SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

Brain-Computer Interface (BCI) facilitates bidirectional communication between the human brain and a computer through biosensing and electronic signal exchange. This project is designed to use BCI technology to create an aid tool for individuals who lose movement and/or oral expression, especially those in Locked-in State, to communicate their basic needs by building a classification model to interpret Electroencephalogram (EEG) signals to identify the two basic needs: hunger and thirst, and temperature differentiation.
A three-phase experiment was designed based on OpenBCI. The first phase of experiment collected the baseline data of normal state as the comparison group. While the second phase collected data in hunger and thirst state respectively; the third phase collecting data for temperature differentiation. A total of 168 sets of biological data were collected from approximately 40 undergraduate students.
Our classification model successfully analyzes EEG signals and identifies among the physiological states of hunger, thirst, and temperature differentiation. Eventually it reaches the accuracy of 91.9% in training sets and 71.7% in testing sets.