Spatial Visual Imagery (SVI)-Based Electroencephalograph Discrimination for Natural CAD Manipulation.

Area of Science:

• Human-Computer Interaction
• Neuroscience
• Computer-Aided Design

Background:

• Current Computer-Aided Design (CAD) interactions can be complex, despite advancements in gaze and gesture controls.
• Implicit design intentions reside in the brain, offering a potential avenue for more intuitive control.
• Bridging the brain and computer via a Brain-Computer Interface (BCI) can enable mental transmission of design intentions for CAD manipulation.

Purpose of the Study:

• To investigate the feasibility of using spatial visual imagery (SVI) for mental control in CAD.
• To develop and validate an electroencephalograph (EEG)-based BCI for intuitive CAD manipulation.
• To provide theoretical support for BCI applications in reducing CAD operational complexity.

Main Methods:

• Focused on a 1-D translation task within the CAD environment.
• Utilized a spatial visual imagery (SVI) paradigm to elicit brain signals.
• Analyzed three spatial EEG features: common spatial patterns, cross-correlation, and coherence.
• Developed a multi-feature fusion model for discriminating SVI-based intentions.

Main Results:

• The multi-feature fusion model achieved an average intent discrimination accuracy of 86% across 10 subjects.
• The highest recorded accuracy for intent discrimination reached 93%.
• The proposed method demonstrated feasibility and good classification performance for discriminating CAD object translation intentions.

Conclusions:

• The study validates the potential of EEG-based BCI for intuitive CAD manipulation.
• Spatial visual imagery (SVI) can be effectively used to decode design intentions for CAD control.
• BCI technology offers a promising pathway towards more natural and efficient CAD interactions.