1成果简介
脑电图（EEG）是脑功能研究和脑机接口（BCI）的核心技术手段。然而，传统的湿电极需要涂抹导电凝胶，准备时间长、使用不便、且不适合长期佩戴，严重制约了EEG在日常可穿戴和高密度非侵入式脑机接口中的应用。尽管近年来干电极技术取得了长足进步，但低噪声、低电位漂移、高信号质量的干电极仍然是该领域亟需突破的瓶颈。
本文，天津理工大学李明吉教授、李红姬教授/天津大学王坤副教授、许敏鹏教授等在《ACS Applied Materials & Interfaces》期刊发表名为"Application of Graphene Dry Electrode in 512-Lead EEG Cap and Real-Time Monitoring EEG System"的论文。该研究开发了一种钠（Na）掺杂垂直石墨烯干电极（直径仅2.8 mm），并将其成功集成于512导联超高密度脑电帽及无线8导联和32导联头戴式设备中。
Na掺杂的垂直石墨烯层具有独特的三维针状结构，能够吸收头皮汗液并将其转化为Na?介导的固态电解质，从而在电极与头皮之间建立稳定的电连接。优化后的石墨烯干电极表现出低头皮接触阻抗（干燥：3.8–6.5 kΩ，加水：4.5 kΩ）、低自噪声（11.1 μV）、低直流偏置电压（15.6 mV）和低电位漂移（189.9 μV）。由512个石墨烯干电极组成的EEG脑电帽能够以高信噪比记录不同节律的脑电信号，并在103天的长期使用中展现出优异的重复性和稳定性。此外，研究还设计了一种任务态策略，结合快慢波强度比与频域事件相关电位，验证了干电极头戴系统在日常佩戴中进行快速注意力分析的可靠性。
2图文导读
图1. Schematic diagrams of Na-doped graphene dry electrodes, EEG systems, and their applications. (a) Preparation of the Na-VG electrode and electrode arrays. (b) Dry scalp-contact mechanism of the Na-VG electrode. (c) Design of the 512-lead EEG cap. (d) Application of daily wearable wireless EEG headbands and attention analysis software.
图2. Characterization of the graphene dry electrode. (a) Surface SEM images of Na-doped graphene samples with different growth times of 20 s, 40 s, and 4 min. (b) TEM, HRTEM, and EDS element mapping of Na-VG nanosheets grown for 4 min. (c) High-resolution C 1s XPS data of Na-doped samples grown for 20 s, 40 s, and 4 min. (d) High-resolution Na 1s XPS data of the Na-doped sample grown for 40 s. (e) Atomic content of various elements in the samples. (f) Raman spectra of Na-VG samples, (g) FWHM of 2D peaks and intensity ratio of 2D and G peaks (I2D/IG), and (h) intensity ratio of D and G peaks (ID/IG) and intensity ratio of D′ and G peaks (ID′/IG) of different metal-doped graphene samples.
图3. Performance optimization of graphene dry electrodes. (a, b) Amplitude-time curves of the Na-VG and Ag/AgCl electrodes, and (c) total noise (NT), self-noise (NE), potential drift (Vdrift), and DC offset voltage (Voffset) of various EEG electrodes (n = 5; mean ± SD). (d) Nyquist plots, (e) three-electrode system on the back of the hand and equivalent circuits, (f) Bode plots, and (g) various contact resistances (n = 5; mean ± SD). Ultrapure water (10 μL) was dropped onto the surface of the VG electrodes. (h) Time-domain and (i) frequency-domain spectra of signals from EEG for eyes closed/open. The Na-VG and Ag/AgCl electrodes were used synchronously, and both electrodes were positioned at the FP1 in the left frontal lobe. (j) Peak frequency of EEG-α rhythm, Rcc, and SNR for different VG electrodes (n = 5; mean ± SD). Asterisks *, ** and ns represent p < 0.05 (significant), p < 0.01 (highly significant), and no statistical significance, respectively.
