Machine Learning to Classify Relative Seizure Frequency From Chronic Electrocorticography
PURPOSE/OBJECTIVE:Brain responsive neurostimulation (NeuroPace) treats patients with refractory focal epilepsy and provides chronic electrocorticography (ECoG). We explored how machine learning algorithms applied to interictal ECoG could assess clinical response to changes in neurostimulation parameters. METHODS:We identified five responsive neurostimulation patients each with ≥200 continuous days of stable medication and detection settings (median, 358 days per patient). For each patient, interictal ECoG segments for each month were labeled as "high" or "low" to represent relatively high or low long-episode (i.e., seizure) count compared with the median monthly long-episode count. Power from six conventional frequency bands from four responsive neurostimulation channels were extracted as features. For each patient, five machine learning algorithms were trained on 80% of ECoG, then tested on the remaining 20%. Classifiers were scored by the area-under-the-receiver-operating-characteristic curve. We explored how individual circadian cycles of seizure activity could inform classifier building. RESULTS:Support vector machine or gradient boosting models achieved the best performance, ranging from 0.705 (fair) to 0.892 (excellent) across patients. High gamma power was the most important feature, tending to decrease during low-seizure-frequency epochs. For two subjects, training on ECoG recorded during the circadian ictal peak resulted in comparable model performance, despite less data used. CONCLUSIONS:Machine learning analysis on retrospective background ECoG can classify relative seizure frequency for an individual patient. High gamma power was the most informative, whereas individual circadian patterns of seizure activity can guide model building. Machine learning classifiers built on interictal ECoG may guide stimulation programming.
Rare Genetic Variation and Outcome of Surgery for Mesial Temporal Lobe Epilepsy
OBJECTIVE:Genetic factors have long been debated as a cause of failure of surgery for mesial temporal lobe epilepsy (MTLE). We investigated whether rare genetic variation influences seizure outcomes of MTLE surgery. METHODS:We performed an international, multicenter, whole exome sequencing study of patients who underwent surgery for drug-resistant, unilateral MTLE with normal magnetic resonance imaging (MRI) or MRI evidence of hippocampal sclerosis and ≥2-year postsurgical follow-up. Patients with either sustained seizure freedom (favorable outcome) or ongoing uncontrolled seizures since surgery (unfavorable outcome) were included. Exomes of controls without epilepsy were also included. Gene set burden analyses were carried out to identify genes with significant enrichment of rare deleterious variants in patients compared to controls. RESULTS:Nine centers from 3 continents contributed 206 patients operated for drug-resistant unilateral MTLE, of whom 196 (149 with favorable outcome and 47 with unfavorable outcome) were included after stringent quality control. Compared to 8,718 controls, MTLE cases carried a higher burden of ultrarare missense variants in constrained genes that are intolerant to loss-of-function (LoF) variants (odds ratio [OR] = 2.6, 95% confidence interval [CI] = 1.9-3.5, p = 1.3E-09) and in genes encoding voltage-gated cation channels (OR = 2.4, 95% CI = 1.4-3.8, p = 2.7E-04). Proportions of subjects with such variants were comparable between patients with favorable outcome and those with unfavorable outcome, with no significant between-group differences. INTERPRETATION/CONCLUSIONS:Rare variation contributes to the genetic architecture of MTLE, but does not appear to have a major role in failure of MTLE surgery. These findings can be incorporated into presurgical decision-making and counseling. ANN NEUROL 2022.
