Parallel Implementation of Ensemble Empirical Mode Decomposition in EEG Signal Processing

This paper presents the design and evaluation of a Parallel Ensemble Empirical Mode Decomposition (PEEMD) algorithm aimed at accelerating EEG signal processing. The proposed approach addresses the computational intensity of traditional EEMD by leveraging a multiprocessing framework on a multicore CPU. The algorithm's performance is assessed through computational time measurements on both synthetic signals and EEG datasets. Results indicate that PEEMD achieves a maximum speedup of 1.75 for synthetic signals and 1.69 for EEG signal processing. The parallel implementation demonstrates significant speedup, particularly for larger ensemble sizes, highlighting its potential for efficient EEG signal analysis.