 Hz), as representative of HFAs, were evaluated along with PAC. The synchronization index (SI), representing PAC, was used to analyze the coupling between the amplitude of ripples and the phase of lower frequencies. We created a video in which the intracranial electrode contacts were represented by circles that were scaled linearly to the power changes of ripple.ResultsThe main low frequency band modulating ictal-ripple activities was the θ band (4–8 Hz), and after completion of ictal-ripple burst, δ (1–4 Hz)-ripple PAC occurred. The video showed that fluctuation of the diameter of these circles indicated the rhythmic changes during significant high values of θ-ripple PAC.ConclusionsWe inferred that ripple activities occurring during seizure evolution were modulated by θ rhythm. In addition, we concluded that rhythmic circles’ fluctuation presented in the video represents the PAC phenomenon. Our video is thus a useful tool for understanding how ripple activity is modulated by the low-frequency phase in relation with PAC.

Utility of depth electrode placement in the neurosurgical management of bottom-of-sulcus lesions: technical note

OBJECTIVESmall lesions at the depth of the sulcus, such as with bottom-of-sulcus focal cortical dysplasia, are not visible from the surface of the brain and can therefore be technically challenging to resect. In this technical note, the authors describe their method of using depth electrodes as landmarks for the subsequent resection of these exacting lesions.METHODSA retrospective review was performed on pediatric patients who had undergone invasive electroencephalography with depth electrodes that were subsequently used as guides for resection in the period between July 2015 and June 2017.RESULTSTen patients (3–15 years old) met the criteria for this study. At the same time as invasive subdural grid and/or strip insertion, between 2 and 4 depth electrodes were placed using a hand-held frameless neuronavigation technique. Of the total 28 depth electrodes inserted, all were found within the targeted locations on postoperative imaging. There was 1 patient in whom an asymptomatic subarachnoid hemorrhage was demonstrated on postprocedural imaging. Depth electrodes aided in target identification in all 10 cases.CONCLUSIONSDepth electrodes placed at the time of invasive intracranial electrode implantation can be used to help localize, target, and resect primary zones of epileptogenesis caused by bottom-of-sulcus lesions.

Tackling Big Data in Human Intracranial Electroencephalography Recordings

INTRODUCTION Intracranial electrode recordings (ECoG [electrocorticography], sEEG [stereoelectroencephalography], and iEEG [intracranial electroencephalography]) are increasingly common across neurosurgery. Intracranial recordings during epilepsy monitoring, awake craniotomy, and deep brain stimulation are revolutionizing our understanding of basic brain function, providing access to direct recordings from human neurons. However analyzing this data is daunting. Hundreds of electrodes, each recording thousands of measurements each second, record for hours, days, or weeks. The resulting datasets easily reach hundreds of gigabytes in size, with trillions of datapoints representing a scale of data difficult to tackle. Presently, most labs write custom in-house code to analyze these datasets, making peer review of results next to impossible and exacerbating a reproducibility crisis. We present a software package, RAVE (R Analysis and Visualization of iEEG), equipped with dimension reduction techniques to allow users to more easily analyze giant iEEG datasets. METHODS RAVE is written in R using the Shiny package, enabling it to run from any web browser. Data is notch-filtered then wavelet transformed to obtain power and phase components. Both common average and bipolar re-referencing are supported. Data is then temporally down-sampled, maintaining high frequency information but vastly reduces the size of datasets. Power and phase information is then interactively displayed across single or multiple electrodes. RESULTS RAVE is a freely available software package with a growing user base, available at https://github.com/beauchamplab/rave/tree/master#rave. Inclusion of a preprocessing pipeline, signal decomposition, dimension reduction, and graphical display of iEEG data achieves our goals of standardizing data analysis, allowing users with limited programming background to analyze iEEG data, and increased transparency of published results. CONCLUSION We have developed a dimension reduction, analysis, and visualization pipeline in RAVE that allows users to take large complex datasets and easily create publication quality images in a rigorous, transparent, and easily shareable way.