{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,14]],"date-time":"2026-11-14T09:40:08Z","timestamp":1794649208987,"version":"4.9.1"},"reference-count":55,"publisher":"IOP Publishing","issue":"1","license":[{"start":{"date-parts":[[2023,1,18]],"date-time":"2023-01-18T00:00:00Z","timestamp":1674000000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2023,1,18]],"date-time":"2023-01-18T00:00:00Z","timestamp":1674000000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["J. Neural Eng."],"published-print":{"date-parts":[[2023,2,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>\n                  <jats:italic>Objective.<\/jats:italic> Epilepsy is one of the most common neurological disorders and can have a devastating effect on a person\u2019s quality of life. As such, the search for markers which indicate an upcoming seizure is a critically important area of research which would allow either on-demand treatment or early warning for people suffering with these disorders. There is a growing body of work which uses machine learning methods to detect pre-seizure biomarkers from electroencephalography (EEG), however the high prediction rates published do not translate into the clinical setting. Our objective is to investigate a potential reason for this. <jats:italic>Approach.<\/jats:italic> We conduct an empirical study of a commonly used data labelling method for EEG seizure prediction which relies on labelling small windows of EEG data in temporal groups then selecting randomly from those windows to validate results. We investigate a confound for this approach for seizure prediction and demonstrate the ease at which it can be inadvertently learned by a machine learning system. <jats:italic>Main results.<\/jats:italic> We find that non-seizure signals can create decision surfaces for machine learning approaches which can result in false high prediction accuracy on validation datasets. We prove this by training an artificial neural network to learn fake seizures (fully decoupled from biology) in real EEG. <jats:italic>Significance.<\/jats:italic> The significance of our findings is that many existing works may be reporting results based on this confound and that future work should adhere to stricter requirements in mitigating this confound. The problematic, but commonly accepted approach in the literature for seizure prediction labelling is potentially preventing real advances in developing solutions for these sufferers. By adhering to the guidelines in this paper future work in machine learning seizure prediction is more likely to be clinically relevant.<\/jats:p>","DOI":"10.1088\/1741-2552\/acae09","type":"journal-article","created":{"date-parts":[[2022,12,22]],"date-time":"2022-12-22T22:30:46Z","timestamp":1671748246000},"page":"016008","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":26,"title":["Machine learning seizure prediction: one problematic but accepted practice"],"prefix":"10.1088","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1676-2758","authenticated-orcid":true,"given":"Joseph","family":"West","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1595-3934","authenticated-orcid":false,"given":"Zahra","family":"Dasht Bozorgi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9813-0094","authenticated-orcid":true,"given":"Jeffrey","family":"Herron","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1612-7420","authenticated-orcid":false,"given":"Howard J","family":"Chizeck","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6760-2079","authenticated-orcid":false,"given":"Jordan D","family":"Chambers","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5840-8645","authenticated-orcid":false,"given":"Lyra","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"266","published-online":{"date-parts":[[2023,1,18]]},"reference":[{"key":"jneacae09bib1","doi-asserted-by":"publisher","DOI":"10.1101\/cshperspect.a022426","article-title":"Seizures and epilepsy: an overview for neuroscientists","volume":"5","author":"Stafstrom","year":"2015","journal-title":"Cold Spring Harbor Perspect. Med."},{"key":"jneacae09bib2","author":"Varsavsky","year":"2011"},{"key":"jneacae09bib3","doi-asserted-by":"publisher","first-page":"425","DOI":"10.1586\/14737175.2015.1025382","article-title":"Non-invasive EEG evaluation in epilepsy diagnosis","volume":"15","author":"Rosenow","year":"2015","journal-title":"Expert Rev. Neurother."},{"key":"jneacae09bib4","doi-asserted-by":"publisher","first-page":"101","DOI":"10.1097\/WNP.0000000000000601","article-title":"Ambulatory EEG usefulness in epilepsy management","volume":"38","author":"Hasan","year":"2021","journal-title":"J. Clin. Neurophysiol."},{"key":"jneacae09bib5","doi-asserted-by":"publisher","first-page":"563","DOI":"10.1016\/S1474-4422(13)70075-9","article-title":"Prediction of seizure likelihood with a long-term, implanted seizure advisory system in patients with drug-resistant epilepsy: a first-in-man study","volume":"12","author":"Cook","year":"2013","journal-title":"Lancet Neurol."