EMG Analysis

Prince E. Adjei

Kwame Nkrumah University of Science and Technology

Topic: EMG Analysis Module 3: Biomedical Signal Processing

Biosignal Processes And Analysis (BME 366)

Topics:

(1). EMG Physiology and Motor Unit Recruitment

(2). Signal Processing for EMG: Rectification and Envelope Detection

(3). Fatigue Detection Using Frequency-Domain Analysis

(4). Comparing EMG Patterns: Healthy vs Fatigued Muscle

(5). Applications in Sports ,and Rehabilitation.

EMG Analysis

Learning Objectives

•Explain how EMG signals reflect motor unit activity and muscle

recruitment.

•Apply rectification and envelope methods (RMS, AR) to

preprocess EMG signals.

•Analyze fatigue-related changes in EMG using spectral features

like median frequency.

•Compare EMG characteristics in healthy vs fatigued muscle using

envelope and spectral analysis.

•Identify practical applications of EMG fatigue tracking in sports

science and rehabilitation.

Review

•Signals generated by muscle

activity, typically recorded using

electromyogram (EMG).

•The EMG reflects the electrical

activity of the muscle.

•The Central Nervous System

(CNS), controls the action of

muscle fibres that typically

results in muscle movements.

•The EMG aids in the diagnosis of neuromuscular diseases.

Motor Unit Action Potentials (MUAPs)

•A motor unit consists of a single motor neuron and the muscle fibers it

innervates.

•MUAPs are the electrical signals generated when a motor unit is activated.

•Recorded via surface EMG (sEMG) or intramuscular EMG.

MUAP waveform characteristics:

⚬Duration: ~5–15 ms

⚬Amplitude: Typically 50 μV to 5mV (intramuscular), ~0.1–1mV (surface)

⚬Shape depends on fiber diameter, distance from electrode, and

conduction velocity.

Amplitude Variation

•EMG amplitude reflects motor unit recruitment and firing rate.

Increases with:

•Voluntary muscle contraction intensity.

•Closer proximity of active fibers to the recording electrode.

Amplitude variability arises due to:

•Muscle fatigue (may increase due to synchronization, then decline).

•Electrode placement.

•Physiological cross-talk from nearby muscles.

Electrode Spacing

Inter-electrode distance significantly affects signal quality:

•Closer spacing (e.g., <10 mm):Better spatial resolution, but may miss

deeper units.

•Wider spacing (>20 mm): Greater signal amplitude, but more prone to

cross-talk.

•Ideal spacing for sEMG: ~10–20 mm (depends on muscle size and

depth).

•Proper alignment along muscle fiber direction improves MUAP

detection and reduces phase cancellation.

Rectification

•Rectification is a signal processing step where all the negative values of

a signal are converted to positive values.

•Raw EMG signals oscillate around zero because motor unit action

potentials (MUAPs) have both positive and negative phases.

Since these phases can cancel each other when averaging, rectification is

needed to:

•Shift the mean above zero (zero-mean fix)

•Accurately represent the muscle activation level

•Prepare the signal for envelope detection and feature extraction.

•Make the signal's magnitude usable for analysis (e.g., RMS, mean

absolute value).

Half- Wave Rectification

•Zeros out the negative part of the signal, keeping only the positive.

•Only positive deflections are kept; negative deflections are discarded.

•Sometimes used for quick visualizations.

Drawbacks:

•Loses signal energy and waveform symmetry.

•Underestimates muscle activity.

Full-Wave Rectification

•Converts all negative values to positive by taking the absolute value of

the entire signal.

•Preserves the signal’s energy and shape while making it unipolar.

•Ideal for calculating features like mean absolute value,RMS, and

envelope.

Advantages:

•Better representation of overall muscle activity.

•Retains more information than half-wave.

Full vs Half-Wave Rectification

EMG Envelope

•The EMG envelope represents the overall muscle activation trend over

time.

•It is typically extracted after rectification, using either a linear or RMS

smoothing method.

The envelope:

•Makes it easier to quantify activation

•Allows visual inspection of bursts and timing

•Enables downstream feature extraction (e.g., area under the curve)

Linear Envelope

Computed by:

•Full-wave rectifying the EMG signal

•Applying a low-pass filter (e.g., Butterworth filter at 3–10 Hz)

Pros:

•Simple and fast.

•Smooth trend of muscle activity

Cons:

•Sensitive to noise and outliers.

•May miss sharp changes in EMG bursts.

RMS Envelope

Computed by:

•Taking the root mean square over a moving window:

Pros:

•More robust to noise.

•Better reflects the true signal energy.

Cons:

•Slightly higher computation time.

•Window size must be tuned based on task dynamics.

Summary Comparison

Feature

Linear Envelope

RMS Envelope

Computation

Rectify + Low

-pass filter

Moving RMS over rectified

signal

Smoothness

Moderate

High (especially with 200

window)

Noise Robustness

Lower

Higher

Temporal accuracy

High (for short windows)

Tunable via window size

Preferred for

Quick trends, low

-latency

display

Detailed analysis, fatigue

EMG Envelope: Linear vs RMS

AR Envelope in EMG

•An AR envelope uses an autoregressive (AR) model to smooth the rectified

EMG signal, producing an estimate of muscle activity that adapts to the

local dynamics of the signal.

•Instead of afixed-window smoothing (like RMS or low-pass filtering), the

AR envelope models the signal as being generated by a recursive system.

Autoregressive Model of Order 2 (AR(2))

•An AR(2) process assumes that each sample of the signal depends on its

two previous values:

•Where: y(t): Current output (envelope), a1,a2: AR coefficients and e(t):

noise or input (in this case, the rectified EMG signal)

Questions

1) What is the purpose of rectifying an

EMG signal?

