用轻量图模块让大模型高效学脑电空间特征。
Graph Adapter of EEG Foundation Models for Parameter Efficient Fine Tuning
- 在预训练时序模型上加图网络适配器，只训练新模块。
- 在抑郁和异常检测任务中，F1得分提升最高达16.1%。
- 适合数据少、算力有限的脑电医疗场景。
在从脑电图（EEG）数据诊断神经疾病时，如Transformer等基础模型可捕捉时间动态，而图神经网络（GNN）对表示电极间空间关系至关重要。然而，为同时学习时空特征而微调这些大规模模型，计算成本极高，尤其在标注数据稀缺时。本文提出EEG-GraphAdapter（EGA），一种参数高效的微调方法。EGA作为基于GNN的模块集成到预训练时序主干模型中，冻结主干网络，仅微调适配器部分，从而有效获取脑电空间表示，显著降低计算开销与数据需求。在两个医疗相关下游任务——重度抑郁症（MDD）与异常检测（TUAB）上的实验表明，相较于主干模型BENDR，EGA在F1-score上最高提升16.1%，展现出其在可扩展且高精度脑电预测中的潜力。
原文摘要 · Abstract (English)
In diagnosing neurological disorders from electroencephalography (EEG) data, foundation models such as Transformers have been employed to capture temporal dynamics. Additionally, Graph Neural Networks (GNNs) are critical for representing the spatial relationships among EEG sensors. However, fine-tuning these large-scale models for both temporal and spatial features can be prohibitively large in computational cost, especially under the limited availability of labeled EEG datasets. We propose EEG-GraphAdapter (EGA), a parameter-efficient fine-tuning (PEFT) approach designed to address these challenges. EGA is integrated into a pre-trained temporal backbone model as a GNN-based module, freezing the backbone and allowing only the adapter to be fine-tuned. This enables the effective acquisition of EEG spatial representations, significantly reducing computational overhead and data requirements. Experimental evaluations on two healthcare-related downstream tasks-Major Depressive Disorder (MDD) and Abnormality Detection (TUAB)-show that EGA improves performance by up to 16.1% in F1-score compared with the backbone BENDR model, highlighting its potential for scalable and accurate EEG-based predictions.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。