This research presents a real-time AI-based seizure detection model for pediatric epilepsy patients using edge devices. The embedded system enhances safety through immediate alerts, supporting clinical staff in continuous monitoring with minimal latency and optimized performance for sensitive patient environments.

Caroline Ann Fletcher, Benjamin George Wainwright, Sophie Eleanor Chadwick, Henry Thomas Colville, Charlotte Abigail Bentley

This study proposes a federated learning architecture tailored for distributed biomedical imaging diagnostics. It ensures data privacy across institutions while allowing collaborative model training, preserving sensitive health information and accelerating shared intelligence in medical AI applications without centralizing patient data.

Julian Robert Ashford, Matilda Frances Prescott, George William Linwood, Florence Isabelle Gifford, Tobias Alexander Maynard

This paper presents a deep learning-based digital therapeutic system that decodes facial micro-expressions for emotional state detection. The system aids in behavioral health monitoring, providing real-time psychological interventions by recognizing subtle emotional cues captured through standard front-facing camera sensors.

Lucinda Evelyn Marlowe, Douglas Harry Pennington, Amelia Georgina Langley, Oliver Thomas Rowntree, Hannah Grace Telford

This study introduces a hybrid deep learning model that combines transformers and residual CNNs for detecting anomalies in cardiovascular time-series data. The architecture improves early identification of irregular heart rhythms, offering superior interpretability and clinical value in digital cardiology tools.

Frederick Michael Ashcroft, Georgina Rose Ellington, Jack Alexander Melville, Emily Jane Carrington, Nathan Charles Rowe

This work presents a personalized cancer prognosis model using graph neural networks with attention layers to integrate multi-omics data. The framework enables robust tumor subtype differentiation, offering precision insights that support individualized treatment planning and improved clinical outcomes in oncology research.

Isabelle Florence Radcliffe, Christopher Dean Hepworth, Megan Olivia Hargreaves, Samuel Thomas Winslow, Beatrice Hannah Forsyth

This study introduces a cross-modality deep learning model integrating genomic profiles with clinical imaging data to improve rare disease diagnosis. The model enhances early detection, supporting multi-source interpretability in biomedical informatics with applications in personalized and precision medicine frameworks.

Florence Abigail Whittaker, Thomas Edward Caldwell, Rupert Jonathan Mayfield, Emily Grace Sutherland, Owen Maxwell Rowden

This paper presents a blockchain-based framework for securely sharing biomedical research data across multiple cloud environments. The architecture ensures transparency, immutability, and access control, promoting collaborative research without compromising patient data privacy or the integrity of decentralized scientific workflows.

Gabriel Henry Fenwick, Imogen Sophia Blackwell, Edward James Loxley, Matilda Grace Marchant, Alexander George Redfern