{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T06:04:55Z","timestamp":1790661895270,"version":"4.1.0"},"reference-count":33,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T00:00:00Z","timestamp":1790640000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neuroinform."],"abstract":"<jats:p>\n                    Windowed electroencephalography (EEG) pipelines can produce misleadingly high performance when segments from the same participant occur in both training and testing. We define this contamination as\n                    <jats:italic>window leakage<\/jats:italic>\n                    and test it in two clinically and technically distinct public cohorts: 121 children (61 attention-deficit\/hyperactivity disorder [ADHD], 60 controls) and 149 adults (100 Parkinson's disease, 49 controls). Signals were harmonized to 18 common channels and 128 Hz. All primary models used stratified 8-fold participant-disjoint evaluation, 50%-overlapping windows, participant-level probability aggregation, and validation participants drawn exclusively from each outer-training fold. A deliberately contaminated Random Forest audit contrasted random-window and participant-disjoint splitting. Random-window balanced accuracy reached 95.59% in ADHD and 92.02% in Parkinson's, but fell to 60.98% and 56.17%, respectively, when participants were held out. Thus, the inflation replicated across disorders, ages, recording tasks, and acquisition systems. Under correct evaluation, CNN-LSTM achieved the highest ADHD subject-level balanced accuracy (73.21%), whereas EEGNet achieved the highest Parkinson's balanced accuracy (63.75%). No architecture comparison remained significant after Bonferroni correction. One-, 2-, 4-, and 8-s windows produced no consistent cross-cohort ranking. Removing ATCNN-SA components also yielded dataset-dependent changes, and no ablation difference was significant after correction. Participant-bootstrap intervals remained wide despite tens of thousands of windows, indicating that inferential precision was governed by participant count rather than segment count. Window-level random splits contaminate clinical EEG performance estimates and must not be used to claim generalization to unseen individuals. Participants should be separated before model fitting, results should be reported at subject level, and split, windowing, aggregation, and uncertainty procedures should be disclosed explicitly.\n                  <\/jats:p>","DOI":"10.3389\/fninf.2026.1930795","type":"journal-article","created":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:46:30Z","timestamp":1790660790000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Window leakage in clinical EEG machine learning: matched subject-independent evaluation across ADHD and Parkinson cohorts"],"prefix":"10.3389","volume":"20","author":[{"given":"Burak","family":"Aggul","sequence":"first","affiliation":[{"name":"Department of Database, Network Design and Management, Plato Vocational School, Istanbul Topkapi University","place":["Istanbul, T\u00fcrkiye"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,29]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"102227","DOI":"10.1016\/j.bspc.2020.102227","article-title":"Computer aided diagnosis system using deep convolutional neural networks for ADHD subtypes","volume":"63","author":"Ahmadi","year":"2021","journal-title":"Biomed. Sig. Process. Control"},{"key":"B2","first-page":"1","article-title":"\u201cNeural machine translation by jointly learning to align and translate,\u201d","volume-title":"Proceedings of International Conference on Learning Representations (ICLR)","author":"Bahdanau","year":"2015"},{"key":"B3","doi-asserted-by":"publisher","first-page":"95","DOI":"10.3390\/brainsci15010095","article-title":"EEG-based ADHD classification using autoencoder feature extraction and ResNet with double augmented attention mechanism","volume":"15","author":"Bansal","year":"2025","journal-title":"Brain Sci"},{"key":"B4","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach. Learn"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1373515","DOI":"10.3389\/fnins.2024.1373515","article-title":"Data