Decoding subjective preference from single-trial near-infrared spectroscopy signals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Cognitive Science

Background:

• Brain-computer interfaces (BCIs) aim to translate neural activity into commands.
• Non-invasive methods like Near-infrared spectroscopy (NIRS) offer portable and cost-effective BCI solutions.
• Decoding cognitive processes like decision-making is crucial for advanced BCI applications.

Purpose of the Study:

• To develop and validate a novel NIRS-based BCI paradigm.
• To decode neural correlates of subjective preference evaluation during decision-making.
• To assess the feasibility of single-trial classification for cognitive intent.

Main Methods:

• Utilized frequency-domain Near-infrared spectroscopy (NIRS) to monitor prefrontal cortex activity.
• Employed a cognitive task involving subjective preference evaluation between two options.
• Extracted mean signal amplitudes as features for classification.
• Applied linear discriminant analysis (LDA) for decoding preferred choice.

Main Results:

• Achieved an average decoding accuracy of 80% for preferred drink selection on a single-trial basis.
• Demonstrated successful identification of neural patterns associated with subjective preference.
• NIRS effectively captured relevant brain activity for BCI control.

Conclusions:

• NIRS is a viable tool for non-invasive BCI development, particularly for decoding decision-making.
• The proposed NIRS-BCI paradigm shows high potential for real-time thought-based communication.
• Further research can expand this approach to more complex cognitive tasks.