DQEMG: DQEMG: Decomposition-based Quantitative Decomposition-based Quantitative

EMGEMG

Daniel W. Stashuk, PhDDaniel W. Stashuk, PhDUniversity of WaterlooUniversity of Waterloo

Timothy J. Doherty MD, PhD, FRCPCTimothy J. Doherty MD, PhD, FRCPCThe University of Western OntarioThe University of Western Ontario

Copyright Daniel W. Stashuk and Timothy J. Doherty, 2007. Some rights reserved. Content in this presentation is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 License. This license is more fully described at:

http://creativecommons.org/licenses/by-nc-sa/3.0/.

ObjectivesObjectives

► Brief review of quantitative EMGBrief review of quantitative EMG► Principles of decomposition EMGPrinciples of decomposition EMG► Basic overview of DQEMGBasic overview of DQEMG► How can DQEMG be used for clinical and How can DQEMG be used for clinical and

research applicationsresearch applications► Validation and reliability of the methodValidation and reliability of the method

Important Neuromuscular Important Neuromuscular InformationInformation

Numbers of MUs Relative sizes of MUs Morphology of MUs Recruitment of MUs Firing patterns of MUs Functional stability of neuromuscular junctions

Why Develop Quantitative EMG Methods?

• Objectivity

• Increased Sensitivity

• Increased Specificity

• Ability to determine degree of involvement

• Improved ability provide longitudinal

assessments

Quantitative EMG methodsQuantitative EMG methods

►Single Fiber EMGSingle Fiber EMG►Quantitative MUP analysisQuantitative MUP analysis

Semi-automated MUP analysisSemi-automated MUP analysis Automated MUP analysisAutomated MUP analysis

► Interference pattern analysisInterference pattern analysis►Motor unit number estimationMotor unit number estimation

Quantitative EMG methodsQuantitative EMG methods

►Single Fiber EMGSingle Fiber EMG►Quantitative MUP analysisQuantitative MUP analysis

Semi-automated MUP analysisSemi-automated MUP analysis Automated MUP analysisAutomated MUP analysis

► Interference pattern analysisInterference pattern analysis►Motor unit number estimationMotor unit number estimation

What is an EMG Signal?What is an EMG Signal?

MFPMFP - muscle fibre potential - muscle fibre potential

MUPMUP - motor unit potential - motor unit potential

MUPT - motor unit potential trainMUPT - motor unit potential train

Composite EMGComposite EMG

EMG Signal CompositionEMG Signal Composition

MUPT MUPT - components of a MUPT- components of a MUPT

kN

1ikikik -tMUPtMUPT

What is an Electromyographic (EMG) What is an Electromyographic (EMG) Signal?Signal?

)t(t)t(1

nMUPTEMGmN

mm

Example Composite EMG SignalExample Composite EMG Signal

Concepts of EMG Signal Concepts of EMG Signal DecompositionDecomposition

Basic AssumptionsBasic Assumptions:: Each MUP can be detectedEach MUP can be detected Each detected MUP can be recognizedEach detected MUP can be recognized

Basic Requirements: Common MUP feature is available for detectionCommon MUP feature is available for detection

MUPs within the same MUPT are more similar MUPs within the same MUPT are more similar

than MUPs from different MUPTsthan MUPs from different MUPTs

Typical MUPs can be determined for each MUPT Typical MUPs can be determined for each MUPT

(i.e., MUPs must occur in isolation)(i.e., MUPs must occur in isolation)

How can an EMG Signal Be Decomposed?

Steps in EMG Signal DecompositionSteps in EMG Signal Decomposition

Signal AcquisitionSignal Acquisition

Segmentation (Detecting MUPs)Segmentation (Detecting MUPs)

Feature ExtractionFeature Extraction

Clustering of Detected MUPsClustering of Detected MUPs

Supervised Classification of Detected MUPsSupervised Classification of Detected MUPs

Discovering MUPT Temporal RelationshipsDiscovering MUPT Temporal Relationships

Resolving Superimposed MUPsResolving Superimposed MUPs

Steps In Signal DecompositionSteps In Signal Decomposition

Steps In Signal DecompositionSteps In Signal Decomposition

Steps In Signal DecompositionSteps In Signal Decomposition

DQEMGDQEMG► Performs multiple passes through an EMG signal to complete a Performs multiple passes through an EMG signal to complete a

partialpartial decomposition decomposition► Detection passDetection pass

absolute or relative criteriaabsolute or relative criteria multiply and disparately detected MUPsmultiply and disparately detected MUPs

