Electromyography: Recording D. Gordon E. Robertson, PhD, FCSB Biomechanics Laboratory, School of Human Kinetics, University of Ottawa, Ottawa, Canada D.

Electromyography:Recording

Electromyography:Recording

D. Gordon E. Robertson, PhD, FCSB

Biomechanics Laboratory,

School of Human Kinetics,

University of Ottawa, Ottawa, Canada

D. Gordon E. Robertson, PhD, FCSB

Biomechanics Laboratory,

School of Human Kinetics,

University of Ottawa, Ottawa, Canada

Biomechanics Laboratory, University of Ottawa 2

EMG Recording: Topics

• Surface or indwelling• Electrode placement• Type of amplifier• Common Mode Rejection Ratio

(CMRR)• Dynamic range and Gain• Input impedance and skin resistance• Frequency response• Telemetry versus directly wired

Biomechanics Laboratory, University of Ottawa 3

Types of Electrodes

Bipolar surface

Needle

Fine-wire

Biomechanics Laboratory, University of Ottawa 4

Surface Electrodes

• lower frequency spectrum

(20 to 500 Hz)

• relatively noninvasive, cabling does encumber subject, telemetry helps

• skin preparation usually necessary

• surface muscles only

• global pickup (whole muscle)

• inexpensive and easy to apply

Surface Electrodes

• pre-gelled disposable electrodes are most common and inexpensive

• MLS pre-amplified electrodes reduce movement artifact

• Delsys Trigno includes 3D accelerometers

Biomechanics Laboratory, University of Ottawa 5

Biomechanics Laboratory, University of Ottawa 6

Indwelling Electrodes

• fine wire or needle

• localized pickup

• difficult to insert

• invasive, possible nerve injury

• produces higher frequency

spectrum (10 to 2000 Hz)

• can record deep muscles

Biomechanics Laboratory, University of Ottawa 7

Electrode Placement

• electrode pairs in parallel with fibres

• midway between motor point and myotendinous junction (or near belly of muscle)

• approximately 2 cm apart, better if electrodes are fixed together to reduce relative movement

Biomechanics Laboratory, University of Ottawa 8

Surface Electrode Placement

motor point frequency spectra

strongest EMG

best

myotendinous junctions

Biomechanics Laboratory, University of Ottawa 9

Noise Reduction and Grounding

• leads should be immobilized to skin

• surgical webbing can help reduce movement artifacts

• ground electrode placed over

electrically neutral area usually

bone

• N.B. there should be only one ground electrode per person to prevent “ground loops” that could cause an electrical shock

Biomechanics Laboratory, University of Ottawa 10

Surface Electrode System(preamplifier type)

Differential amplifier

Leads

Electrodes

Ground or reference electrode

Cable

Biomechanics Laboratory, University of Ottawa 11

Type of Amplifier

• because EMG signals are small (< 5 mV) and external signals (radio, electrical cables, fluorescent lighting, television, etc.) are relatively large, EMG signals cannot be distinguished from background noise

• background noise (hum) is a “common mode signal” (i.e., arrives at all electrodes simultaneously)

• common mode signals can be removed by differential amplifiers

• single-ended (SE) amplifiers may be used after differential preamplified electrodes

Biomechanics Laboratory, University of Ottawa 12

Common Mode Rejection Ratio (CMRR)

• ability of a differential amplifier to perform accurate subtractions (attenuate common mode noise)

• usually measured in decibels (y = 20 log10 x)

• EMG amplifiers should be >80 dB (i.e., S/N of 10 000:1, the difference between two identical 1 mV sine waves would be 0.1 V)

• most modern EMG amplifiers are >100 dB

Biomechanics Laboratory, University of Ottawa 13

Dynamic Range and Gain

• dynamic range is the range of linear amplification of an electrical device

• typical A/D computers use +/–10 V or +/–5 V

• amplifiers usually have +/–10 V or more, oscilloscopes and multimeters +/–200 V or more

