Sep 2026· Health Science Inquiry· Vol 15, pp. 104-112· 0 citations· 64 references
TL;DR
MetaCAM is discussed, a novel XAI framework that integrates multiple techniques to produce more stable, robust and clinically meaningful visual explanations and underscores that future progress in medical and sports‑medicine AI depends not only on predictive accuracy but also on transparent reasoning that clinicians, practitioners and end‑users can trust.
Abstract
The rapid growth of multimodal health and performance data offers unprecedented opportunities to apply machine learning and personalize care in sports‑medicine settings. However, as machine‑learning models become increasingly complex, their decision processes often become difficult to interpret, limiting trust, safe deployment and clinical adoption. This paper provides an overview of explainable artificial intelligence (XAI) methods relevant to medical and sports‑medicine applications, highlighting the unique challenges posed by high‑dimensional, heterogeneous, and noise‑prone data. We examine feature‑attribution techniques for tabular inputs, intrinsically interpretable and attention‑based models for structured and unstructured data, and visual explanation methods for medical imaging, emphasizing both their strengths and limitations in real‑world use. We further discuss emerging trends in medical XAI, including hybrid interpretable deep learning systems, interpretability tools for large multimodal foundation models and approaches such as sparse autoencoders. As a specific example, we discuss MetaCAM, our novel XAI framework that integrates multiple techniques to produce more stable, robust and clinically meaningful visual explanations. By demonstrating the benefit of consistent, reproducible and domain‑aligned interpretability, this work underscores that future progress in medical and sports‑medicine AI depends not only on predictive accuracy but also on transparent reasoning that clinicians, practitioners and end‑users can trust.
The past decade has seen machine learning play an increasingly important role in medicine. However, many of the state-of-the-art models being developed and deployed suffer from a lack of transparency, limiting integration into the clinical decision-making process and raising concerns about patient safety. Here we call...
Michael Ferguson, Fei-Fan Liu, Hong-Huang Lin et al.· Communications Medicine· 0 citations
Deep learning (DL) has shown great promise for clinical decision support, particularly for classifying medical images, but its limited transparency may limit clinicians’ understanding and trust. Alternative explainable AI (XAI) techniques include attention mechanisms, Grad-CAM, SHAP and LIME, all of which provide compl...
James Onyilo, A. C. Ikegwu, F. Uloko· Nature Journal of Emerging S...· 0 citations
In the field of medicine, ML and deep learning models offer excellent diagnostic results, though their lack of transparency creates obstacles in the area of safety, ethics, and regulations. XAI tackles this transparency issue by explaining how the intelligent systems operate. The paper provides a systematic overview of...
Pratiksha Gaddamwar, Suhani Gawate, Yash Pardhi et al.· 2026 10th International Conf...· 0 citations
Machine and deep learning algorithms have been proven their efficiency in the medical domain. This is particularly noticeable in medical imaging, where these algorithms have demonstrated exceptional capabilities, often surpassing human experts in detecting diseases, segmenting anatomical structures, and in aiding diagn...
A. Hamrouni, Thuraya N. I. Alrumaih, Mahmoud Ltaeif et al.· Journal of Visualized Experi...· 0 citations
T-MoXAI (Temporal Multimodal eXplainable AI) provides a baseline for temporal multimodal XAI, addressing fragmentation in the field and supporting applications where understanding decisions is as important as predictive accuracy.
This study proposes an explainable machine learning pipeline that integrates feature extraction techniques, namely Sparse PCA, PLS, and MCR-ALS, with Shapley Additive exPlanations (SHAP) for post hoc interpretation, while preserving the ability to project explanations back into the original input space so that practiti...
Ming-Xing Zhang· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.