相关实验视频

Updated: Jan 20, 2026

Simultaneous Electroencephalography and Magnetoencephalography to Identify Seizure-Prone Brain Regions

Simultaneous Electroencephalography and Magnetoencephalography to Identify Seizure-Prone Brain Regions

Published on: May 29, 2025

开发和验证一个深度生存模型,以预测从常规脑电图中查获的时间

Émile Lemoine^1,2,3, An Qi Xu³, Mezen Jemel^1,3

¹Department of Neuroscience, Université de Montréal, Montreal, Quebec, Canada.

|January 19, 2026

概括

一个新的深度学习模型,EEGSurvNet,准确地预测了常规脑电图 (EEG) 的个人风险,长达两年. 这种先进的预测工具的性能优于传统方法,特别是在没有明显异常的情况下.

科学领域:

人工智能在医学中的应用
神经学和临床神经生理学
机器学习用于医疗保健

背景情况:

准确的发作风险预测对于的管理和治疗决策至关重要.
传统的临床预测指标,如间接性发作性泄露 (IED),在预测个人随时间推移的发作风险方面存在局限性.
常规的脑电图 (EEG) 包含了可能未充分利用的预后信息.

研究的目的:

开发和验证一个深度生存模型,EEGSurvNet,用于使用常规EEG数据预测个体发作风险.
将EEGSurvNet的性能与传统的临床预测指标和基线发作风险进行比较.
调查模型在可见形异常之外提取预后信息的能力.

主要方法:

对994名患者的1014例常规EEG进行了回顾性队列研究.
开发EEGSurvNet,这是一个深度学习模型,可以在2年时间范围内预测下一次扣押的时间.
使用依赖时间的AUROC,集成的AUROC (iAUROC) 和C指数的性能评估,与Cox模型和随机模型相比.

主要成果:

EEGSurvNet的两年iAUROC为0.69和C指数为0.66,超过了临床和随机模型的表现.
模型性能在EEG后2个月达到峰值 (AUROC=.80) 并与临床预测因素结合使用时得到改善 (iAUROC=.70).
EEGSurvNet 对于没有 IED 的 EEG 进行了优异的歧视 (iAUROC=.78 vs. .53),突出显示了椎部区域和 6-15 Hz 频率作为关键预测指标.

结论:

像EEGSurvNet这样的深度学习模型可以从常规EEG中提取重要的预后信息,即使没有明显的形异常.
EEGSurvNet显示有潜力提高患者咨询,并为管理中的治疗决策提供信息.
需要进一步的前性研究来验证这些发现,并评估EEGSurvNet.Net的临床实用性.

关键词:

人工智能的人工智能是人工智能.生物标志物生物标志物电脑脑电图 (EEG) 是一种电脑电图.预后预后预后生存分析,生存分析.

更多相关视频

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Published on: April 15, 2015

Author Spotlight: Streamlining Protein Target Prediction and Validation via Molecular Docking and CETSA

Author Spotlight: Streamlining Protein Target Prediction and Validation via Molecular Docking and CETSA

Published on: February 23, 2024

相关概念视频

Simultaneous Electroencephalography and Magnetoencephalography to Identify Seizure-Prone Brain Regions

Simultaneous Electroencephalography and Magnetoencephalography to Identify Seizure-Prone Brain Regions

Source: Papadelis, C., et al., Interictal High Frequency Oscillations Detected with Simultaneous Magnetoencephalography and Electroencephalography as Biomarker of Pediatric Epilepsy. J. Vis. Exp. (2016). This video demonstrates the method of using electroencephalography (EEG) and magnetoencephalography (MEG) recordings to identify interictal epileptiform discharges (IEDs) and high-frequency oscillations (HFOs) to locate seizure-prone regions in the...

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Stereotactic Electroencephalography (SEEG) is an operative technique used in epilepsy surgery to help localize seizure foci. It also affords a unique opportunity to investigate brain function. Here we describe how SEEG can be used to investigate cognitive processes in human...

Recording of Induced Seizures through Electroencephalography in Freely Moving Rat Pups

Recording of Induced Seizures through Electroencephalography in Freely Moving Rat Pups

The video demonstrates the recording of electroencephalogram (EEG) signals from seizure-induced rat pups. Take a freely moving pup implanted with electrodes connected to a data acquisition system. A neurotoxin is administered to induce epileptic seizures, and the resulting electrical bursting activity is monitored using the EEG...

Protein Target Prediction and Validation of Small Molecule Compound

Protein Target Prediction and Validation of Small Molecule Compound

The experiment used here shows a method of molecular docking combined with cellular thermal shift assay to predict and validate the interaction between small molecules and protein...

Implanting an On-Head Electrode System in a Rat Model for Real-Time Seizure Detection

Implanting an On-Head Electrode System in a Rat Model for Real-Time Seizure Detection

This video demonstrates the method of implanting an on-head electrode system in an anesthetized rat for real-time seizure detection. The procedure includes placing stimulation and recording electrodes, securing them with dental cement, and connecting them to a head-mounted...

Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery

Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery

The study found that male gender, poor tumor grade, and advanced Tumor Node Metastasis stage were associated with poorer cancer-specific survival (CSS) in multiple primary colorectal cancer (MPCC) patients after surgery. We developed a nomogram to predict the CSS of MPCC patients and contribute to clinical treatment...