Related Experiment Video
Updated: Dec 11, 2025
12:09
Network Analysis of the Default Mode Network Using Functional Connectivity MRI in Temporal Lobe Epilepsy
Published on: August 5, 2014
18.4K
Classification of Task-State fMRI Data Based on Circle-EMD and Machine Learning.
Renzhou Gui1, Tongjie Chen1, Han Nie1
1The Department of Information and Communication Engineering, Tongji University, Shanghai 201804, China.
Computational Intelligence and Neuroscience
|August 18, 2020
Summary
This study introduces an improved circle-Empirical Mode Decomposition (circle-EMD) algorithm to enhance functional Magnetic Resonance Imaging (fMRI) data analysis for brain-computer interfaces. The enhanced method improves task-state classification accuracy using deep neural networks.
Area of Science:
- Neuroscience
- Signal Processing
- Machine Learning
Background:
- Functional Magnetic Resonance Imaging (fMRI) data analysis for brain-computer interfaces faces challenges due to nonstationary signals and noise.
- Accurate classification of human brain task states from fMRI is crucial for BCI research.
Purpose of the Study:
- To propose an improved circle-Empirical Mode Decomposition (circle-EMD) algorithm to suppress end effects in fMRI signal analysis.
- To enhance the accuracy of classifying different human brain task states using processed fMRI data.
Main Methods:
- An improved circle-EMD algorithm was developed to decompose fMRI data and filter out noise, building upon the Hilbert-Huang Transform (HHT).
- Filtered fMRI signals were classified using three machine learning models: logistic regression (LR), support vector machine (SVM), and deep neural network (DNN).
Main Results:
- The improved circle-EMD algorithm effectively suppressed noise and extracted intrinsic mode functions (IMFs) from fMRI data.
- Deep neural networks (DNNs) achieved the highest accuracy in classifying task-state fMRI data compared to LR and SVM.
Conclusions:
- The proposed improved circle-EMD algorithm is effective for preprocessing fMRI data in BCI applications.
- Deep neural networks demonstrate superior performance for task-state classification of preprocessed fMRI data, highlighting the potential for advanced BCI systems.
