Category Ranking
98%
Total Visits
921
Avg Visit Duration
2 minutes
Citations
20
Article Abstract
Objective: Our objective is to compare different pre-training strategies for electronic health record (EHR) foundation models.
Materials And Methods: We evaluated three approaches using a transformer-based architecture: baseline (no pre-training), self-supervised pre-training with masked language modeling, and supervised pre-training. The models were assessed on their ability to predict both major adverse cardiac events and mortality occurring within 12 months. The pre-training cohort was 405 679 patients prescribed antihypertensives and the fine tuning cohort was 5525 patients who received doxorubicin.
Results: Task-specific supervised pre-training achieved superior performance (AUROC 0.70, AUPRC 0.23), outperforming both self-supervised pre-training and the baseline. However, when the model was evaluated on the task of 12-month mortality prediction, the self-supervised model performed best.
Discussion: While supervised pre-training excels when aligned with downstream tasks, self-supervised approaches offer more generalized utility.
Conclusion: Pre-training strategy selection should consider intended applications, data availability, and transferability requirements.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC12349770 | PMC |
| http://dx.doi.org/10.1093/jamiaopen/ooaf090 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Self-supervised representation learning with continuous training data improves the feel and performance of myoelectric control.
Comput Biol Med
September 2025
Department of Electrical and Computer Engineering and the Institute of Biomedical Engineering, University of New Brunswick, Fredericton, E3B 5A3, NB, Canada.
Pattern recognition-based myoelectric control is traditionally trained with static or ramp contractions, but this fails to capture the dynamic nature of real-world movements. This study investigated the benefits of training classifiers with continuous dynamic data, encompassing transitions between various movement classes. We employed both conventional (LDA) and deep learning (LSTM) classifiers, comparing their performance when trained with ramp data, continuous dynamic data, and an LSTM pre-trained with a self-supervised learning technique (VICReg).
View Article and Find Full Text PDFEpilepsyFM: A domain-specific foundation model for epileptic representation learning using EEG signals.
Neural Netw
September 2025
School of Automation, Northwestern Polytechnical University, Xi'an, 710072, Shaanxi, China. Electronic address:
Epilepsy with its complex seizure mechanisms and diverse clinical manifestations, presents numerous challenges for clinical diagnosis and treatment, while electroencephalography (EEG) plays a crucial and irreplaceable role in its diagnosis. Although general-purpose foundation models have demonstrated some capability in knowledge processing, they still face challenges in capturing specific disease features and dealing with data scarcity in highly specialized domains such as epilepsy. To address these issues, we propose a domain-specific foundation model for epilepsy-EpilepsyFM, designed to learn generalized representations of epilepsy to support various downstream tasks.
View Article and Find Full Text PDFCorrection: Multi-label remote sensing classification with self-supervised gated multi-modal transformers.
Front Comput Neurosci
August 2025
Origin Dynamics Intelligent Robot Co., Ltd., Zhengzhou, China.
[This corrects the article DOI: 10.3389/fncom.2024.
View Article and Find Full Text PDFEnhancing employee well-being through a culturally adapted training program: a mixed-methods study in South Africa.
Front Public Health
September 2025
Department of Industrial Psychology and People Management, College of Business and Economics, University of Johannesburg, Johannesburg, South Africa.
Introduction: Structured, well-being interventions are under-researched in non-Western workplaces. This study evaluates training program-a participatory, multi-component training intervention-on employee well-being, engagement and stress in South Africa.
Methods: Employing an exploratory, quasi-experimental, explanatory sequential mixed-methods design, we collected quantitative data from 50 South African respondents across three delivery formats (four half-days online, two full-days in classroom, and four half-days in classroom) at pre-training and 3 months post-training using five validated scales (PSS-4, UWES-3, SWLS, FS, WEMWBS-14).
Development and validation of a deep learning model for early detection and screening of diabetic retinopathy.
BMC Med Inform Decis Mak
August 2025
Ningbo Municipal Hospital of TCM, Affiliated Hospital of Zhejiang Chinese Medical University, Ningbo, P.R. China.
Early diagnosis and screening of diabetic retinopathy (DR) are crucial for reducing medical burdens and conserving healthcare resources. This study introduces an advanced AI-assisted recognition system designed to enhance the detection of DR lesions through innovative automatic learning methods. Central to our approach are agnostic text instruction templates, which facilitate zero-shot DR detection by integrating text embeddings with visual information.
View Article and Find Full Text PDF