Skip to content

Latest commit

 

History

History
51 lines (37 loc) · 2.51 KB

README.md

File metadata and controls

51 lines (37 loc) · 2.51 KB
Raw

README

Finite Impulse Response (FIR) Band-Pass Filter for EEG Data Processing

This code demonstrates the implementation of a Finite Impulse Response (FIR) band-pass filter for processing time-series data obtained from brain parcellation. It utilizes various Python libraries for signal processing, machine learning, and data visualization.

Requirements

  • Python 3.x
  • Required libraries: numpy, matplotlib, mne, scipy, sklearn

Instructions

  1. Data Loading:

    • Ensure that you have EEG data in the BrainVision format (.vhdr file).
    • Uncomment and modify the vhdr_file variable to specify the path to your .vhdr file.
    • Uncomment the relevant code to load the EEG data using mne.io.read_raw_brainvision().

  2. Data Visualization:

    • Visualize the EEG signals using raw.plot() after loading the data.

  3. Synthetic EEG Data Generation:

    • Synthetic EEG data is generated for demonstration purposes.
    • You can adjust parameters such as sampling frequency, duration, and frequency components of the synthetic signal.

  4. Filter Design:

    • Define the sinc function to create filter coefficients using sinc_filter().
    • Specify the desired frequency band using low_cutoff and high_cutoff.
    • Number of filter taps (numtaps) can be adjusted for desired filter characteristics.

  5. Signal Filtering:

    • Apply the band-pass filter to the synthetic EEG signal using np.convolve().

  6. Feature Extraction:

    • Extract features from the filtered signal, such as mean, standard deviation, skewness, and kurtosis of the Power Spectral Density (PSD) values.

  7. Model Training and Evaluation:

    • Train and evaluate machine learning models (Logistic Regression, Random Forest, SVM, Neural Network) using extracted features.
    • Perform hyperparameter tuning for the Random Forest classifier.
    • Evaluate model performance using accuracy metrics and cross-validation.

  8. Plotting:

    • Visualize the original synthetic EEG signal, the filtered signal, and the impulse response of the band-pass FIR filter.

Usage

  • Ensure that all required libraries are installed (numpy, matplotlib, mne, scipy, sklearn).
  • Modify the code as necessary to adapt it to your specific EEG data and analysis requirements.
  • Execute the code in a Python environment.

Note

  • This code serves as a demonstration and may require modifications for use with actual EEG data.
  • Ensure proper understanding of the algorithms and methods employed before applying them to real-world data.