Neural Signals Generate Clinical Notes in the Wild

TL;DR

This paper introduces CELM, a novel EEG-to-language foundation model that automatically generates comprehensive clinical reports from long-term EEG recordings, improving efficiency and accuracy in clinical settings.

Contribution

CELM is the first multimodal foundation model capable of summarizing long-duration EEG data into clinical reports, integrating pretrained EEG and language models, and establishing a large-scale dataset and benchmark.

Findings

CELM outperforms existing methods across all evaluation metrics.

Human experts find CELM-generated reports more clinically coherent and reliable.

The authors release the model and benchmark pipeline for future research.

Abstract

Generating clinical reports that summarize abnormal patterns, diagnostic findings, and clinical interpretations from long-term EEG recordings remains labor-intensive. We present CELM, the first clinical EEG-to-Language foundation model capable of summarizing long-duration, variable-length EEG recordings and performing end-to-end clinical report generation at multiple scales. CELM integrates pretrained EEG foundation models with language models to enable scalable multimodal learning. We curate a large-scale clinical EEG dataset containing 9,922 reports paired with approximately 11,000 hours of EEG recordings from 9,048 patients to train CELM, and release the benchmark with an automated report-structuring pipeline to facilitate future research. Experimental results show that CELM consistently outperforms existing methods across all evaluation settings. Importantly, we further conduct human evaluation with clinical experts, demonstrating that CELM generates reports that are more clinically coherent, diagnostically reliable, and better aligned with expert interpretation. We release our model and benchmark construction pipeline at https://github.com/Jathurshan0330/CELM.