Related Experiment Video
Updated: May 9, 2026
11:15
Applications of EEG Neuroimaging Data: Event-related Potentials, Spectral Power, and Multiscale Entropy
Published on: June 27, 2013
Cross-Approximate Entropy parallel computation on GPUs for biomedical signal analysis. Application to MEG recordings
Mario Martínez-Zarzuela1, Carlos Gómez, Francisco Javier Díaz-Pernas
1Imaging and Telematics Group, E.T.S. Ingenieros de Telecomunicación, University of Valladolid, Paseo de Belén 15, 47011 Valladolid, Spain.
Computer Methods and Programs in Biomedicine
|August 7, 2013
Summary
We developed a fast GPU implementation for Cross-Approximate Entropy (Cross-ApEn) to analyze complex biomedical signals. This significantly accelerates computation, making advanced time series analysis more accessible for large datasets.
Area of Science:
- Biomedical Signal Processing
- Computational Neuroscience
- Time Series Analysis
Background:
- Cross-Approximate Entropy (Cross-ApEn) quantifies time series dissimilarity but is computationally intensive.
- Limited application in biomedical signals due to high computational cost.
- Existing methods are slow for large-scale multidimensional data analysis.
Purpose of the Study:
- To introduce a fast, scalable GPU-based implementation of Cross-ApEn.
- To enable efficient analysis of large multidimensional biomedical datasets.
- To overcome the computational bottleneck of Cross-ApEn.
Main Methods:
- Developed a fully scalable GPU-based algorithm for Cross-ApEn computation.
- Implemented parallel processing of multiple trials/epochs simultaneously.
- Utilized Graphics Processing Units (GPUs) for accelerated computation.
Main Results:
- Achieved an average speedup greater than 250× compared to a multi-core CPU implementation.
- Enabled analysis of a large MEG dataset (30 subjects, 148 channels) in approximately 30 minutes.
- Demonstrated significant computational efficiency over sequential and parallel CPU methods.
Conclusions:
- The proposed GPU implementation drastically reduces Cross-ApEn computation time.
- This advancement makes Cross-ApEn a feasible tool for extensive biomedical signal processing.
- GPU hardware offers the optimal solution for efficient Cross-ApEn analysis of biomedical data.
