Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Feb 24, 2026

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

No Black Box Anymore: Demystifying Clinical Predictive Modeling with Temporal-Feature Cross Attention Mechanism.

Yubo Li¹, Xinyu Yao¹, Rema Padman¹

¹Carnegie Mellon University, Pittsburgh, PA, USA.

AMIA ... Annual Symposium Proceedings. AMIA Symposium

|February 23, 2026

Summary

We developed a new deep learning method, Temporal-Feature Cross Attention Mechanism (TFCAM), to improve clinical prediction and explainability. TFCAM accurately predicts Chronic Kidney Disease progression, offering transparent insights for clinicians.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Adaptive Robust EKF with NARX-Based Velocity Prediction for High Precision AUV Navigation Under DVL Outages.

Sensors (Basel, Switzerland)·2026

Same author

Structural insights into the binding of CX-5461 and MTR-106 to G-quadruplex DNA.

Nucleic acids research·2026

Same author

Dependencies of ecological drought on climatic drivers in diverse ecosystems: Implications for water resources management.

Journal of environmental management·2026

Same author

Giant Topological Hall Effect Across a Broad Temperature Window in Co-Doped Mn<sub>3</sub>Sn Noncoplanar Antiferromagnets.

Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2026

Same author

Reply to: Reassessing the evidence linking clinical leadership to AI deployment outcomes.

NPJ digital medicine·2026

Same author

A simple LC-MS/MS method for the simultaneous determination of 18 free fatty acids in human serum.

Journal of chromatography. B, Analytical technologies in the biomedical and life sciences·2026

Area of Science:

* Artificial Intelligence in Medicine
* Clinical Informatics
* Biomedical Data Science

Background:

* Deep learning models excel in clinical prediction but often lack transparency.
* Explainability is crucial for clinical adoption and trust in AI-driven healthcare.
* Existing methods struggle to capture complex temporal dynamics of disease progression.

Purpose of the Study:

* Introduce the Temporal-Feature Cross Attention Mechanism (TFCAM) for enhanced clinical prediction.
* Improve interpretability of deep learning models in healthcare.
* Capture dynamic feature interactions over time for better disease progression prediction.

Main Methods:

* Developed TFCAM, a novel deep learning framework inspired by transformer architectures.

More Related Videos

Cross-Modal Multivariate Pattern Analysis

Cross-Modal Multivariate Pattern Analysis

Published on: November 9, 2011

Related Experiment Videos

Last Updated: Feb 24, 2026

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Measuring Attention and Visual Processing Speed by Model-based Analysis of Temporal-order Judgments

Published on: January 23, 2017

Cross-Modal Multivariate Pattern Analysis

Cross-Modal Multivariate Pattern Analysis

Published on: November 9, 2011

* Applied TFCAM to predict End-Stage Renal Disease progression in 1,422 Chronic Kidney Disease patients.

* Compared TFCAM against LSTM and RE-TAIN baselines.

Main Results:

* TFCAM achieved superior predictive performance with an AUROC of 0.95 and F1-score of 0.69.
* The model outperformed established LSTM and RE-TAIN methods.
* TFCAM provided multi-level explainability, identifying critical time periods and feature importance.

Conclusions:

* TFCAM effectively addresses the 'black box' problem in clinical deep learning.
* The framework offers clinicians transparent insights into disease progression.
* TFCAM enhances predictive accuracy while providing interpretable results for healthcare applications.