2026-10-11T16:58:13Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34035608 2026-09-30T17:36:04Z category_146 category_15 category_21 category_734 category_931 category_64 portal_5 item_type_1 month_year_09_2026

<p>Hypnograms over 8 hours comparing human expert annotations to our model predictions.</p> Amir Ali Vakili (25145354) Salar Jahanshiri (25145357) Armin Salimi-Badr (25145360) Space Science Neuroscience Biotechnology Biological Sciences not elsewhere classified Information Systems not elsewhere classified Cancer work proposes temposleep widely available signal weighted loss functions scale feature extraction lacking clearly defined hierarchical sequence learning enhancing contextual representation scale temporal encoding proposed framework achieves automatic sleep staging address data imbalance challenging n1 stage aware temporal modeling temporal modeling aware framework n1 stage sleep disorders data augmentation field modeling unified multi substantial improvement study focuses stacked layers sleepedf datasets results indicate range dependencies providing insights previous methods prediction behavior potential application particular emphasis overall accuracy limited receptive insufficient interpretability improving detection healthcare due global prevalence final predictions epoch chunks eeg signals eeg ), critical task class imbalance channel electroencephalography black boxes 46 %. <p>The comparison of sleep stage sequences annotated by a human expert with those predicted by the proposed model across a complete overnight recording from a single subject in the SleepEDF-20 dataset.</p> 2026-09-30T17:35:55Z Image Figure 10.1371/journal.pone.0358241.g007 https://figshare.com/articles/figure/_p_Hypnograms_over_8_hours_comparing_human_expert_annotations_to_our_model_predictions_p_/34035608 CC BY 4.0