Covert Speech vs. Motor Imagery: a comparative study of class separability in identical environments

Area of Science:

• Neuroscience
• Cognitive Science
• Brain-Computer Interfaces

Background:

• Comparing brain activity patterns for different cognitive tasks is crucial for understanding neural processes.
• Motor imagery (MI) and covert speech tasks are both non-invasive paradigms used in brain-computer interfaces (BCIs).
• Standardized experimental environments are needed to accurately compare distinct cognitive paradigms.

Purpose of the Study:

• To identify class-dependent features and important neural activities for covert speech and motor imagery (MI) tasks.
• To provide an identical experimental environment for both paradigms to enable direct comparison.
• To investigate the differences in neural processes underlying covert speech versus MI.

Main Methods:

• A single experimental protocol and analysis pipeline were applied to both motor imagery (MI) and covert speech tasks.
• Four volunteers participated in the study.
• Classification accuracy was assessed for both paradigms.

Main Results:

• Average classification accuracy for covert speech tasks was 82.5%, and for motor imagery (MI) tasks was 77.2%.
• Both paradigms showed performance significantly above chance level.
• Key differences in neural activity were identified: motor imagery involved imagined movements and executive control, while covert speech involved linguistic processing stages prior to articulation.

Conclusions:

• Covert speech tasks exhibit higher classification accuracy than motor imagery tasks within an identical experimental setup.
• Linguistic processing stages in covert speech, particularly high-gamma activity, are highly class-dependent and contribute to superior performance.
• The findings highlight distinct neural correlates for language production and motor imagination, offering insights for BCI development.