图4. Reliability evaluation of dry electrode array and 512-lead EEG cap. (a) Design structure of the Na-VG electrode array, photographs of volunteers wearing the 512-lead EEG cap, and layout of the electrode arrays on the 512-lead EEG cap. (b) Rscalp, amplitude of EEG-α rhythm, and SNR of the five Na-VG electrode arrays at Fp1 and Oz positions. (c) Rscalp, amplitude, and SNR values of the eight electrodes in the array (n = 5; mean ± SD). Statistical significance was assessed using one-way ANOVA, where *, **, and ns were used to denote significance (p < 0.05), high significance (p < 0.01), and no statistical significance, respectively. (d) The Rscalp, amplitude, and SNR of the frontal lobe Fp1 position of six volunteers were recorded using an electrode array, as well as the Rscalp distribution of the 512-lead EEG cap. (e) Long-term stability of Rscalp in a 512-lead EEG cap over 103 days.
图5. Ultrahigh-density 512–lead EEG system. (a) Experimental scenario, distribution of 512 electrodes on the UHD EEG cap, and Rscalp distribution. (b) Frequency–channel–amplitude mapping distributions of δ, θ, α, β, and γ-rhythms in EEG signals recorded during the closed/open-eye paradigm. (c) Mean EEG amplitudes at 4.5, 11.75, 29, and 62 Hz. (d) Mean SNR values for θ, α, β, and γ-rhythms.
图6. Application of Na-VG nail electrode in wireless EEG headband BCI systems. (a) Schematic diagram of the electrode distribution, application scenarios, and task states of the 32- and 8-lead EEG headbands. (b) Time-domain spectra selected from 32 leads. (c) Distribution of the three EEG rhythms (θ, α, and β) in different brain regions. (d) Time-domain spectra were collected at different time intervals during the volunteer’s viewing of a ～150 s movie. The EEG signals were collected at channel 2 of Fp1(B). (e) Pβ/Pα, Pβ/Pθ, and ERP-amplitude under silent-1, sound, and silent-2 conditions. (f) Five indicator values for different brain regions (n = 5; mean ± SD). (g) Distribution of attention indicators for 50 young healthy volunteers. (h) Average values of 50 people for the five indicators (n = 50; mean ± SD). (i) Confusion matrix for binary-classification recognition. (j) Accuracy, precision, recall, and F1-score values (n = 5; mean ± SD). The data in (e), (g), and (h)–(j) were collected using 8-lead h eadbands. Statistical significance was assessed using one-way ANOVA, where *, **, and ns were used to denote significance (p < 0.05), high significance (p < 0.01), and no statistical significance, respectively.
3小结
在本研究中，开发了一套512通道超高清（UHD）脑电图（EEG）干电极系统，以及无线头带式8通道和32通道EEG采集系统，其中采用了一种新型Na-VG干电极作为通用组件，用于全脑、脑区及电极位置的EEG数据相关性分析。该 512 通道 UHD-BCI 由 512 个直径为 2.8 毫米的 Na-VG 干电极组成，每个脑区分布有 32 至 88 个电极，用于捕捉脑电图的细节。在无线 32 通道 EEG 头带中，电极分布于额叶、颞叶、顶叶和枕叶的信噪比（SNR）较高的位置，左右半球对称放置了 4 个电极阵列。首先，Na-VG 干电极体积小，自噪声和头皮接触电阻较低。因此，在干式测试中，组装好的8电极阵列和512导联脑电帽展现了出色的可重复性和长期稳定性。其次，UHD 512导联脑机接口系统的闭眼静息状态脑电信号证实了θ、α、β和γ脑波节律的高信噪比响应，呈现了初步绘制的脑电神经影像。该512导联脑电帽采用八电极阵列设计策略，不仅比单电极更便于采集有效信号，还提升了脑机接口（BCI）的日常佩戴舒适度。这款干式脑电帽是首个采用8倍脑电细节采集策略的BCI系统，可在10–10国际电极系统中覆盖64个电极位置。此外，在任务导向型应用方面，无线32导联Na-VG干电极脑电头带在视听范式中记录到了快波和慢波的正常变化以及ERP特征，而配套的前额叶8导联脑电头带在注意力评估中达到了90%的准确率。这些结果证实了这两种BCI系统在采集诱发电脑波方面的有效性。本研究提供了新型干式石墨烯电极及有效的BCI硬件设计策略，旨在提升BCI的空间分辨率并构建适用于日常使用的可穿戴硬件。随着高通道数脑电放大器的商业化日益普及，这款超高清512通道干式电极脑电帽在神经科学研究和临床实践中展现出广阔的应用前景。
文献：
