Development of an open-source EEG headset, mobile platform, and software suite for biometrics-enhanced user experience research

Signal quality in uncontrolled environments

Challenge

Maintaining adequate EEG signal quality outside laboratory settings is extremely difficult. Field research environments introduce substantial sources of electrical noise, movement artifacts, and variable electrode contact quality that can severely degrade signal-to-noise ratios.

Our approach

• Research advanced adaptive filtering algorithms specifically optimized for field conditions in published literature.
• Implement real-time sensor monitoring with user feedback for electrode adjustment.
• Source dry electrode arrays with multiple contact points per measurement site to improve reliability
• Incorporate additional reference sensors for environmental electrical noise cancellation.
• Research, develop and test machine learning models to detect and compensate for common field artifacts.

Hardware usability vs. research-grade performance

Challenge

Creating EEG hardware that is simultaneously easy to use, comfortable, affordable, and capable of producing research-grade data involves significant engineering tradeoffs. Traditional research EEG systems use conductive gels and extensive preparation to achieve optimal signal quality, while consumer-grade systems sacrifice data quality for convenience.

Our approach

• Use industry standard 10-20 electrode placement.
• Research UX-relevant brain regions to reduce sensor channel count requirements.
• Implement adaptive amplification and filtering tailored to each electrode’s contact quality.
• Design modular hardware allowing researchers to scale complexity based on research needs.
• Create validation protocols comparing signals against research-grade systems.
• Provide clear documentation of signal quality limitations and appropriate use cases.

Cross-platform software integration

Challenge

Developing a cohesive software ecosystem that functions reliably across desktop and mobile platforms while handling real-time signal processing and analysis requires addressing significant platform differences, performance constraints, and UI/UX considerations.

Our approach

• Utilize a modular architecture with platform-specific optimizations for critical components.
• Implement a tiered processing approach with lightweight preprocessing on mobile devices.
• Develop clear API contracts between system components to ensure interoperability.
• Create comprehensive automated testing across platforms and devices.
• Design UI components with responsive principles appropriate to research contexts.

Real-time processing of complex neurophysiological metrics

Challenge

Translating raw EEG data into meaningful UX metrics in real-time requires computationally intensive signal processing, feature extraction, and pattern recognition algorithms that must operate within tight performance constraints, particularly on mobile devices.

Our approach

• Implement optimized versions of critical signal processing algorithms.
• Utilize GPU acceleration where available for parallel processing tasks.
• Develop progressive computation strategies that refine metrics as more processing time is available.
• Create specialized optimized algorithms for UX-specific metrics rather than general-purpose approaches.
• Implement post processing strategy for compute constrained devices, and add configuration options for researchers.
• Research appropriate downsampling strategies for less critical metrics.

Eye-tracking integration and synchronization

Challenge

Integrating eye-tracking with EEG presents challenges in temporal synchronization, spatial co-registration, and combined analysis of these distinct data streams.

Our approach

• Develop robust timestamp reconciliation algorithms accounting for variable latencies.
• Create integrated calibration procedures ensuring spatial alignment between modalities.
• Plot eye tracking data onto screen recording.
• Develop specialized metrics leveraging the complementary nature of both data sources.

Open-Source hardware manufacturing complexity

Challenge

Creating truly reproducible open-source hardware poses component sourcing challenges, and assembly complexity that can impact data quality and consistency.

Our approach

• Develop comprehensive component tolerance specifications and testing protocols.
• Create modular designs allowing for interface and component substitution.
• Implement software calibration routines compensating for hardware variations.
• Establish relationships with reliable component suppliers for recommended parts.
• Create assembly guides with quality control methods.

Privacy and data security in field research

Challenge

Collecting and wirelessly transmitting biometric data in field settings introduces significant privacy, security, and ethical considerations, particularly with mobile devices that may have varying security models.

Our approach

• Try to implement end-to-end encryption for all participant data when sufficient compute power is available.
• Develop secure local storage with optional centralized synchronization
• Implement automatic purging of raw data after extraction of anonymized metrics.
• Create audit logging for all data access and processing.
• Design offline operation modes minimizing data exposure.
• Develop compliance documentation for common regulatory frameworks (GDPR, HIPAA, etc.).