PhysioEMG

EMG Analysis Functions for PhysioExperiment Objects

PhysioEMG provides 21 exported functions for electromyography (EMG) analysis, built on top of PhysioCore. It covers the complete EMG analysis pipeline from signal conditioning through clinical interpretation: envelope extraction and amplitude normalization, spectral analysis, muscle activation onset detection, fatigue monitoring, muscle synergy decomposition, and inter-muscular connectivity network analysis -- all operating directly on PhysioExperiment objects.

Installation

You can install PhysioEMG from r-universe:

# the containers build on Bioconductor, so its repositories are needed too
install.packages("BiocManager", repos = "https://cloud.r-project.org")
install.packages("PhysioEMG",
  repos = c("https://x-biosignal.r-universe.dev", BiocManager::repositories()))

Or install the development version from GitHub:

# install.packages("remotes", repos = "https://cloud.r-project.org")
remotes::install_github("x-biosignal/PhysioEMG")

Quick Start

library(PhysioEMG)

# Generate simulated EMG with muscle bursts
pe <- make_emg(n_time = 5000, n_channels = 4, sr = 1000)

# Detect muscle activation onsets
onsets <- emgOnsetDetect(pe, method = "hodges_bui")

# Extract RMS envelope and normalize
pe_env <- emgEnvelope(pe, method = "rms", window_ms = 50)
pe_norm <- emgAmplitudeNormalize(pe_env, method = "peak")

# Decompose into muscle synergies
syn <- muscleSynergy(pe, n_synergies = 3, method = "nmf")

# Analyze inter-muscular coordination network
net <- emgCoherenceNetwork(pe, freq_band = c(10, 50))
coord <- emgCoordinationStructure(net)

Features

Envelope Extraction and Amplitude Normalization

Signal conditioning for amplitude analysis:

emgEnvelope() -- extract signal envelope using RMS (sliding window), Hilbert transform, or lowpass rectification
emgAmplitudeNormalize() -- normalize amplitude to maximum voluntary contraction (MVC) or peak value

Spectral Analysis

Frequency-domain characterization of EMG signals:

emgSpectralMoments() -- compute spectral moments (mean frequency, median frequency, bandwidth) over sliding windows

Onset Detection

Automatic identification of muscle activation timing:

emgOnsetDetect() -- detect muscle activation onsets with two algorithms:
- Hodges-Bui: threshold-based detection on the rectified/smoothed signal
- Teager-Kaiser: energy operator for improved sensitivity to rapid onsets

Fatigue Analysis

Monitor neuromuscular fatigue during sustained or repeated contractions:

emgFatigue() -- track median frequency shift over time in sliding windows (progressive decrease indicates fatigue)
emgFatigueIndex() -- compute fatigue index by comparing spectral properties between initial and final contraction segments

Muscle Synergy Decomposition

Extract coordinated muscle activation patterns underlying motor control:

muscleSynergy() -- decompose multi-channel EMG into synergies using:
- NMF: non-negative matrix factorization (physiologically interpretable, non-negative weights)
- PCA: principal component analysis (orthogonal decomposition)
- ICA: independent component analysis (statistically independent sources)
synergyReconstruct() -- reconstruct EMG signals from a reduced set of synergies (assess reconstruction quality)
synergyCompare() -- compare synergy structures between conditions, sessions, or subjects using similarity metrics

Inter-Muscular Network Analysis

Characterize functional connectivity and coordination between muscles:

emgCoherenceNetwork() -- magnitude-squared coherence networks within specified frequency bands
emgWPLINetwork() -- weighted phase lag index networks (robust to volume conduction artifacts)
emgPartialCoherenceNetwork() -- partial coherence networks controlling for common input effects
emgDirectedGCNetwork() -- directed Granger causality networks for causal inter-muscular coupling
emgDynamicWaveletNetwork() -- time-varying connectivity using wavelet coherence (track coordination changes during movement)
emgCoordinationStructure() -- extract network topology metrics (modularity, hub muscles, clustering coefficient)
emgInterpretNetworkKG() -- interpret network results using anatomical and functional knowledge graphs

Simulated Data Generators

Ready-to-use data for testing, demonstration, and teaching:

make_emg() -- multi-channel EMG with realistic burst patterns
make_emg_contraction() -- EMG with controlled contraction-relaxation cycles
make_emg_fatigue() -- EMG with progressive fatigue characteristics (spectral shift)

Dependencies

R (>= 4.2)
PhysioCore
SummarizedExperiment
stats

PhysioExperiment Ecosystem

PhysioEMG is the EMG analysis layer of the PhysioExperiment ecosystem, a suite of R packages for multi-modal physiological signal analysis:

Package	Description
PhysioCore	Core data structures and accessors
PhysioIO	File I/O (EDF, HDF5, BIDS, CSV, MAT)
PhysioPreprocess	Preprocessing (filters, ICA, resampling)
PhysioAnalysis	Analysis and visualization
PhysioEEG	EEG analysis (ICA, ERP, source, BCI, sleep)
PhysioEMG	EMG analysis (synergy, fatigue, onset)
PhysioECG	ECG and HRV analysis

Visit the r-universe page to browse all available packages.

License

MIT License. See LICENSE for details.

Author

Yusuke Matsui

Governance & support

Part of the Physio ecosystem. Community and policy documents live in the umbrella repository: