Inner speech in motor cortex and implications for speech neuroprostheses

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Speech brain-computer interfaces (BCIs) aim to restore communication for individuals with paralysis.
• Current BCIs often require physical speech attempts, which can be slow and fatiguing.
• Concerns exist about BCIs potentially decoding private inner speech.

Purpose of the Study:

• To investigate the neural representation of inner speech in the motor cortex.
• To determine if inner speech can be decoded in real time.
• To explore methods for preventing unintended decoding of private inner speech.

Main Methods:

• Utilized multi-unit recordings from four participants.
• Analyzed neural activity during attempted and imagined speech tasks.
• Developed and tested strategies to prevent private inner speech decoding.

Main Results:

• Inner speech is robustly represented in the motor cortex and can be decoded in real time.
• Inner speech representations correlate with attempted speech, with a distinct neural "motor-intent" dimension identified.
• Decoding of free-form inner speech was achieved during specific cognitive tasks (sequence recall, counting).

Conclusions:

• Inner speech is a viable target for brain-computer interfaces, potentially offering a less fatiguing communication method.
• Distinguishing between attempted and inner speech is possible through neural decoding.
• Effective strategies exist to safeguard private thoughts from unintended BCI decoding.