Using the Cocktail Party Effect to Add the Coding Dimension of Auditory Event Related Potential Brain-Computer...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Auditory event-related potential based brain-computer interface (aERP-BCI) is a standard method for brain-computer communication.
• Improving the coding efficiency of aERP-BCI is crucial for practical applications.

Purpose of the Study:

• To enhance the coding efficiency of aERP-BCI by introducing parallel auditory channels.
• To leverage the cocktail party effect for improved brain-computer communication.

Main Methods:

• A novel paradigm using two parallel voice channels (male and female) for oddball sound stimuli was developed.
• Offline and online experiments compared the double voice condition (DVC) with a single voice condition (SVC).
• Electroencephalography (EEG) signals and BCI performance metrics were analyzed.

Main Results:

• The DVC demonstrated a significant difference in EEG responses between focused and ignored vocal channels (97.2% recognition accuracy).
• Target recognition rates for DVC (82.3%) were comparable to SVC (85%).
• The information transfer rate (ITR) for DVC reached 15.3 bits/min, significantly higher than SVC.

Conclusions:

• The cocktail party effect enhances aERP-BCI coding efficiency by enabling parallel channels.
• This method improves performance without reducing stimulus recognition in the focused channel.
• The findings offer a new direction for advancing aERP-BCI technology.