Abstract:Objective.Decoding different types of movements noninvasively from electroencephalography (EEG) is an essential topic in neural engineering, especially in brain-computer interface. Although the widely used sensorimotor rhythm (SMR) is efficient in limb decoding, it lacks efficacy in decoding movement frequencies. Accumulating evidence supports the notion that the movement frequency is encoded in the steady-state movement-related rhythm (SSMRR). Our study has two primary objectives: firstly, to investigate the spatial-spectral representation of SSMRR in EEG during voluntary movements; secondly, to assess whether movement frequencies and limbs can be effectively decoded based on SSMRR.Approach.To comprehensively examine the representation of SSMRR, we investigated the frequency characteristics and spatial patterns associated with various rhythmic finger movements. Coherence analysis was performed between the sensor or source domain EEG and finger movements recorded by data gloves. A fusion model based on spectral SNR features and filter-bank common spatial pattern features was utilized to decode movement frequencies and limbs.Main results.At the group-level, sensor domain, and source domain coherence maps demonstrated that the accurate movement frequency (f0) and its first harmonic (f1) were encoded in the contralateral motor cortex. For the four-class classification, including two movement frequencies for both hands, the decoding accuracies for externally paced and internally paced movements were 73.14 ± 15.86% and 66.30 ± 17.26% (averaged across ten subjects, chance levels at 31.05% and 30.96%). Notably, the average results of five subjects with the highest decoding accuracies reached 87.21 ± 7.44% and 80.44 ± 7.99%.Significance.Our results verified the EEG representation of SSMRR and proved that the movement frequency and limb could be effectively decoded based on spatial-spectral features extracted from SSMRR. We suggest that SSMRR can serve as a complement to SMR to expand the range of decodable movement types and the approaches of limb decoding.

Enhance Decoding of Functional Lower-Limb Movements by Combining Sensory Motor Rhythm and Movement-Related Cortical Potential Features

Reduce brain computer interface inefficiency by combining sensory motor rhythm and movement-related cortical potential features

Enhanced pre-movement detection of sitting and standing intention based on movement-related cortical potential.

Fast and Accurate Decoding of Finger Movements from ECoG Through Riemannian Features and Modern Machine Learning Techniques

Tactile Stimulation Improves Sensorimotor Rhythm-Based BCI Performance in Stroke Patients.

A Study on Combining Local Field Potential and Single Unit Activity for Better Neural Decoding

Decoding of Self-paced Lower-Limb Movement Intention: A Case Study on the Influence Factors

An Evaluation of Hybrid Deep Learning Models for Classifying Multiple Lower Limb Actions

Improving Movement-Related Cortical Potential Detection at the EEG Source Domain

Decoding movement frequencies and limbs based on steady-state movement-related rhythms from noninvasive EEG

A New Compound-Limbs Paradigm: Integrating Upper-Limb Swing Improves Lower-Limb Stepping Intention Decoding From EEG

Decoding multiclass motor imagery EEG from the same upper limb by combining Riemannian geometry features and partial least squares regression

Stepwise Discriminant Analysis based Optimal Frequency Band Selection and Ensemble Learning for Same Limb MI Recognition

Detection of Lower-Limb Movement Intention from EEG Signals

Enhance Decoding of Pre-Movement EEG Patterns for Brain-Computer Interfaces.

Enhancement of lower limb motor imagery ability via dual-level multimodal stimulation and sparse spatial pattern decoding method

Decoding Premovement Patterns with Task-Related Component Analysis

Pre-movement pattern decoding from motor evoked potentials for reducing interaction delay

Application of Movement-Related Cortical Potentials in the Field of Motor Rehabilitation

Classification of lower limb motor imagery based on iterative EEG source localization and feature fusion

Unilateral movement decoding of upper and lower limbs using magnetoencephalography