leakage in deep learning studies of translational EEG","volume":"18","author":"Brookshire","year":"2024","journal-title":"Front. Neurosci"},{"key":"B6","doi-asserted-by":"publisher","first-page":"107153","DOI":"10.1016\/j.bspc.2024.107153","article-title":"Topological feature search method for multichannel EEG: application in ADHD classification","volume":"100","author":"Cai","year":"2025","journal-title":"Biomed. Sig. Process. Control"},{"key":"B7","doi-asserted-by":"publisher","first-page":"2098","DOI":"10.1016\/S1388-2457(01)00668-X","article-title":"EEG-defined subtypes of children with attention-deficit\/hyperactivity disorder","volume":"112","author":"Clarke","year":"2001","journal-title":"Clin. Neurophysiol"},{"key":"B8","doi-asserted-by":"publisher","first-page":"045055","DOI":"10.1088\/2057-1976\/ae8fdb","article-title":"Unmasking data leakage in EEG-ADHD literature: a rigorous, interpretable SOTA framework (DSAEN)","volume":"12","author":"Das","year":"2026","journal-title":"Biomed. Phys. Eng. Exp"},{"key":"B9","doi-asserted-by":"publisher","first-page":"975","DOI":"10.1007\/s11571-021-09680-3","article-title":"Direction of information flow between brain regions in ADHD and healthy children based on EEG by using directed phase transfer entropy","volume":"15","author":"Ekhlasi","year":"2021","journal-title":"Cogn. Neurodyn"},{"key":"B10","doi-asserted-by":"publisher","first-page":"109305","DOI":"10.1016\/j.compbiomed.2024.109305","article-title":"Refining ADHD diagnosis with EEG: the impact of preprocessing and temporal segmentation on classification accuracy","volume":"183","author":"Garc\u00eda-Ponsoda","year":"2024","journal-title":"Comput. Biol. Med"},{"key":"B11","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1007\/s13755-024-00305-7","article-title":"Convolutional neural network framework for EEG-based ADHD diagnosis in children","volume":"12","author":"Hassan","year":"2024","journal-title":"Health Inf. Sci. Syst"},{"key":"B12","doi-asserted-by":"publisher","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput"},{"key":"B13","doi-asserted-by":"publisher","first-page":"7132","DOI":"10.1109\/CVPR.2018.00745","article-title":"\u201cSqueeze-and-excitation networks,\u201d","author":"Hu","year":"2018","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B14","doi-asserted-by":"publisher","first-page":"100804","DOI":"10.1016\/j.patter.2023.100804","article-title":"Leakage and the reproducibility crisis in machine-learning-based science","volume":"4","author":"Kapoor","year":"2023","journal-title":"Patterns"},{"key":"B15","first-page":"1","article-title":"\u201cAdam: a method for stochastic optimization,\u201d","volume-title":"Proceeedings of International Conference on Learning Representations","author":"Kingma","year":"2015"},{"key":"B16","doi-asserted-by":"publisher","first-page":"653659","DOI":"10.3389\/fnhum.2021.653659","article-title":"BENDR: Using transformers and a contrastive self-supervised learning task to learn from massive amounts of EEG data","volume":"15","author":"Kostas","year":"2021","journal-title":"Front. Hum. Neurosci"},{"key":"B17","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/EMBC40787.2023.10340700","article-title":"\u201cMachine learning-based detection of Parkinson's disease from resting-state EEG: a multi-center study,\u201d","author":"Kurbatskaya","year":"2023","journal-title":"Proceedings of 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)"},{"key":"B18","doi-asserted-by":"publisher","first-page":"109092","DOI":"10.1016\/j.compbiomed.2024.109092","article-title":"Siamese based deep neural network for ADHD detection using EEG signal","volume":"182","author":"Latifi","year":"2024","journal-title":"Comput. Biol. Med"},{"key":"B19","doi-asserted-by":"publisher","first-page":"056013","DOI":"10.1088\/1741-2552\/aace8c","article-title":"EEGNet: a compact convolutional neural network for EEG-based brain-computer interfaces","volume":"15","author":"Lawhern","year":"2018","journal-title":"J. Neural Eng"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1561994","DOI":"10.3389\/fnins.2025.1561994","article-title":"ADHD detection from EEG signals using GCN based on multi-domain features","volume":"19","author":"Li","year":"2025","journal-title":"Front. Neurosci"},{"key":"B21","doi-asserted-by":"publisher","first-page":"66","DOI":"10.1007\/s13534-016-0218-2","article-title":"EEG classification of ADHD and normal children using non-linear features and neural network","volume":"6","author":"Mohammadi","year":"2016","journal-title":"Biomed. Eng. Lett"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.21227\/rzfh-zn36","article-title":"EEG Data for ADHD\/Control Children","author":"Nasrabadi","year":"2020","journal-title":"IEEE DataPort."