},{"key":"jneacae09bib6","doi-asserted-by":"publisher","DOI":"10.3389\/fneur.2021.713794","article-title":"Seizure forecasting using a novel sub-scalp ultra-long term EEG monitoring system","volume":"12","author":"Stirling","year":"2021","journal-title":"Front. Neurol."},{"key":"jneacae09bib7","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1007\/978-94-017-8914-1_1","article-title":"How can we identify ictal and interictal abnormal activity?","volume":"813","author":"Fisher","year":"2014","journal-title":"Adv. Exp. Med. Biol."},{"key":"jneacae09bib8","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-020-78784-3","article-title":"Automatic seizure detection based on imaged-EEG signals through fully convolutional networks","volume":"10","author":"G\u00f3mez","year":"2020","journal-title":"Sci. Rep."},{"key":"jneacae09bib9","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s11760-022-02318-9","article-title":"Automated epilepsy seizure detection from EEG signal based on hybrid CNN and LSTM model","author":"Pandey","year":"2022","journal-title":"Signal Image Video Process."},{"key":"jneacae09bib10","doi-asserted-by":"publisher","first-page":"111","DOI":"10.1186\/s12911-018-0693-8","article-title":"Automatic seizure detection using three-dimensional CNN based on multi-channel EEG","volume":"18","author":"Wei","year":"2018","journal-title":"BMC Med. Inform. Decis. Mak."},{"key":"jneacae09bib11","doi-asserted-by":"publisher","first-page":"314","DOI":"10.1093\/brain\/awl241","article-title":"Seizure prediction: the long and winding road","volume":"130","author":"Mormann","year":"2007","journal-title":"Brain"},{"key":"jneacae09bib12","doi-asserted-by":"publisher","first-page":"618","DOI":"10.1038\/s41582-018-0055-2","article-title":"Seizure prediction\u2014ready for a new era","volume":"14","author":"Kuhlmann","year":"2018","journal-title":"Nat. Rev. Neurol."},{"key":"jneacae09bib13","doi-asserted-by":"publisher","first-page":"2172","DOI":"10.1038\/s41467-020-15908-3","article-title":"Critical slowing down as a biomarker for seizure susceptibility","volume":"11","author":"Maturana","year":"2020","journal-title":"Nat. Commun."},{"key":"jneacae09bib14","doi-asserted-by":"publisher","first-page":"267","DOI":"10.1038\/s41582-021-00464-1","article-title":"Cycles in epilepsy","volume":"17","author":"Karoly","year":"2021","journal-title":"Nat. Rev. Neurol."},{"key":"jneacae09bib15","doi-asserted-by":"publisher","first-page":"39998","DOI":"10.1109\/ACCESS.2020.2976866","article-title":"Epileptic seizures prediction using deep learning techniques","volume":"8","author":"Usman","year":"2020","journal-title":"IEEE Access"},{"key":"jneacae09bib16","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2022.967116","article-title":"Efficient graph convolutional networks for seizure prediction using scalp EEG","volume":"16","author":"Jia","year":"2022","journal-title":"Front. Neurosci."},{"key":"jneacae09bib17","doi-asserted-by":"publisher","first-page":"272","DOI":"10.3345\/cep.2021.00766","article-title":"Big data analysis and artificial intelligence in epilepsy\u2014common data model analysis and machine learning-based seizure detection and forecasting","volume":"65","author":"Chung","year":"2021","journal-title":"Clin. Exp. Pediatrics"},{"key":"jneacae09bib18","first-page":"pp 266","article-title":"An end-to-end deep learning approach for epileptic seizure prediction","author":"Yankun","year":"2020"},{"key":"jneacae09bib19","doi-asserted-by":"publisher","first-page":"4420","DOI":"10.1038\/s41598-022-08322-w","article-title":"Interpretable EEG seizure prediction using a multiobjective evolutionary algorithm","volume":"12","author":"Pinto","year":"2022","journal-title":"Sci. Rep."},{"key":"jneacae09bib20","article-title":"C2SP-Net: joint compression and classification network for epilepsy seizure prediction","author":"Wu","year":"2022"},{"key":"jneacae09bib21","first-page":"pp 1","article-title":"Towards memristive deep learning systems for real-time mobile epileptic seizure prediction","author":"Lammie","year":"2021"},{"key":"jneacae09bib22","first-page":"pp 1","article-title":"Binary single-dimensional convolutional neural network for seizure prediction","author":"Zhao","year":"2020"},{"key":"jneacae09bib23","author":"Abu-Mostafa","year":"2012"},{"key":"jneacae09bib24","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2022.108508","article-title":"EEG-based classification of normal and seizure types using relaxed local neighbour difference pattern and artificial neural network","volume":"249","author":"Sairamya","year":"2022","journal-title":"Knowl.-Based Syst."