This project aims to reproduct Akkaradet Experiment EEG Chronic stress which classify stress and non-stress under pre-examination condition

• To classify chronic stress (stressed and non-stressed) based on EEG and compare it with Perceived Stress Scale (PSS) score test.
• To explore feature importance to identify which frequency channels are highly significant in classifying chronic stress.

• Chronic Stress can be classified based on EEG data
• Some of the frequency band are more significant in Chronic Stress classification.

• High PSS scores can indicate chronic stress.
• EEG data doesn’t vary with personal information such as age, gender, etc.

st123012	Todsavad Tangtortan
st123459	Anjana Tissera
st122053	Wanchanok Sunthorn
st123010	Tonson Praphabkul
st122876	Aiman Lameesa

• mne library

• 01 ETL
  • select 16-channels out of 32-channels
  • notch filter power line
  • filter
• 02 EDA
  • Feature Extraction (Alpha Beta Gamma)
    • Power spectral density (PSD)
    • Asymmetry
• 03 ML Model
  • SVM
  • NB
  • KNN
  • LR
• 04 DL Model
  • CNN1D
  • LSTM

• Out of the ML models we trained, Support Vector Machine has higher test accuracy.
• Generally, When Relative Gamma frequency band is included as a feature, Model testing accuracy is high.
• Deep Learning Models didn’t result with good accuracy.
• Accuracy of testing Akkaradet’s dataset
  • Low accuracy compared to our test dataset

• Hypothesis 1: Some of the frequency band are more significant
  • The results show that Relative Gamma has a higher significance in classifying chronic stress
• Hypothesis 2 : Chronic Stress can be classified based on EEG data
  • Model accuracy varies with the dataset
  • Our experiment doesn't completely support this hypothesis.

• Bias Stress
• No Consultant
• Small Dataset

• Recording Post Examination
  • Collecting the same participants after post-examination

• Saeed, S. M. U., Anwar, S. M., Khalid, H., Majid, M., & Bagci, U. (2020). EEG based classification of long-term stress using psychological labeling. Sensors, 20(7), 1886.

• Minguillon, J., Lopez-Gordo, M. A., & Pelayo, F. (2016). Stress assessment by prefrontal relative gamma. Frontiers in computational neuroscience, 10, 101.

• Zhang, Y., Wang, Q., Chin, Z. Y., & Ang, K. K. (2020, July). Investigating different stress-relief methods using Electroencephalogram (EEG). In 2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) (pp. 2999-3002). IEEE.