► Clustering passClustering pass STBC (shape and temporal based clustering)STBC (shape and temporal based clustering) Analyzes a selected portion of the signal (3 to 8 s long)Analyzes a selected portion of the signal (3 to 8 s long)

► Multiple Supervised Classification PassesMultiple Supervised Classification Passes Certainty-based classificationCertainty-based classification Analyzes complete signalAnalyzes complete signal Robustly and actively uses firing pattern information Robustly and actively uses firing pattern information

► Temporal relationships passTemporal relationships pass accounts foraccounts for multiply (linked MUPTs)multiply (linked MUPTs)

and disparately (exclusive MUPTs) detected MUPsand disparately (exclusive MUPTs) detected MUPs► Superimposed MUPs are Superimposed MUPs are notnot resolved resolved

DQEMGDQEMG► Concentric or Concentric or

mono-polar needle and mono-polar needle and surface or “macro” surface or “macro” EMG signals acquiredEMG signals acquired

► 30 - 60 sec 30 - 60 sec isometric contractionsisometric contractions

► Mild to moderate intensityMild to moderate intensity

► Twenty or more MUs Twenty or more MUs from 4 – 6 contractionsfrom 4 – 6 contractions

Acquiring EMG SignalsAcquiring EMG Signals

► Apply surface electrode configuration as per Apply surface electrode configuration as per standard motor study with active electrode standard motor study with active electrode over motor point.over motor point.

► Acquire maximal CMAPAcquire maximal CMAP

Acquiring EMG SignalsAcquiring EMG Signals

Acquiring EMG SignalsAcquiring EMG Signals► Apply surface electrode configuration as Apply surface electrode configuration as

per standard motor study with active per standard motor study with active electrode over motor point.electrode over motor point.

► Acquire maximal CMAPAcquire maximal CMAP

► Perform MVC and calculate MVCRMSPerform MVC and calculate MVCRMS

Acquiring EMG SignalsAcquiring EMG Signals

Acquiring EMG SignalsAcquiring EMG Signals► Apply surface electrode configuration as per Apply surface electrode configuration as per

standard motor study with active electrode over standard motor study with active electrode over motor point.motor point.

► Acquire maximal CMAPAcquire maximal CMAP► Perform MVC and calculate MVCRMSPerform MVC and calculate MVCRMS

► Insert needle electrode and position for sufficient Insert needle electrode and position for sufficient signal quality (signal quality monitor V/s or kV/ssignal quality (signal quality monitor V/s or kV/s22).).

► Have subject isometrically contract to desired Have subject isometrically contract to desired level of effort or signal intensity level of effort or signal intensity (%MVCRMS effort and pps intensity monitors) (%MVCRMS effort and pps intensity monitors)

Acquiring EMG SignalsAcquiring EMG Signals

Acquiring EMG SignalsAcquiring EMG Signals► Apply surface electrode configuration as per Apply surface electrode configuration as per

standard motor study with active electrode over motor standard motor study with active electrode over motor point.point.

► Acquire maximal CMAPAcquire maximal CMAP► Perform MVC and calculate MVCRMSPerform MVC and calculate MVCRMS► Insert needle electrode and position for sufficient Insert needle electrode and position for sufficient

signal quality (signal quality monitor V/s or kV/ssignal quality (signal quality monitor V/s or kV/s22). ). ► Have subject isometrically contract to desired Have subject isometrically contract to desired

level of effort or signal intensity level of effort or signal intensity (%MVCRMS effort and pps intensity monitors) (%MVCRMS effort and pps intensity monitors)

► Decompose needle acquired signal Decompose needle acquired signal and calculate available SMUPs and calculate available SMUPs

► Reposition needle and repeat during a Reposition needle and repeat during a subsequent contractionsubsequent contraction

► Continue until sufficient number of SMUPs acquired (20 –35)Continue until sufficient number of SMUPs acquired (20 –35)