},{"key":"jneacae09bib25","doi-asserted-by":"publisher","first-page":"1927","DOI":"10.1016\/j.clinph.2009.09.002","article-title":"Classification of patterns of EEG synchronization for seizure prediction","volume":"120","author":"Mirowski","year":"2009","journal-title":"Clin. Neurophysiol."},{"key":"jneacae09bib26","doi-asserted-by":"publisher","first-page":"61","DOI":"10.1016\/j.eswa.2018.04.021","article-title":"An automated system for epilepsy detection using EEG brain signals based on deep learning approach","volume":"107","author":"Ullah","year":"2018","journal-title":"Expert Syst. Appl."},{"key":"jneacae09bib27","doi-asserted-by":"publisher","DOI":"10.3389\/fnins.2021.760987","article-title":"Automatic seizure classification based on domain-invariant deep representation of EEG","volume":"15","author":"Cao","year":"2021","journal-title":"Front. Neurosci."},{"key":"jneacae09bib28","doi-asserted-by":"publisher","DOI":"10.1016\/j.jneumeth.2019.108395","article-title":"Early prediction of epileptic seizures using a long-term recurrent convolutional network","volume":"327","author":"Wei","year":"2019","journal-title":"J. Neurosci. Methods"},{"key":"jneacae09bib29","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1016\/j.compbiomed.2018.05.019","article-title":"A long short-term memory deep learning network for the prediction of epileptic seizures using EEG signals","volume":"99","author":"Tsiouris","year":"2018","journal-title":"Comput. Biol. Med."},{"key":"jneacae09bib30","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2021.102767","article-title":"Deep learning based efficient epileptic seizure prediction with EEG channel optimization","volume":"68","author":"Jana","year":"2021","journal-title":"Biomed. Signal Process. Control"},{"key":"jneacae09bib31","doi-asserted-by":"publisher","first-page":"95","DOI":"10.3389\/fncom.2021.650050","article-title":"A deep learning approach for automatic seizure detection in children with epilepsy","volume":"15","author":"Abdelhameed","year":"2021","journal-title":"Front. Comput. Neurosci."},{"key":"jneacae09bib32","doi-asserted-by":"publisher","first-page":"3415","DOI":"10.1038\/s41598-021-82828-7","article-title":"A personalized and evolutionary algorithm for interpretable EEG epilepsy seizure prediction","volume":"11","author":"Pinto","year":"2021","journal-title":"Sci. Rep."},{"key":"jneacae09bib33","doi-asserted-by":"publisher","first-page":"2029","DOI":"10.1007\/s12652-018-0805-4","article-title":"Mean amplitude spectrum based epileptic state classification for seizure prediction using convolutional neural networks","volume":"10","author":"Guan","year":"2019","journal-title":"J. Ambient Intell. Human. Comput."},{"key":"jneacae09bib34","doi-asserted-by":"publisher","first-page":"192","DOI":"10.4103\/2228-7477.161495","article-title":"Noise estimation in electroencephalogram signal by using Volterra series coefficients","volume":"5","author":"Hassani","year":"2015","journal-title":"J. Med. Signals Sens."},{"key":"jneacae09bib35","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2021.104710","article-title":"A deep learning based ensemble learning method for epileptic seizure prediction","volume":"136","author":"Usman","year":"2021","journal-title":"Comput. Biol. Med."},{"key":"jneacae09bib36","doi-asserted-by":"publisher","first-page":"2405","DOI":"10.1007\/s40747-021-00627-z","article-title":"Two-layer LSTM network-based prediction of epileptic seizures using EEG spectral features","volume":"8","author":"Singh","year":"2022","journal-title":"Complex Intell. Syst."},{"key":"jneacae09bib37","doi-asserted-by":"publisher","first-page":"324","DOI":"10.1016\/j.cmpb.2014.02.007","article-title":"Epileptic seizure predictors based on computational intelligence techniques: a comparative study with 278 patients","volume":"114","author":"Teixeira","year":"2014","journal-title":"Comput. Methods Progr. Biomed."},{"key":"jneacae09bib38","doi-asserted-by":"publisher","first-page":"1299","DOI":"10.1016\/j.clinph.2017.04.026","article-title":"Seizure prediction in patients with focal hippocampal epilepsy","volume":"128","author":"Aarabi","year":"2017","journal-title":"Clin. Neurophysiol."},{"key":"jneacae09bib39","doi-asserted-by":"publisher","first-page":"2711","DOI":"10.1109\/TNSRE.2020.3035836","article-title":"Seizure prediction using directed transfer function and convolution neural network on intracranial EEG","volume":"28","author":"Wang","year":"2020","journal-title":"IEEE Trans. Neural Syst. Rehabil. Eng."