We help you find the perfect fit.

Swiss Ai Research Overview Platform

Neuronal biomarkers of seizure activity in human epilepsy

Lay summary

Le « Human Neuron Lab » cherche à comprendre comment l’activité coordonnée des neurones du cerveau humain sous-tend les fonctions cognitives, et comment les troubles de la coordination neuronale mènent à des maladies comme l’épilepsie.

Les progrès récents des neurotechnologies ont rendu possible l’enregistrement, pendant plusieurs jours de suite, de l’activité de nombreux neurones individuels chez des patients humains, au moyen de microélectrodes implantées chirurgicalement dans le cerveau. Grâce à ce point de vue unique, nous allons identifier des séquences d’activation au sein de la population neuronale qui seront typiques des crises épileptiques de chaque patient. Cela nous permettra de développer des biomarqueurs des crises spécifiques à chaque patient, ce qui offrira alors la possibilité de détecter automatiquement les crises quand elles surviennent, ainsi que de les prévoir à l’avance, au moyen d’algorithmes automatiques.

Par ailleurs, nous allons mettre au point des expériences de neurosciences cognitives personnalisées, en nous basant sur la localisation des sites d’implantation des microélectrodes et en fonction des leurs capacités cognitives de chaque patient ; ainsi, nous révèlerons comment leurs neurones représentent les informations sensorielles et les réponses comportementales. Nous nous focaliserons en particulier sur la perception et la production de la parole et du langage. Ces expériences nous permettrons de mieux comprendre comment l’activité des neurones du cerveau est à la base des fonctions cognitives normales des êtres humains.

Abstract

Background and rationale: The unpredictable nature of seizures is a major issue for patients with epilepsy and their caregivers. Detecting seizures as soon as they start, or anticipating their occurrence, could enable on-demand treatment and significantly improve patient care. Today, however, seizure detection and prediction remain challenges without a solution. Data from animal models suggest that seizures are characterized by stereotypical sequences of activation among cortical neurons. The dynamics of neuronal activity across a population of neurons, rather than changes in the activity of single cells, might thus represent the long-elusive marker of pathologic activity during seizures. Microelectrode recordings, which give access to cerebral activity at the level of individual neurons, have recently become possible for long periods of time in humans.Overall objectives: This project aims at identifying novel biomarkers for seizure detection and prediction at the neuronal level using microelectrode recordings in patients with epilepsy. Taking inspiration from animal model data, I will investigate the degree to which human seizures are characterized by stereotypical sequences of activity among populations of cortical neurons. For Detection, I will develop metrics that summarize activity in the neuronal population during seizures, and I will test the performance of these metrics in the prospective detection of seizure onsets. For Prediction, I will assess whether measuring the activity of neuronal populations helps tackle the difficult challenge of predicting seizures, compared to the current state-of-the-art based on intracranial EEG.Methods: Additional microelectrodes will be implanted in the brain of patients considering epilepsy surgery and undergoing intracranial EEG monitoring. According to the clinical situation, and in order to sample the diversity of human epilepsy, two types of microelectrodes will be used: silicon-based microelectrode arrays implanted in the neocortex, or microwire electrodes targeting deep structures like the hippocampus and amygdala. I will acquire continuous, weeks-long recordings of neuronal activity, both during seizures and in between them. To analyze these complex datasets, I will establish an unsupervised spike-sorting pipeline to automatically identify the activity of individual neurons. I will use machine learning to develop metrics of neuronal population activity and to test the performance of these metrics in seizure detection and prediction.Expected results: Identifying stereotypical sequences of activation across neuronal populations will provide novel, patient-specific biomarkers for the detection of epilepsy at the microscopic level. Simultaneous micro- and macroelectrode recordings will reveal the time-frequency signatures of these neuronal biomarkers in the more conventional intracranial and scalp EEG.Impact: The reliable identification of neuronal biomarkers of seizures through microelectrodes will stimulate the development of medical devices that use microelectrodes to count and forecast seizures, a major unmet need in clinical research and patient care. Accurate seizure detection will significantly improve on-demand therapies like deep-brain stimulation, and could also prove important in future clinical trials of antiepileptic drugs.

  Pierre Megevand