BMI

This activity submodule enables the MiME software to simultaneously run the trained model in real time while presenting the predefined inner-speech protocol to the user. During this process, EEG signals are recorded in parallel and automatically tagged with the appropriate labels based on the protocol timing. This functionality supports accurate dataset expansion, real-time validation, and continuous improvement of the inner-speech decoding models by ensuring synchronized stimulus presentation, EEG acquisition, and label generation.
Additionally, the labeled dataset generated through this module can be used in the future to train and refine advanced AI models for more accurate and robust inner-speech decoding.