• audio tape or minidisk recorders have +/–1.25 V

• EMG signals must be amplified by usually 1000x or more but not too high to cause amplifier “saturation” (signal overload)

• if too low, numerical resolution will comprised (too few significant digits)

Biomechanics Laboratory, University of Ottawa 14

Input Impedance

• impedance is the combination of electrical resistance and capacitance

• all devices must have a high input impedance to prevent “loading” of the input signal

• if loading occurs the signal strength is reduced

• typically amplifiers have a 1 M(megohm) input resistance, EMG amplifiers need 10 M or greater

• 10 G bioamplifiers need no skin preparation

Biomechanics Laboratory, University of Ottawa 15

Skin Impedance

• dry skin provides insulation from static electricity, 9-V battery discharge, etc.

• unprepared skin resistance can be 2 M or greater except when wet or “sweaty”

• if using electrodes with < 1 G input resistances, skin resistance should be reduced to < 100 k

Vinput = [ Rinput / (Rinput + Rskin) ] VEMG

Biomechanics Laboratory, University of Ottawa 16

Skin Impedance: Example

Vinput = [ Rinput / (Rinput + Rskin) ] VEMG

• If skin resistance is 2 M (megohm) and input resistance is 10 M then voltage at amplifier will be [10/(10 + 2) = 0.833] 83.3% of its true value.

• By reducing skin resistance to 100 k this can be improved to 99%.

• By also using a 100 M resistance amplifier the signal will be 99.9%.

Biomechanics Laboratory, University of Ottawa 17

Frequency Response

• frequency responses of amplifier and recording systems must match frequency spectrum of the EMG signal

• since “raw” surface EMGs have a frequency spectrum from 20 to 500 Hz, amplifiers and recording systems must have same frequency response or wider

• since relative movements of electrodes cause low frequency “artifacts,” high-pass filtering is necessary (10 to 20 Hz cutoff)

• since surface EMG signals only have frequencies as high as 500 Hz, low-pass filtering is desirable (500 to 1000 Hz cutoff)

• therefore use a “band-pass filter” (e.g., 20 to 500 Hz)

Biomechanics Laboratory, University of Ottawa 18

Frequency Response

• Typical frequency spectrum of surface EMG

Biomechanics Laboratory, University of Ottawa 19

Typical Band WidthsEMG 20–500 Hz

10–1000 Hz

surface

indwelling

ECG 0.05–30 Hz

0.05–100 Hz

standard

diagnostic

EEG 1–3 Hz

4–7 Hz

8–12 Hz

12–30 Hz

30–100 Hz

delta waves

theta waves

alpha waves

beta waves

gamma waves

muscle forces or

human movements

DC–10 Hz muscle moments

joint trajectories

audio 20–8000 Hz

20–15 000 Hz

20–20 000 Hz

voice

tape

CD

Biomechanics Laboratory, University of Ottawa 20

EMG Sampling Rate

• since highest frequency in surface EMG signal is 500 Hz, A/D (computer) sampling rates should be 1000 Hz or greater (>2 times maximum frequency)

• raw EMGs cannot be correctly recorded by pen recorders since pen recorders are essentially 50 Hz low-pass filters

• mean or median frequencies of unfatigued muscles are around 70 to 80 Hz

• “notch” filters should not be used to remove 50/60 cycle (line frequency) interference because much of the EMG signal strength is in this range

Biomechanics Laboratory, University of Ottawa 21

Telemetry versus Direct Wire

• telemetry has less encumbrance and permits greater movement volumes

• radio telemetry can be affected by interference and external radio sources

• radio telemetry may have limited range due to legislation (e.g., IC, FCC, CRTC)

• cable telemetry (e.g., Bortec) can reduce interference from electrical sources

• telemetry is usually more expensive than directly wired systems

• telemetry has limited bandwidth (more channels reduce frequency bandwidths)

Biomechanics Laboratory, University of Ottawa 22

Telemetered EMG

• Delsys’s Trigno EMG and accelerometry telemetry system