Brief

Researchers have developed UniECG, a novel unified model designed for interactive electrocardiogram (ECG) education. This model can generate evidence-based explanations for given ECG signals or images and, conversely, create corresponding ECG signals based on textual learning objectives. UniECG employs a two-stage design, first learning grounded ECG explanations from a dataset of ECG signals, images, and text, and then incorporating special ECG generation tokens aligned with a text-conditioned diffusion model for controllable signal generation. The system is intended as an educational aid to enhance case-based learning and interactive AI-assisted ECG education, rather than a clinical diagnostic tool. AI

Researchers have developed ECG-R1, a new multimodal large language model (MLLM) specifically designed for reliable electrocardiogram (ECG) interpretation. The model incorporates protocol-guided instruction generation, a modality-decoupled architecture to handle missing data, and reinforcement learning with diagnostic evidence rewards. Evaluations indicate that many existing MLLMs, including proprietary and open-source versions, suffer from widespread hallucinations when interpreting ECGs, underscoring the need for independent verification of their outputs. AI