January 1, 2023IEEE AccessOpen Access

The proposed technique improved classification accuracy significantly, with K-nearest neighbor (KNN) achieving a mean accuracy of 91.28% and up to 96.71% by raising the feature dimension.

View Full Paper

Ask AI

Bookmark

Why the study?

Attaining higher classification performance by extracting discriminative features from motor imagery-based electroencephalogram signals remains a key focus in brain-computer interface systems.

Population

BCI competition IV (2a) dataset

Comparison

EMD with correlation-based IMF selection plus MCCSP vs MCCSP applied directly to original EEG data

Authors

NANarges Alizadeh SASajjad Afrakhteh MMM. R. Mosavi

Discussion

Member takes

Overview

Proposed EMD-MCCSP approach may enhance multi-class MI-EEG classification; leaves open clinical validation and real-world BCI deployment.

Structured PICO

Population

BCI competition IV (2a) dataset (motor imagery-based EEG signals)

Intervention

Signal processing technique based on empirical mode decomposition (EMD) with correlation-based intrinsic mode functions (IMF) selection and multi-class common spatial patterns (MCCSP)

Comparator

Applying MCCSP directly to the original EEG channel data

Outcome

Classification accuracy

The proposed EMD and MCCSP-based signal processing technique significantly improves the classification accuracy of multi-class motor imagery EEG signals for brain-computer interfaces.

Cite This Study

Alizadeh et al. (2023) studied this question.

synapsesocial.com/papers/6a7d403df1daba58ecd5af9e https://doi.org/10.1109/access.2023.3274704

View Full Paper

Ask AI

Bookmark

January 1, 2023IEEE AccessOpen Access

The proposed technique improved classification accuracy significantly, with K-nearest neighbor (KNN) achieving a mean accuracy of 91.28% and up to 96.71% by raising the feature dimension.

View Full Paper

Ask AI

Bookmark

Why the study?

Attaining higher classification performance by extracting discriminative features from motor imagery-based electroencephalogram signals remains a key focus in brain-computer interface systems.

Population

BCI competition IV (2a) dataset

Comparison

EMD with correlation-based IMF selection plus MCCSP vs MCCSP applied directly to original EEG data

Authors

NANarges Alizadeh SASajjad Afrakhteh MMM. R. Mosavi

Discussion

Member takes

Overview

Proposed EMD-MCCSP approach may enhance multi-class MI-EEG classification; leaves open clinical validation and real-world BCI deployment.

Structured PICO

Population

BCI competition IV (2a) dataset (motor imagery-based EEG signals)

Intervention

Signal processing technique based on empirical mode decomposition (EMD) with correlation-based intrinsic mode functions (IMF) selection and multi-class common spatial patterns (MCCSP)

Comparator

Applying MCCSP directly to the original EEG channel data

Outcome

Classification accuracy

The proposed EMD and MCCSP-based signal processing technique significantly improves the classification accuracy of multi-class motor imagery EEG signals for brain-computer interfaces.

Cite This Study

Alizadeh et al. (2023) studied this question.

synapsesocial.com/papers/6a7d403df1daba58ecd5af9e https://doi.org/10.1109/access.2023.3274704

View Full Paper

Ask AI

Bookmark