},{"key":"B23","first-page":"2825","article-title":"Scikit-learn: Machine learning in python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res"},{"key":"B24","doi-asserted-by":"publisher","first-page":"5391","DOI":"10.1002\/hbm.23730","article-title":"Deep learning with convolutional neural networks for eeg decoding and visualization","volume":"38","author":"Schirrmeister","year":"2017","journal-title":"Hum. Brain Map"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.82901\/nemar.on004584","author":"Singh","year":"2026","journal-title":"Rest eyes open (Version v1.0.0)"},{"key":"B26","doi-asserted-by":"publisher","first-page":"945","DOI":"10.1136\/jnnp-2022-330154","article-title":"Evoked mid-frontal activity predicts cognitive dysfunction in Parkinson's disease","volume":"94","author":"Singh","year":"2023","journal-title":"J. Neurol. Neurosurg. Psych"},{"key":"B27","doi-asserted-by":"publisher","first-page":"440","DOI":"10.1097\/01.wnp.0000221363.12503.78","article-title":"A meta-analysis of quantitative EEG power associated with attention-deficit hyperactivity disorder","volume":"23","author":"Snyder","year":"2006","journal-title":"J. Clin. Neurophysiol"},{"key":"B28","doi-asserted-by":"publisher","first-page":"710","DOI":"10.1109\/TNSRE.2022.3230250","article-title":"EEG conformer: convolutional transformer for EEG decoding and visualization","volume":"31","author":"Song","year":"2023","journal-title":"IEEE Trans. Neural Syst. Rehabil. Eng"},{"key":"B29","doi-asserted-by":"publisher","first-page":"101352","DOI":"10.1016\/j.imu.2023.101352","article-title":"Generalizable electroencephalographic classification of Parkinson's disease using deep learning","volume":"42","author":"Sugden","year":"2023","journal-title":"Informat. Med. Unlock"},{"key":"B30","doi-asserted-by":"publisher","first-page":"91","DOI":"10.1186\/1471-2105-7-91","article-title":"Bias in error estimation when using cross-validation for model selection","volume":"7","author":"Varma","year":"2006","journal-title":"BMC Bioinf"},{"key":"B31","doi-asserted-by":"publisher","first-page":"68","DOI":"10.1016\/j.neuroimage.2017.06.061","article-title":"Cross-validation failure: small sample sizes lead to large error bars","volume":"180","author":"Varoquaux","year":"2018","journal-title":"NeuroImage"},{"key":"B32","doi-asserted-by":"publisher","first-page":"e20192528","DOI":"10.1542\/peds.2019-2528","article-title":"Clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit\/hyperactivity disorder in children and adolescents","volume":"144","author":"Wolraich","year":"2019","journal-title":"Pediatrics"},{"key":"B33","doi-asserted-by":"publisher","first-page":"1703","DOI":"10.1609\/aaai.v32i1.11496","article-title":"Cascade and parallel convolutional recurrent neural networks on eeg-based intention recognition for brain computer interface","volume":"32","author":"Zhang","year":"2018","journal-title":"Proc. AAAI Conf. Artif. Intell"}],"container-title":["Frontiers in Neuroinformatics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2026.1930795\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,29]],"date-time":"2026-09-29T05:46:32Z","timestamp":1790660792000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fninf.2026.1930795\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,29]]},"references-count":33,"alternative-id":["10.3389\/fninf.2026.1930795"],"URL":"https:\/\/doi.org\/10.3389\/fninf.2026.1930795","relation":{},"ISSN":["1662-5196"],"issn-type":[{"value":"1662-5196","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,29]]},"article-number":"1930795"}}