Sem Hoogteijling , Eline V. Schaft , Evi H.M. Dirks , Sven Straumann , Matteo Demuru , Pieter van Eijsden , Tineke Gebbink , Willem M. Otte , Geertjan M. Huiskamp , Maryse A. van ’t Klooster , Maeike Zijlmans , the RESPect database group
{"title":"Machine learning for (non–)epileptic tissue detection from the intraoperative electrocorticogram","authors":"Sem Hoogteijling , Eline V. Schaft , Evi H.M. Dirks , Sven Straumann , Matteo Demuru , Pieter van Eijsden , Tineke Gebbink , Willem M. Otte , Geertjan M. Huiskamp , Maryse A. van ’t Klooster , Maeike Zijlmans , the RESPect database group","doi":"10.1016/j.clinph.2024.08.012","DOIUrl":null,"url":null,"abstract":"<div><h3>Objective</h3><p>Clinical visual intraoperative electrocorticography (ioECoG) reading intends to localize epileptic tissue and improve epilepsy surgery outcome. We aimed to understand whether machine learning (ML) could complement ioECoG reading, how subgroups affected performance, and which ioECoG features were most important.</p></div><div><h3>Methods</h3><p>We included 91 ioECoG-guided epilepsy surgery patients with Engel 1A outcome. We allocated 71 training and 20 test set patients. We trained an extra trees classifier (ETC) with 14 spectral features to classify ioECoG channels as covering resected or non-resected tissue. We compared the ETC’s performance with clinical ioECoG reading and assessed whether patient subgroups affected performance. Explainable artificial intelligence (xAI) unveiled the most important ioECoG features learnt by the ETC.</p></div><div><h3>Results</h3><p>The ETC outperformed clinical reading in five test set patients, was inferior in six, and both were inconclusive in nine. The ETC performed best in the tumor subgroup (area under ROC curve: 0.84 [95%CI 0.79–0.89]). xAI revealed predictors of resected (relative theta, alpha, and fast ripple power) and non-resected tissue (relative beta and gamma power).</p></div><div><h3>Conclusions</h3><p>Combinations of subtle spectral ioECoG changes, imperceptible by the human eye, can aid healthy and pathological tissue discrimination.</p></div><div><h3>Significance</h3><p>ML with spectral ioECoG features can support, rather than replace, clinical ioECoG reading, particularly in tumors.</p></div>","PeriodicalId":10671,"journal":{"name":"Clinical Neurophysiology","volume":"167 ","pages":"Pages 14-25"},"PeriodicalIF":3.7000,"publicationDate":"2024-08-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Clinical Neurophysiology","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1388245724002396","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CLINICAL NEUROLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Objective
Clinical visual intraoperative electrocorticography (ioECoG) reading intends to localize epileptic tissue and improve epilepsy surgery outcome. We aimed to understand whether machine learning (ML) could complement ioECoG reading, how subgroups affected performance, and which ioECoG features were most important.
Methods
We included 91 ioECoG-guided epilepsy surgery patients with Engel 1A outcome. We allocated 71 training and 20 test set patients. We trained an extra trees classifier (ETC) with 14 spectral features to classify ioECoG channels as covering resected or non-resected tissue. We compared the ETC’s performance with clinical ioECoG reading and assessed whether patient subgroups affected performance. Explainable artificial intelligence (xAI) unveiled the most important ioECoG features learnt by the ETC.
Results
The ETC outperformed clinical reading in five test set patients, was inferior in six, and both were inconclusive in nine. The ETC performed best in the tumor subgroup (area under ROC curve: 0.84 [95%CI 0.79–0.89]). xAI revealed predictors of resected (relative theta, alpha, and fast ripple power) and non-resected tissue (relative beta and gamma power).
Conclusions
Combinations of subtle spectral ioECoG changes, imperceptible by the human eye, can aid healthy and pathological tissue discrimination.
Significance
ML with spectral ioECoG features can support, rather than replace, clinical ioECoG reading, particularly in tumors.
期刊介绍:
As of January 1999, The journal Electroencephalography and Clinical Neurophysiology, and its two sections Electromyography and Motor Control and Evoked Potentials have amalgamated to become this journal - Clinical Neurophysiology.
Clinical Neurophysiology is the official journal of the International Federation of Clinical Neurophysiology, the Brazilian Society of Clinical Neurophysiology, the Czech Society of Clinical Neurophysiology, the Italian Clinical Neurophysiology Society and the International Society of Intraoperative Neurophysiology.The journal is dedicated to fostering research and disseminating information on all aspects of both normal and abnormal functioning of the nervous system. The key aim of the publication is to disseminate scholarly reports on the pathophysiology underlying diseases of the central and peripheral nervous system of human patients. Clinical trials that use neurophysiological measures to document change are encouraged, as are manuscripts reporting data on integrated neuroimaging of central nervous function including, but not limited to, functional MRI, MEG, EEG, PET and other neuroimaging modalities.