# HeartLLM: Discretized ECG Tokenization for LLM-Based Diagnostic Reasoning

arXiv: https://arxiv.org/abs/2508.15338 · HTML: https://mlsanity.com/paper/2508.15338/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 2 of its 42 references that are in the map.
- First posted on arXiv 21 Aug 2025.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)

## Builds on (most cited first, 2 of 2)

- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Estimating or Propagating Gradients Through Stochastic Neurons](https://mlsanity.com/paper/1305.2982/) (2013)

## Built on by (most cited first, 3 of 3)

- [Large language models in medical time series analysis](https://mlsanity.com/paper/2609.22262/) (2026)
- [ShapeLex: Decoupling Local Shape Symbolization and Global Scale Modeling for Text-Controlled Time Series Generation](https://mlsanity.com/paper/2609.24003/) (2026)
- [MedTokenBudget: Lesion-Preserving Token Routing for Dermoscopic Image Classification](https://mlsanity.com/paper/2609.30613/) (2026)

## Abstract (arXiv)

Electrocardiography (ECG) plays a central role in cardiovascular diagnostics, yet existing automated approaches often struggle to generalize across clinical tasks and offer limited support for open-ended reasoning. We present HeartLLM, a novel framework that integrates time-series (TS) and language modeling by enabling large language models (LLMs) to process 12-lead ECG signals for clinical text generation tasks. Our approach discretizes continuous ECG embeddings into quantized codes using a lead-wise encoder and quantization module. These quantized codes are then mapped to an extended ECG vocabulary to form ECG tokens, enabling the model to process both ECG and natural language inputs within a unified framework. To bridge the modality gap, we pretrain the model on an autoregressive ECG token forecasting task, allowing the LLM to capture temporal dynamics through its inherent language modeling capability. Finally, we perform instruction tuning on both ECG question answering and diagnostic report generation. Without modifying the core model, HeartLLM achieves strong performance across tasks while maintaining generalization to out-of-distribution settings. Extensive experiments demonstrate the effectiveness of each component and highlight the potential of integrating discretized ECG tokens into LLMs for medical reasoning.