Temporal dynamics of neural responses in human visual cortex
Neural responses to visual stimuli exhibit complex temporal dynamics, including sub-additive temporal summation, response reduction with repeated or sustained stimuli (adaptation), and slower dynamics at low contrast. These phenomena are often studied independently. Here, we demonstrate these phenomena within the same experiment and model the underlying neural computations with a single computational model. We extracted time-varying responses from electrocorticographic (ECoG) recordings from patients presented with stimuli that varied in contrast, duration, and inter-stimulus interval (ISI). Aggregating data across patients from both sexes yielded 98 electrodes with robust visual responses, covering both earlier (V1-V3) and higher-order (V3a/b, LO, TO, IPS) retinotopic maps. In all regions, the temporal dynamics of neural responses exhibit several non-linear features: peak response amplitude saturates with high contrast and longer stimulus durations; the response to a second stimulus is suppressed for short ISIs and recovers for longer ISIs; response latency decreases with increasing contrast. These features are accurately captured by a computational model comprised of a small set of canonical neuronal operations: linear filtering, rectification, exponentiation, and a delayed divisive normalization. We find that an increased normalization term captures both contrast- and adaptation-related response reductions, suggesting potentially shared underlying mechanisms. We additionally demonstrate both changes and invariance in temporal response dynamics between earlier and higher-order visual areas. Together, our results reveal the presence of a wide range of temporal and contrast-dependent neuronal dynamics in the human visual cortex, and demonstrate that a simple model captures these dynamics at millisecond resolution.SIGNIFICANCE STATEMENTSensory inputs and neural responses change continuously over time. It is especially challenging to understand a system that has both dynamic inputs and outputs. Here we use a computational modeling approach that specifies computations to convert a time-varying input stimulus to a neural response time course, and use this to predict neural activity measured in the human visual cortex. We show that this computational model predicts a wide variety of complex neural response shapes that we induced experimentally by manipulating the duration, repetition and contrast of visual stimuli. By comparing data and model predictions, we uncover systematic properties of temporal dynamics of neural signals, allowing us to better understand how the brain processes dynamic sensory information.
Spatiotemporal dynamics of human high gamma discriminate naturalistic behavioral states
In analyzing the neural correlates of naturalistic and unstructured behaviors, features of neural activity that are ignored in a trial-based experimental paradigm can be more fully studied and investigated. Here, we analyze neural activity from two patients using electrocorticography (ECoG) and stereo-electroencephalography (sEEG) recordings, and reveal that multiple neural signal characteristics exist that discriminate between unstructured and naturalistic behavioral states such as "engaging in dialogue" and "using electronics". Using the high gamma amplitude as an estimate of neuronal firing rate, we demonstrate that behavioral states in a naturalistic setting are discriminable based on long-term mean shifts, variance shifts, and differences in the specific neural activity's covariance structure. Both the rapid and slow changes in high gamma band activity separate unstructured behavioral states. We also use Gaussian process factor analysis (GPFA) to show the existence of salient spatiotemporal features with variable smoothness in time. Further, we demonstrate that both temporally smooth and stochastic spatiotemporal activity can be used to differentiate unstructured behavioral states. This is the first attempt to elucidate how different neural signal features contain information about behavioral states collected outside the conventional experimental paradigm.
Impact of the COVID-19 pandemic on people with epilepsy: findings from the US arm of the COV-E study
OBJECTIVES/OBJECTIVE:As part of the COVID-19 and Epilepsy (COV-E) global study, we aimed to understand the impact of COVID-19 on the medical care and well-being of people with epilepsy (PWE) in the United States, based on their perspectives and those of their caregivers. METHODS:Separate surveys designed for PWE and their caregivers were circulated from April 2020 to July 2021; modifications in March 2021 included a question about COVID-19 vaccination status. RESULTS:We received 788 responses, 71% from PWE (nÂ =Â 559) and 29% (n=229) from caregivers of persons with epilepsy. A third (nÂ =Â 308) of respondents reported a change in their health or in the health of the person they care for. Twenty-seven percent (nÂ =Â 210) reported issues related to worsening mental health. Of respondents taking ASMs (nÂ =Â 769), 10% (n=â€‰78) reported difficulty taking medications on time, mostly due to stress causing forgetfulness. Less than half of respondents received counseling on mental health and stress. Less than half of the PWE reported having discussions with their healthcare providers about sleep, ASMs and potential side effects, while a larger proportion of caregivers (81%) reported having had discussions with their healthcare providers on the same topics. More PWE and caregivers reported that COVID-19 related measures caused adverse impact on their health in the post-vaccine period than during the pre-vaccine period, citing mental health issues as the primary reason. SIGNIFICANCE/CONCLUSIONS:Our findings indicate that the impact of the COVID-19 pandemic in the US on PWE is multifaceted. Apart from the increased risk of poor COVID-19 outcomes, the pandemic has also had negative effects on mental health and self-management. Healthcare providers must be vigilant for increased emotional distress in PWE during the pandemic and consider the importance of effective counseling to diminish risks related to exacerbated treatment gaps.
