NeuroCanvas: VLLM-Powered Robust Seizure Detection by Reformulating Multichannel EEG as Image

TL;DR

NeuroCanvas addresses the core challenge of robust, real-time seizure detection from multichannel EEG by reformulating EEG signals as dense visual canvases for vision-language models. It introduces Entropy-guided Channel Selector to pick informative channels and Canvas of Neuron Signal to convert selected channels into compact intensity maps, enabling efficient VLLM-based classification. Across TUSZ and CHB-MIT, NeuroCanvas delivers significant improvements in binary F1 and precision, with substantial reductions in inference latency and token usage, demonstrating practical potential for real-time clinical deployment. The framework leverages prompt-guided adaptation and demonstrates robust performance even under reduced channel availability, underscoring its scalability and applicability to heterogeneous EEG data.

Abstract

Accurate and timely seizure detection from Electroencephalography (EEG) is critical for clinical intervention, yet manual review of long-term recordings is labor-intensive. Recent efforts to encode EEG signals into large language models (LLMs) show promise in handling neural signals across diverse patients, but two significant challenges remain: (1) multi-channel heterogeneity, as seizure-relevant information varies substantially across EEG channels, and (2) computing inefficiency, as the EEG signals need to be encoded into a massive number of tokens for the prediction. To address these issues, we draw the EEG signal and propose the novel NeuroCanvas framework. Specifically, NeuroCanvas consists of two modules: (i) The Entropy-guided Channel Selector (ECS) selects the seizure-relevant channels input to LLM and (ii) the following Canvas of Neuron Signal (CNS) converts selected multi-channel heterogeneous EEG signals into structured visual representations. The ECS module alleviates the multi-channel heterogeneity issue, and the CNS uses compact visual tokens to represent the EEG signals that improve the computing efficiency. We evaluate NeuroCanvas across multiple seizure detection datasets, demonstrating a significant improvement of $20\%$ in F1 score and reductions of $88\%$ in inference latency. These results highlight NeuroCanvas as a scalable and effective solution for real-time and resource-efficient seizure detection in clinical practice.The code will be released at https://github.com/Yanchen30247/seizure_detect.

Figures (7)

• Figure 1: Overview of the NeuroCanvas framework.
• Figure 2: Attention heatmap on CNS intensity map and direct time series visualization figure.
• Figure 3: (a) Layer-wise Pearson correlation coefficients between "CNS" and Direct Time-series. (b) Comparison of average attention scores across 19 EEG channels.
• Figure 4: Number of input tokens across EEG representation pipelines. Compared with numeric and direct visualization inputs, CNS achieves up to a 95.56% reduction in input tokens.
• Figure 5: Why CNS helps or fails, the attention distribution analysis. The "TP" denotes the true positive, and the "FN" means the false positive. The blue represents the Direct Visualization, and the red denotes CNS.