Detection and labeling of neural connectivity patterns of educational chatbot users

This project will be developed within the framework of the Avanza UC project “Neural Dynamics of Learning: Brain Signals Insights for Next-Generation Educational Chatbots,” whose general objective is to measure the cognitive performance of educational chatbot users in order to improve active learning and reduce the phenomenon of “cognitive offloading,” a dynamic in which the user tends to become more passive during interactions, delegating the cognitive load to the chatbot.

During the project experiments, a multi-modal framework will be used, and signals will be recorded from:
- electroencephalogram (EEG),
- video (facial expression, posture),
- audio (voice, with verbal and non-verbal components, since chatbots with voice detection will also be used),
- gaze direction detection of users while reading/interacting with the chatbot.

Temporal correlations related to the text of the interaction with the chatbot will be sought from these modalities, and cognitive load indicators will be labeled.

This IPRE proposes to contribute to the project by developing metrics of neuronal connectivity and causality from electroencephalogram (EEG) signals using different libraries already available, and will correlate them with the text generated by the user-chatbot interaction.