Acquiring EMG SignalsAcquiring EMG Signals

Quantitative MUP analysisQuantitative MUP analysis

► Buchthal method – 1950’sBuchthal method – 1950’s► Concentic needle MUPs collected one-by-one Concentic needle MUPs collected one-by-one

from minimally contracting musclefrom minimally contracting muscle SlowSlow ++ patient and operator interaction++ patient and operator interaction Biased population of MUPsBiased population of MUPs Relatively limited information provided – Relatively limited information provided –

only MUP parametersonly MUP parameters

Quantitative Needle EMGQuantitative Needle EMGMorphological parameters - prototypical MUP of each MUPT

duration, number of phases, turns, Vpp, area, area/amplitude ratio

Spike-triggered AveragingSpike-triggered Averaging

Motor Unit Number Motor Unit Number EstimationEstimation

Size of M-potential

Mean S-MUP Size

= MUNE

McComas et al. 1971

Mean S-MUP template MUNE

CMAP

Measurement of MUP Stability

Jitter Measurement

Jitter MCD: 17.9 SBlocking: 0%Mean IPI: 219 S

Methods ofMethods ofacquiring sampleacquiring sample

of S-MUPsof S-MUPs

IncrementalStimulation

MPS STA

Statistical

Quantitative Needle EMGQuantitative Needle EMG

Decomposition-Based S-MUPsDecomposition-Based S-MUPs

Decomposition-Based S-MUPsDecomposition-Based S-MUPs

Decomposition-based mS-MUAPDecomposition-based mS-MUAP

Size PrincipleSize Principle

MUs with smallertwitch tensions andslower contraction timesare recruited beforelarger, faster MUs

Will this impact the sizesof MUPs collected withD-QEMG?

Effect of force onEffect of force on needle detected MUP and needle detected MUP and surface MUP sizesurface MUP size

Boe et al. 2005 Muscle and Nerve

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 50 100 150 200 250 300

Force (N)

RM

S (

mV

)

0

0.25

0.5

0.75

1

1.25

1.5

0 20 40 60 80 100

Force (N)

RM

S (

mV

)

0

0.5

1

1.5

2

2.5

3

0 100 200 300 400 500

Force (N)

RM

S (

mV

)

0

0.5

1

1.5

2

2.5

3

0 50 100 150 200 250 300 350 400

Force (N)

RM

S (

mV

)

Biceps Force-EMG RelationshipsBiceps Force-EMG Relationships

Data from 2 ALS subjects, r values of 0.96 (top) and 0.99

(bottom)

Data from 2 control subjects, r values of 0.99 (top) and

0.99 (bottom)

Force-EMG RelationshipsForce-EMG Relationships

Pearson Correlation ( r )Pearson Correlation ( r )

SubjectSubject BicepsBiceps FDIFDI

11 0.98620.9862 0.93050.9305

22 0.99240.9924 0.96210.9621

33 0.99200.9920 0.77930.7793

44 0.97290.9729 0.94180.9418

55 0.97190.9719 0.98470.9847

66 0.98050.9805 0.73320.7332

77 0.98590.9859 0.74840.7484

88 0.95540.9554 0.98450.9845

99 0.96280.9628 0.98640.9864

1010 0.96630.9663 0.89420.8942

MinimumMinimum 0.95540.9554 0.73320.7332

MaximumMaximum 0.99240.9924 0.98470.9847

MeanMean 0.97660.9766 0.89450.8945

Application of DQEMGApplication of DQEMGto the study of Agingto the study of Aging

McNeil, Doherty, Stashuk, Rice (2005) Motor unit number McNeil, Doherty, Stashuk, Rice (2005) Motor unit number estimates in the tibialis anterior muscle of young, old and very estimates in the tibialis anterior muscle of young, old and very old men. old men. Muscle and NerveMuscle and Nerve. 31: 461-67. 31: 461-67

► Does progressive MU Does progressive MU loss contribute to loss of loss contribute to loss of strength in the very old?strength in the very old?

► Measured strength, Measured strength, muscle mass, MUNEs in muscle mass, MUNEs in the tibialis anterior the tibialis anterior muscle of three groups muscle of three groups of menof men

► Age GroupsAge Groups Young 27 Young 27 3 yrs 3 yrs Old 66 Old 66 3 yrs 3 yrs Very Old 82 Very Old 82 4 yrs 4 yrs

012345678

Young Old Very Old

Age Group

Nega

tive

Peak

Am

plitu

de (m

V)