},{"key":"jneacae09bib40","article-title":"Targeting EEG\/LFP synchrony with neural nets","volume":"vol 30","author":"Yitong","year":"2017"},{"key":"jneacae09bib41","doi-asserted-by":"publisher","first-page":"9217","DOI":"10.1109\/TPAMI.2021.3121268","article-title":"Confounds in the data\u2014Comments on \u2018Decoding brain representations by multimodal learning of neural activity and visual features\u2019","volume":"44","author":"Ahmed","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"jneacae09bib42","doi-asserted-by":"publisher","first-page":"316","DOI":"10.1109\/TPAMI.2020.2973153","article-title":"The perils and pitfalls of block design for EEG classification experiments","volume":"43","author":"Li","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"jneacae09bib43","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pcbi.1005674","article-title":"Noise correlations in the human brain and their impact on pattern classification","volume":"13","author":"Bejjanki","year":"2017","journal-title":"PLoS Comput. Biol."},{"key":"jneacae09bib44","doi-asserted-by":"publisher","first-page":"1594","DOI":"10.1038\/nn.2439","article-title":"Attention improves performance primarily by reducing interneuronal correlations","volume":"12","author":"Cohen","year":"2009","journal-title":"Nat. Neurosci."},{"key":"jneacae09bib45","doi-asserted-by":"publisher","first-page":"140","DOI":"10.1038\/370140a0","article-title":"Correlated neuronal discharge rate and its implications for psychophysical performance","volume":"370","author":"Zohary","year":"1994","journal-title":"Nature"},{"key":"jneacae09bib46","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevE.64.051904","article-title":"Population coding in neuronal systems with correlated noise","volume":"64","author":"Sompolinsky","year":"2001","journal-title":"Phys. Rev. E"},{"key":"jneacae09bib47","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1162\/089976699300016827","article-title":"The effect of correlated variability on the accuracy of a population code","volume":"11","author":"Abbott","year":"1999","journal-title":"Neural Comput."},{"key":"jneacae09bib48","doi-asserted-by":"publisher","first-page":"6010","DOI":"10.1038\/s41467-020-19784-9","article-title":"Training confounder-free deep learning models for medical applications","volume":"11","author":"Zhao","year":"2020","journal-title":"Nat. Commun."},{"key":"jneacae09bib49","doi-asserted-by":"publisher","first-page":"220","DOI":"10.1038\/nature06563","article-title":"Adaptive coding of visual information in neural populations","volume":"452","author":"Gutnisky","year":"2008","journal-title":"Nature"},{"key":"jneacae09bib50","doi-asserted-by":"publisher","first-page":"12591","DOI":"10.1523\/JNEUROSCI.2929-08.2008","article-title":"Spatial and temporal scales of neuronal correlation in primary visual cortex","volume":"28","author":"Smith","year":"2008","journal-title":"J. Neurosci."},{"key":"jneacae09bib51","article-title":"High-fidelity coding with correlated neurons","volume":"10","author":"da Silveira","year":"2014","journal-title":"PLoS Comput. Biol."},{"key":"jneacae09bib52","article-title":"Application of machine learning to epileptic seizure onset detection and treatment","author":"Shoeb","year":"2009"},{"key":"jneacae09bib53","doi-asserted-by":"publisher","first-page":"e215","DOI":"10.116\/01.CIR.101.23.e215","article-title":"PhysioBank, PhysioToolkit and PhysioNet: components of a new research resource for complex physiologic signals","volume":"101","author":"Goldberger","year":"2000","journal-title":"Circulation"},{"key":"jneacae09bib54","doi-asserted-by":"publisher","first-page":"1283","DOI":"10.3390\/en10091283","article-title":"Impacts of power grid frequency deviation on time error of synchronous electric clock and worldwide power system practices on time error correction","volume":"10","author":"Zhang","year":"2017","journal-title":"Energies"},{"key":"jneacae09bib55","doi-asserted-by":"publisher","first-page":"422","DOI":"10.1038\/s41598-020-79444-2","article-title":"Association between long-range temporal correlations in intrinsic EEG activity and subjective sense of identity","volume":"11","author":"Sugimura","year":"2021","journal-title":"Sci. Rep."}],"container-title":["Journal of Neural Engineering"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,1,18]],"date-time":"2023-01-18T14:33:10Z","timestamp":1674052390000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/1741-2552\/acae09"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,1,18]]},"references-count":55,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2023,1,18]]},"published-print":{"date-parts":[[2023,2,1]]}},"URL":"https:\/\/doi.org\/10.1088\/1741-2552\/acae09","relation":{},"ISSN":["1741-2560","1741-2552"],"issn-type":[{"value":"1741-2560","type":"print"},{"value":"1741-2552","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,1,18]]},"assertion":[{"value":"Machine learning seizure prediction: one problematic but accepted practice","name":"article_title","label":"Article Title"},{"value":"Journal of Neural Engineering","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2023 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2022-10-06","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-12-22","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2023-01-18","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}