Embroidered EEG Electrodes

• Embroidered EEG electrodes are textile-based biosensors that use conductive threads in structured embroidery to achieve flexible, gel-free neural monitoring.
• The design employs hierarchical layers with precise stitch dimensions to optimize electrical properties and ensure effective biopotential acquisition.
• Modular integration into wearable systems allows customizable electrode placement and has shown performance comparable to commercial EEG caps in key neurophysiological metrics.

Embroidered electrodes for electroencephalography (EEG) are textile-based, flexible biosensors fabricated by integrating conductive threads into fabric substrates using controlled embroidery techniques. In recent developments, wearable EEG systems employing embroidered electrodes have addressed significant limitations of conventional EEG devices, including the need for abrasive skin preparation, gel application, and rigid electrode arrays. Solutions such as the modular smart headband incorporate adjustable, replaceable embroidered electrodes, yielding mechanically compliant, washable, and comfortable EEG interfaces suitable for real-world, long-term neurophysiological monitoring (Komal et al., 24 Nov 2025).

1. Electrode Structure and Electrical Properties

The canonical design consists of multiple integrated textile layers, each contributing to the electrode’s structural stability, electrical performance, and user comfort. Modules are based on 100% calico cotton (100 × 50 mm), overlaid with non-conductive interfacing for dimensional maintenance, and a 2 mm Puffy Foam™ spacer to enhance scalp contact. The functional electrode region is a 20 mm circular patch, embroidered with Madeira HC-12 stainless-steel threads (linear resistivity $\rho < 100\ \Omega/\mathrm{m}$), achieving a measured DC resistance of approximately 2 kΩ per electrode (four-point probe, 20 mm diameter).

The embroidery pattern is hierarchical:

• Boundary layer: outline stitches fixing the circular geometry.
• Baseline layer: sparse stitches for substrate stabilization.
• Electrode layer: dense cross-stitch optimizing biopotential acquisition.

Stitch dimensions are precisely specified at 2.2 mm length and 2.5 mm spacing. Contact impedance $Z(\omega)$ measured on forehead skin at 10 Hz–1 kHz yields magnitudes in the range 50 kΩ–100 kΩ at 10 Hz, compatible with signal acquisition from scalp potentials. Signal-to-noise ratio (SNR) is expressed as $$\mathrm{SNR} = 20\,\log_{10}({V_{\rm signal}}/{V_{\rm noise}})$$.

2. Integration into Wearable Modular Systems

Embroidered electrodes are engineered as modules within a broader wearable device architecture. The representative headband utilizes a 60 cm neoprene band (8.5 cm edge width, tapering to 5 cm at center, 3 mm thick, total weight ∼20 g) with Velcro closure and circumferential adjustability for head sizes from about 50 to 64 cm. Secure and customizable electrode placement is implemented via 12 horizontal slits (six pairs, 40 mm each, situated 1 cm from the band’s upper/lower edges), facilitating approximately 15 mm of lateral repositioning per electrode.

Each electrode is housed on a fabric patch with a conductive face for skin contact and a rear-side 13 mm snap connector, also produced from HC-12 thread, enabling direct, low-resistance connection to external shielded wiring. Snap-terminated cables link all individual electrodes to the acquisition system (e.g., Neuroelectrics Enobio8 microcontroller), with subsequent wireless transmission via Bluetooth/LSL to controller software (Neuroelectrics Instrument Controller, NIC). Modules are fully removable, supporting regular washing and replacement without discarding the headband or underlying circuitry.

3. Experimental Paradigms and Signal Processing

Functional validation of embroidered EEG systems has been conducted on 10 healthy participants using a tripartite behavioral protocol:

1. Eyes Open/Closed: 1 min with eyes open, 1 min with eyes closed to elicit alpha-band modulation.
2. Auditory Oddball: 360 trials (80% 600 Hz standard, 20% 900 Hz deviant, 100 ms duration, ISI 500–600 ms), with target tone counting.
3. Visual Oddball: 360 images (80% objects, 20% faces, 300 ms duration, ISI 250–350 ms), requiring face stimulus counting.

EEG pre-processing utilized EEGLAB with a 0.3–30 Hz bandpass filter. Data underwent visual inspection, gross artifact rejection, and identification of bad channels via Local Outlier Factor (typically 0–1 channels removed for headband, 2–3 for commercial cap). ICA (with ICLabel classification) isolated blink/muscle artifacts. Channels were interpolated and data rereferenced to Fz, harmonizing montage across systems. Alpha power differences (ΔPα) between conditions were estimated using Welch’s method on power spectral density (8–12 Hz). ERP components were extracted around stimulus timings: P300 (250–500 ms, oddball paradigms) and N170 (150–200 ms, visual oddball). Statistical validation applied cluster-based permutation tests (FieldTrip, dependent-samples $t$, 1000 permutations, $p=0.05$ threshold).

4. Comparative Performance Metrics

Quantitative results demonstrate that embroidered electrode arrays perform equivalently to commercial EEG caps across canonical neurophysiological signatures:

Metric	Headband Value	Cap Value	p-Value
ΔAlpha Power (uV²)	45–50	45–55	0.015 vs. 0.044
P300 (µV)	–4 to +4	–4 to +4	0.014 vs. 0.064
N170 (µV)	–3.5	–4.0	0.013 vs. <0.001

• In the eyes open/closed condition, headband alpha power (45–50 µV²) closely matches commercial cap values (45–55 µV²), each showing significant within-condition differences (headband $p=0.015$, cap $p=0.044$).
• The auditory P300 is detectable with comparable amplitude (–4 to +4 µV), cluster $p=0.014$ at TP7, T7, T8, TP8 (headband), with commercial cap showing similar—but less statistically robust—findings.
• The visual N170 peak is observed with $p=0.013$ (headband), $p<0.001$ (cap), both at temporoparietal electrode sites.

This equivalence extends to data quality, as indicated by similar SNR and statistical discrimination of task effects.

5. Usability, Comfort, and User Assessment

Usability was evaluated via a structured Likert survey addressing overall comfort, skin irritation, controller weight, and device stability. The embroidered headband was distinguished by high user-rated comfort (most responses at the top of the scale), minimal skin irritation (9/10 reported none), and perceptions of the headband as stable and controller weight as “noticeable but not uncomfortable.” Key practical advantages included elimination of gels and skin preparation, reduction of pressure points through soft, breathable fabrics, and the ability to locally dry and machine wash electrodes. The modular approach enables straightforward replacement or repositioning of electrodes for varying monitoring needs.

6. Implications and Future Directions

The integration of embroidered electrodes fabricated from HC-12 stainless-steel thread on calico/foam substrates demonstrates robust measurement of spontaneous (alpha) and event-related (P300, N170) EEG activity, with performance on par with traditional sponge-based caps in both accuracy and user comfort. The modular slit-and-Velcro design supports flexible array configuration, easy maintenance, and sustainable hygienic practices. This suggests that textile-based electrodes could facilitate broader real-world, everyday neurophysiological monitoring across diverse populations.

Planned advancements include scaling to higher channel counts (16–32) for expanded cortical coverage, incorporation of cost-effective microcontrollers (e.g., Arduino) to further democratize and integrate DIY neurotechnology, and exploration of multimodal biosignals (EEG + EMG) within unified wearable platforms for comprehensive brain–body monitoring (Komal et al., 24 Nov 2025).