0

20

40

60

80

100

Young Old Very Old

Age Group

Nega

tive

Peak

Am

plitu

de (µ

V)Maximum M

Mean S-MUP Size

McNeil et al. 2004 Muscle and Nerve

MVIC strength

• Young – 39 Nm• Old – 38 Nm• Very old – 30 Nm

*

*

020406080

100120140160180

Young Old Very Old

Age Group

Num

ber o

f Mot

or U

nits

TA MUNEsTA MUNEs

Loss of muscleMass and StrengthFunctional Decline

*

**

DQEMG in CMTXDQEMG in CMTX

► CMTX is the X-linked form of hereditary motor-sensory CMTX is the X-linked form of hereditary motor-sensory neuropathy – 2neuropathy – 2ndnd most common following CMT1A most common following CMT1A

► Numerous mutations of the GJNumerous mutations of the GJββ1 gene leading to 1 gene leading to abnormalities in Connexin-32 proteinabnormalities in Connexin-32 protein

► Distal wasting and weakness by 3Distal wasting and weakness by 3rdrd decade in males – decade in males – females usually milder coursefemales usually milder course

► Conduction slowing in intermediate range (30 - 40 Conduction slowing in intermediate range (30 - 40 m/s)m/s)

► Reduced motor and sensory amplitudesReduced motor and sensory amplitudes► As part of a longitudinal study we have examined the As part of a longitudinal study we have examined the

hypothenar and biceps/brachialis muscles of 58 hypothenar and biceps/brachialis muscles of 58 subjects at baseline with D-QEMGsubjects at baseline with D-QEMG

CMT-XCMT-X

►Biceps/brachialisBiceps/brachialis ControlsControls CMAP NP ampCMAP NP amp 6.6 6.6 ± 3.6± 3.6 ( 12 ± 2)( 12 ± 2)

S-MUP NP ampS-MUP NP amp 45 ± 2045 ± 20 ( 55 ± 20 )( 55 ± 20 )

MUNEMUNE 144 144 ± 117± 117(272 ± 124)(272 ± 124)

MUP p-pMUP p-p 648 ± 344 648 ± 344 (340(340 ± ± 80)80)

Duration msDuration ms 14.7 ± 5.114.7 ± 5.1 PhasesPhases 2.6 ± 0.42.6 ± 0.4

CMT-XCMT-X

►HypothenarHypothenar ControlsControls CMAP NP ampCMAP NP amp 4.3 4.3 ± 2.3± 2.3 (>5)(>5)

S-MUP NP ampS-MUP NP amp 134 ± 30134 ± 30 (80 ± 30)(80 ± 30)

MUNEMUNE 32 32 ± 32± 32 (125 ± (125 ± 40)40)

MUP p-pMUP p-p 1510 ± 9471510 ± 947 (600(600 ± 50)± 50)

Duration msDuration ms 12.7 ± 2.312.7 ± 2.3 PhasesPhases 2.8 ± 0.42.8 ± 0.4

HypothenarHypothenar

0 1 2 3 4 5 6 7 8 9 100

25

50

75

100

125

150

Hypothenar CMAP neg pk

MU

NE

15/24 female

NormalNormal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

Myopathic 25%Normal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

Myopathic 50%Normal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

Myopathic 75%Normal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

Neuropathic 25%Normal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

Neuropathic 75%Normal

Probability of finding MUP in a muscle of type:

Myopathic

Neuropathic

SummarySummary► D-QEMG is a robust, valid and reliable methodD-QEMG is a robust, valid and reliable method► Further work is needed to determine the Further work is needed to determine the

“usefulness” of these tools“usefulness” of these tools Intra-rater reliability across centersIntra-rater reliability across centers Responsiveness to changeResponsiveness to change Ease of use in clinical trial settingEase of use in clinical trial setting Ability to provide evidence of disease presence Ability to provide evidence of disease presence

or progression earlier, or with greater specificityor progression earlier, or with greater specificity► ALSALS► MyopathiesMyopathies► EntrapmentsEntrapments

Application to measures of NMJ stabilityApplication to measures of NMJ stability

AcknowledgementsAcknowledgements

► Dr. Charles Rice, Dr. Bill BrownDr. Charles Rice, Dr. Bill Brown► Shaun Boe, Chris McNeil, Lou PinoShaun Boe, Chris McNeil, Lou Pino► NSERCNSERC► CIHRCIHR► Dr. Doherty acknowledges the support of Dr. Doherty acknowledges the support of

the Canada Research Chairs Programthe Canada Research Chairs Program

Thank-you