Towards an intuitive communication-BCI: decoding visually imagined characters from the early visual cortex using...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Brain-computer interfaces (BCIs) enable individuals with paralysis to control devices using neural signals.
• Current BCI communication often relies on selecting letters from a digital board, which can be slow.
• Visual mental imagery presents a potential alternative for faster and more intuitive BCI-based spelling.

Purpose of the Study:

• To demonstrate the feasibility of decoding visually imagined characters from the early visual cortex.
• To investigate the potential of visual imagery as a communication method in BCIs.
• To assess the accuracy and temporal characteristics of decoding imagined characters.

Main Methods:

• Utilized 7 Tesla functional MRI (fMRI) to record brain activity.
• Sixteen healthy participants performed visual mental imagery of three distinct characters.
• Employed single-trial classification to decode the imagined characters from fMRI data.

Main Results:

• Successfully decoded imagined characters with an average accuracy of 54%, significantly above the chance level of 33%.
• Classifiable neural signals were detected shortly after cue onset and persisted with prolonged imagery.
• The early visual cortex showed decodable activity related to imagined characters.

Conclusions:

• Visual mental imagery is a viable strategy for BCI communication.
• The findings support further research into intracranial interfacing for visual imagery-based BCIs.
• This approach could lead to more intuitive and efficient communication solutions for people with paralysis.