Uploaded July 2025 | Updated September 2026, 1 day ago
In this technical demonstration, we showcase the end-to-end neuro-engineering workflow for processing Motor Imagery data and P300 Event-Related Potentials (ERP) using g.BSanalyze, g.tec’s comprehensive biosignal analysis and visualization platform.
The pipeline begins with the construction of a 16-channel physical electrode montage, highlighting how to implement advanced source derivations. We detail the setup of Small and Large Laplacian reference selections to cleanly isolate sensory-motor area activity overlying the C3 and C4 cortical regions. To directly address the need for open, reproducible, and automated software workflows, every operation executed within the graphical user interface automatically outputs functional, machine-readable code to the MATLAB command window, allowing developers to instantly build standalone offline batch-processing scripts.
Moving into preprocessing and raw data streaming, we demonstrate robust data management for high DC offsets using custom 5th-order Butterworth bandpass filters (8−30 Hz) optimized for motor imagery paradigms. The continuous data stream is then segmented and epoch-triggered based on Left/Right visual cues. We showcase the platform's analytical capabilities by extracting Event-Related Desynchronization (ERD) time curves, spectral densities, and ERD maps to visually validate strong, localized cortical activation.
To build an adaptive, real-time brain-computer interface (BCI) classifier, we demonstrate how to compute Common Spatial Patterns (CSP) to dynamically handle movement artifacts and filter data down to the most significant feature channels. By applying mathematical variance calculations, normalization, and logarithmic transformations, we establish a robust feature matrix. This data feeds into a Linear Discriminant Analysis (LDA) classifier, which achieves an exceptional 0% total error rate during validation.
Finally, we transition into high-resolution 3D visualization. g.BSanalyze tracks spatial filter patterns and localized P300 voltage shifts over time, mapping them directly onto a 3D reconstructed skull and anatomical brain model. We also leverage the platform's built-in graphics engine to export ultra-high-DPI images perfectly formatted for academic journal papers and technical presentations. The video concludes with a developer-focused discussion covering computational neuroscience prerequisites, engineering skills needed for real-time applications, and the physiological effects of action observation on mirror neuron networks.
Discover g.BSanalyze Software: https://www.gtec.at/product/g-bsanalyze-biosignal-analysis-software/
Main Website & Research Resources: https://www.gtec.at
In this technical demonstration, we showcase the end-to-end neuro-engineering workflow for processing Motor Imagery data and P300 Event-Related Potentials (ERP) using g.BSanalyze, g.tec’s comprehensive biosignal analysis and visualization platform.
The pipeline begins with the construction of a 16-channel physical electrode montage, highlighting how to implement advanced source derivations. We detail the setup of Small and Large Laplacian reference selections to cleanly isolate sensory-motor area activity overlying the C3 and C4 cortical regions. To directly address the need for open, reproducible, and automated software workflows, every operation executed within the graphical user interface automatically outputs functional, machine-readable code to the MATLAB command window, allowing developers to instantly build standalone offline batch-processing scripts.
Moving into preprocessing and raw data streaming, we demonstrate robust data management for high DC offsets using custom 5th-order Butterworth bandpass filters (8−30 Hz) optimized for motor imagery paradigms. The continuous data stream is then segmented and epoch-triggered based on Left/Right visual cues. We showcase the platform's analytical capabilities by extracting Event-Related Desynchronization (ERD) time curves, spectral densities, and ERD maps to visually validate strong, localized cortical activation.
To build an adaptive, real-time brain-computer interface (BCI) classifier, we demonstrate how to compute Common Spatial Patterns (CSP) to dynamically handle movement artifacts and filter data down to the most significant feature channels. By applying mathematical variance calculations, normalization, and logarithmic transformations, we establish a robust feature matrix. This data feeds into a Linear Discriminant Analysis (LDA) classifier, which achieves an exceptional 0% total error rate during validation.
Finally, we transition into high-resolution 3D visualization. g.BSanalyze tracks spatial filter patterns and localized P300 voltage shifts over time, mapping them directly onto a 3D reconstructed skull and anatomical brain model. We also leverage the platform's built-in graphics engine to export ultra-high-DPI images perfectly formatted for academic journal papers and technical presentations. The video concludes with a developer-focused discussion covering computational neuroscience prerequisites, engineering skills needed for real-time applications, and the physiological effects of action observation on mirror neuron networks.
Discover g.BSanalyze Software: https://www.gtec.at/product/g-bsanalyze-biosignal-analysis-software/
Main Website & Research Resources: https://www.gtec.at