Genomics in the presurgical epilepsy evaluation
Epilepsy surgery should be considered in all patients with drug-resistant focal epilepsy. The diagnostic presurgical evaluation aims to delineate the epileptogenic zone and its relationship to eloquent brain regions. Genetic testing is not yet routine in presurgical evaluations, despite many monogenic causes of severe epilepsies, including some focal epilepsies. This review highlights genomic data that may inform decisions regarding epilepsy surgery candidacy and strategy. Focal epilepsies due to pathogenic variants in mechanistic target of rapamycin pathway genes are amenable to surgery if clinical, electroencephalography and imaging data are concordant. Epilepsy surgery outcomes are less favourable in patients with pathogenic variants in ion channel genes such as SCN1A. However, genomic data should not be used in isolation to contraindicate epilepsy surgery and should be considered alongside other diagnostic modalities. The additional role of somatic mosaicism in the pathogenesis of focal epilepsies may have implications for surgical planning and prognostication. Here, we advocate for including genomic data in the presurgical evaluation and multidisciplinary discussion for many epilepsy surgery candidates. We encourage neurologists to perform genetic testing in patients with focal non-lesional epilepsy, epilepsy in the setting of intellectual disability and epilepsy due to specific malformations of cortical development. The integration of genomics into the presurgical evaluation assists selection of patients for resective surgery and fosters a personalised medicine approach, where precision or targeted therapies are considered alongside surgical procedures.
Shared computational principles for language processing in humans and deep language models
Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models generate appropriate linguistic responses in a given context. In the current study, nine participants listened to a 30-min podcast while their brain responses were recorded using electrocorticography (ECoG). We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process the same natural narrative: (1) both are engaged in continuous next-word prediction before word onset; (2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise; (3) both rely on contextual embeddings to represent words in natural contexts. Together, our findings suggest that autoregressive DLMs provide a new and biologically feasible computational framework for studying the neural basis of language.
Genomic analysis of "microphenotypes" in epilepsy
Large international consortia examining the genomic architecture of the epilepsies focus on large diagnostic subgroupings such as "all focal epilepsy" and "all genetic generalized epilepsy". In addition, phenotypic data are generally entered into these large discovery databases in a unidirectional manner at one point in time only. However, there are many smaller phenotypic subgroupings in epilepsy, many of which may have unique genomic risk factors. Such a subgrouping or "microphenotype" may be defined as an uncommon or rare phenotype that is well recognized by epileptologists and the epilepsy community, and which may or may not be formally recognized within the International League Against Epilepsy classification system. Here we examine the genetic structure of a number of such microphenotypes and report in particular on two interesting clinical phenotypes, Jeavons syndrome and pediatric status epilepticus. Although no single gene reached exome-wide statistical significance to be associated with any of the diagnostic categories, we observe enrichment of rare damaging variants in established epilepsy genes among Landau-Kleffner patients (GRIN2A) and pediatric status epilepticus patients (MECP2, SCN1A, SCN2A, SCN8A).
Long-term priors influence visual perception through recruitment of long-range feedback
Perception results from the interplay of sensory input and prior knowledge. Despite behavioral evidence that long-term priors powerfully shape perception, the neural mechanisms underlying these interactions remain poorly understood. We obtained direct cortical recordings in neurosurgical patients as they viewed ambiguous images that elicit constant perceptual switching. We observe top-down influences from the temporal to occipital cortex, during the preferred percept that is congruent with the long-term prior. By contrast, stronger feedforward drive is observed during the non-preferred percept, consistent with a prediction error signal. A computational model based on hierarchical predictive coding and attractor networks reproduces all key experimental findings. These results suggest a pattern of large-scale information flow change underlying long-term priors' influence on perception and provide constraints on theories about long-term priors' influence on perception.
Impact of the COVID-19 pandemic on people with epilepsy: Findings from the Brazilian arm of the COV-E study
The COVID-19 pandemic has had an unprecedented impact on people and healthcare services. The disruption to chronic illnesses, such as epilepsy, may relate to several factors ranging from direct infection to secondary effects from healthcare reorganization and social distancing measures.