ViTexSZ：用于脑电图癫痫发作检测的异构视觉-文本知识蒸馏框架
ViTexSZ: Heterogeneous Vision-Text Knowledge Distillation for EEG Seizure Detection
浏览论文内容
中文总结 AI 辅助
ViTexSZ是用于EEG癫痫发作检测的异构视觉-文本知识蒸馏框架，通过多模态大语言模型对齐异构EEG与临床语义，在四个数据集上实现最高准确率，相对第二优基线提升达12.9%。
中文摘要 AI 辅助
从脑电图（EEG）中自动检测癫痫发作对于连续神经监测至关重要，尤其是针对可能仅表现出微弱电生理变化的亚临床癫痫发作。现有时间序列方法通常针对固定的EEG通道配置设计，因此限制了其在通道布局不规则的异构EEG记录中的适用性。尽管视觉和语言建模提供了有前景的替代方案，但将异构EEG表示与临床语义对齐仍然具有挑战性。我们提出ViTexSZ，这是一种用于EEG癫痫发作检测的异构视觉-文本知识蒸馏框架。ViTexSZ将EEG记录转换为结构化波形图像，并引入基于查询的多通道对齐模块，该模块将依赖于源的视觉特征映射到统一令牌空间。异构教师模型进一步通过多模态大语言模型将对齐的EEG表示与临床提示整合，将高级临床语义与癫痫发作相关证据关联起来。随后，视觉-文本知识蒸馏在检测过程中将教师模型的表示传递给轻量级学生模型。在四个EEG癫痫发作数据集上的实验证明了ViTexSZ在亚临床和普通癫痫发作检测场景中的泛化能力，在所有数据集上均达到最高准确率，相比第二优基线方法的相对提升高达12.9%，显示了其有效性。
英文摘要
Automated seizure detection from electroencephalography (EEG) is essential for continuous neurological monitoring, particularly for subclinical epileptic seizures that may exhibit only subtle electrographic changes. Existing time-series methods are often designed for fixed EEG channel configurations, thereby limiting their applicability to heterogeneous EEG recordings with irregular channel layouts. Although visual and language modeling offer promising alternatives, aligning heterogeneous EEG representations with clinical semantics remains challenging. We introduce ViTexSZ, a heterogeneous Vision-Text knowledge distillation framework for EEG seizure detection. ViTexSZ converts EEG recordings into structured waveform images and introduces a query-based multi-channel alignment module that maps source-dependent visual features into a unified token space. A heterogeneous teacher further integrates the aligned EEG representations with clinical prompts through a multimodal large language model, associating high-level clinical semantics with seizure-related evidence. Vision-text knowledge distillation then transfers the teacher representations to a lightweight student during detection. Experiments on four EEG seizure datasets demonstrate the generalizability of ViTexSZ across both subclinical and general seizure detection scenarios, achieving the highest accuracy on all datasets and relative improvements of up to 12.9% over the second-best baselines, showing its effectiveness.
