NYCU_BCI_2023

HW1_ BCI research review

• Review a BCI research work that was nominated for the BCI award, and create a short video to present your digest of this work.

• Paper：Regulation of arousal via online neurofeedback improves human performance in a demanding sensory-motor task

HW2_EEG Analysis

• Deal with EEG data preprocessing

HW3_Deep Learning for BCI

• Utilize Convolutional Neural Network (CNN) for motor imagery(MI) EEG classification tasks, with experiments on 4 training schemes and 3 model architectures.

• I use conv3D to solve the SCCNet_v2 bonus problem.

Term Project_Motor Imagery Classification

• Select a paper from the candidate list and reproduce the results including data preprocessing on the provided unprocessed BCIC-IV 2a dataset, implementing the proposed model in the paper, and finally compare its performance with the 3 baseline models in HW3.

• Paper：FBCNet: A Multi-view Convolutional Neural Network for Brain-Computer Interface Motor imagery-BCI

Final Car Competition

• For the final car competition, you will be extending your work from the midterm mind-controlled car demo by controlling the car to turn right and left using hybrid-BCI approach.

• We use this github to record and analysis our data ：https://github.com/HeosSacer/SSVEP-Brain-Computer-Interface

• Use alpha wave and utils.py to control our car.

• peak_alpha = the sum of frequency in alpha wave(8-13Hz)/the sum of the frequency in 3-30Hz

peak_alpha = sum of frequency in alpha wave(8-13Hz)/ sum of the frequency in 3-30Hz    

    if (peak_alpha > threshold): 
        action point to 1
    else: 
        action point to 0

• Using 2-bit to control the car : forwad, left, right. ex. 00, 10, 11, 10

Team members : 張瑀芯、許瑋芸、趙品瑄、陳妍沂、楊佳誠