Using a motor imagery questionnaire to estimate the performance of a Brain-Computer Interface based on object...

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Brain-computer interfaces (BCIs) offer potential for motor rehabilitation.
• Motor imagery (MI) is a key BCI control strategy.
• Predicting BCI performance in users is crucial for effective application.

Purpose of the Study:

• To evaluate motor imagery questionnaires for predicting BCI 'illiteracy'.
• To investigate the influence of different MI paradigms on BCI classifier performance.

Main Methods:

• Administered kinaesthetic (KI) and visual (VI) motor imagery questionnaires to 30 healthy volunteers.
• Recorded EEG during simple imagery (SI) and goal-oriented imagery (GOI) tasks.
• Analyzed correlations between questionnaire scores and BCI classification accuracy.

Main Results:

• Strongest correlation found between KI and SI (r(2)=0.53).
• Classification accuracy was similar for SI (71.1%) and GOI (70.5%), but features differed.
• Goal-oriented imagery improved accuracy for 'poor' imagers and reduced it for 'very good' imagers.

Conclusions:

• Kinaesthetic motor imagery scores may predict MI-based BCI performance.
• Physical presence of an action goal enhances motor imagination in novice users.
• Findings are potentially transferable to upper extremity rehabilitation BCIs for patients.