2) What is the difference between full-

wave and half-wave rectification of

EMG signals?

3) Why is a200-ms moving-RMS

envelope preferred over a linear

envelope in EMG analysis?

Frequency Domain Fatigue

•As amuscle fatigues (especially during sustained contractions), the frequency

content of the EMG signal shifts:

•There's a gradual shift from higher to lower frequencies.

This is due to physiological changes such as:

•Decreased muscle fiber conduction velocity

•Increased motor unit synchronization

•Altered recruitment patterns

Key Frequency-Domain Metrics

•Mean Frequency (MNF): The average frequency weighted by power distribution.

•Median Frequency (MDF): The frequency that divides the power spectrum into

two halves.

•During Fatigue, MNF and MDF decrease over time

Frequency Domain Fatigue

•As amuscle fatigues (especially during sustained contractions), the

frequency content of the EMG signal shifts:

•There's a gradual shift from higher to lower frequencies.

This is due to physiological changes such as:

•Decreased muscle fiber conduction velocity

•Increased motor unit synchronization

•Altered recruitment patterns.

•Median Frequency (MDF) is a commonly used metric in frequency-domain

analysis of surface EMG (sEMG) signals.

•It reflects the center of gravity of the signal’s power spectrum and is especially

valuable for tracking muscle fatigue.

•As fatigue progresses, there's a shift in EMG spectral power from higher to

lower frequencies, leading to a decline in MDF over time.

How to Compute MDF for Fatigue Monitoring

Step-by-Step Procedure:

1.Segment the EMG signal into non-overlapping blocks.

2.For each block:

⚬Compute the power spectral density (PSD) using Welch’s method.

Median Frequency in EMG Fatigue Analysis

2. For each block

⚬Calculate the cumulative sum of the PSD.

⚬Determine the frequency (fm) at which the cumulative power reaches

50%of the total power →This is the median frequency (MDF).

3. Repeat for all blocks and store the MDF value over time.

4. Plot MDF values against time to visualize the fatigue trend.

Median Frequency in EMG Fatigue Analysis

2. For each block

⚬Calculate the cumulative sum of the PSD.

⚬Determine the frequency (fm) at which the cumulative power reaches

50%of the total power →This is the median frequency (MDF).

3. Repeat for all blocks and store the MDF value over time.

4. Plot MDF values against time to visualize the fatigue trend.

Median Frequency in EMG Fatigue Analysis

Frequency Domain Fatigue Slope

Comparison: Healthy vs Fatigued EMG

Aspect

Healthy Muscle EMG

Fatigued Muscle EMG

Power Spectrum

Power concentrated in higher

frequencies

(70–100 Hz)

Shift toward lower

frequencies

(30–60 Hz)

Median Frequency

Stable

over time

Declines

over time

Envelope Amplitude

Consistent or

increases

during effort

Decay or flattening

as fatigue

progresses

Signal Shape

Clear burst patterns, high

variability

More uniform, flatter signal with

less variance

Conduction Velocity

Normal

Decreases

Motor Unit Behaviour

Balanced recruitment

Increased synchronization,

altered firing

Comparison: Healthy vs Fatigued EMG

Applications of EMG in Sports and Rehabilitation

Rest–Recruitment Pattern Analysis

•Examines how and when motor units are recruited during muscle activation

•Uses EMG to visualize on/off timing and intensity of different muscles

•Useful in identifying: delayed activation,overcompensation, or muscle

synergy disruption.

EMG Features Used:

•Onset latency: Measure reaction delay

•RMS amplitude: Estimate recruitment intensity

•Co-contraction: Identify simultaneous activation

Use Cases:

•Sports Performance: Assess the efficiency of movement patterns.

•Rehab Therapy: Diagnose abnormal muscle firing in stroke or post-surgical

patients

Applications of EMG in Sports and Rehabilitation

Endurance Mapping

•Tracks muscle fatigue and recovery over time

Uses features like:

•Median Frequency (MDF) slope over repeated contractions

•Envelope decay across sustained holds

Generates muscle endurance profiles by region (e.g., quads vs hamstrings)

Use Cases:

•Athlete Screening: Detect asymmetries or endurance deficits

•Rehabilitation Planning: Tailor workloads to recovery capacity

•Injury Prevention: Spot early fatigue zones to adjust training

Questions

1) What spectral change in the EMG signal

indicates muscle fatigue over time?

2) How is median frequency (MDF) used

to track fatigue in an EMG signal?

3) How does the EMG signal of a fatigued

muscle differ from that of a healthy

muscle in both frequency and

amplitude?

Lab Preview: MF Slope from Biceps EMG

Goal

•Detect muscle fatigue by extracting the median frequency (MDF)

slope from a biceps EMG recording.

Procedure

•Load sample EMG signal (e.g., from .mat or .csv)

•Segment into 1-second blocks

•Compute median frequency (MDF) per block using Welch’s method

•Plot MDF vs time

•Fit a linear regression line →extract slope (Hz/s)

Lab Preview: MF Slope from Biceps EMG

Goal

•Detect muscle fatigue by extracting the median frequency (MDF)

slope from a biceps EMG recording.

Procedure

•Load sample EMG signal (e.g., from .mat or .csv)

•Segment into 1-second blocks

•Compute median frequency (MDF) per block using Welch’s method

•Plot MDF vs time

•Extract slope (Hz/s)

Summary

•EMG rectification converts all signal values to positive, and RMS

envelopes are used to smooth and track muscle activity over time.

•Muscle fatigue causes a spectral shift from high to low frequencies,

which can be detected using the median frequency (MDF).

•MDF is computed using Welch’s method and decreases over time

during sustained contractions, indicating fatigue.

•Healthy muscles show high-frequency, high-amplitude EMG, while

fatigued muscles show envelope decay and low-frequency

dominance.