Escolar Documentos
Profissional Documentos
Cultura Documentos
MUSCLE
M.J. Conaway, Ph.D.
Introduction
Spinal cord injury (SCI) causes many adverse health consequences including paralysis
and muscle atrophy. Rehabilitative limb integrity maintenance strategies are designed to
facilitate optimal care while a cure for paralysis is found. These electrotherapeutic limb
stimulation strategies are based in part on mathematical models of muscle force development.
The models of Ding and colleagues (4-12) succeeded in predicting force responses to
electrical stimulus inputs with frequencies up to 5 pulses per second (pps) in intact and paralyzed
quadriceps muscles. However, the models failed to predict force responses to inputs with
frequencies greater than 5 pps. This is clinically important since most standard
electrotherapeutic protocols prescribe inputs with stimulation frequencies greater than 5 pps. (15)
The Ding model assumed a constant nonlinear summation of divalent calcium (Ca2+)
currents in single muscle fibers in response to extrinsic electrical stimulation. We hypothesize
that the summation of divalent calcium currents from individual muscle fibers is nonlinear and
time-varying throughout a contraction, which in turn leads to time-varying force generated by a
paralyzed muscle in response to an extrinsic input stimulus. The hypothesis was tested by
modifying the Ding model by making the nonlinear summation a function of the frequency of the
input stimulus. Our model describes the hypothesized nonlinear dynamics of calcium recruitment
with a variant of the time-varying Riccati-Bass function. Using unconstrained optimization, the
augmented model was validated and tested by comparing the force of contraction predicted by
the model to the force of contraction measured by stimulating trained and untrained soleus
muscle in paralyzed subjects.
Methods
Human subjects
We collected force profiles in the Clinical Measurement Laboratory at The University of Iowa
Hospitals and Clinics (UIHC). Subjects with spinal cord injury (SCI) and paralysis qualified for
the study if they have had complete paralysis at the level of T12 or above for at least 1.5 years.
Two of the investigators (RKS, LFL) work extensively with SCI subjects and have assisted in the
identification of the subjects from the laboratory. Subjects of either sex and any race were
considered for enrollment; however four Caucasian male subjects were used. Only subjects
greater than 18 years of age were enrolled. These subjects had no other known medical
conditions.
The subjects were recruited by personal contact, telephone, or mail by physical therapy staff after
a review of records to determine their time after injury. Each subject was asked to fill out a
questionnaire that identifies demographics and injury history. Informed consent was obtained
from each subject. The risks of the study were outlined in the consent form.
The Institutional Review Board of The University of Iowa approved the experimental protocol.
Subject confidentiality was strictly honored. The subjects have not been identified by name in
publications or presentations at scientific sessions.
Age (yr)
33.5
23.8
25.1
23.0
Gender
M
M
M
M
Height (cm)
180
185
188
188
Weight (kg)
111
68
77
68
Force Measurement
SCI* Level
T9
T10
T4
T4
ASIA Scale
A
A
A
A
Years Injured
4.24
1.51
2.73
1.68
We measured the plantar flexion torque with a subject in the seated position with the knee and
ankle at 90o. Adjustability of a heel cup, force transducer, and axis of rotation allowed the force
transducer (AWU-250, Genisco Technology) to be positioned under the first metatarsal head. The
foot plate axis of rotation was aligned with the anatomic axis of the ankle. Stabilization of the
heel to the foot plate during torque production was provided by a rigid ankle cuff directing three
vectors of force through the calcaneus via turnbuckles. A double strap secured over the femur
provided additional stabilization of the heel in the foot plate assembly during plantar flexion. A
schematic of the torque measurement system has been previously shown in the literature (14).
Calibration of the torque measurement system was done with known loads. The calibration
yielded a linear regression equation with an r2correlation value of 0.99. The force transducer was
electronically coupled to a multichannel recorder.
Stimulus Instrumentation
Two silver-silver chloride electrodes (8 mm in diameter, with an inter-electrode distance of
20mm) recorded soleus compound muscle action potential activity to verify supramaximal
activation. Each electrode contained an on-site pre-amplifier with a gain of 36. The signal was
further amplified by a GCS 67 amplifier (Therapeutics Unlimited, Iowa City, IA, USA) with
adjustable gain from 500 to 10,000. The amplifier utilized a high impedance circuit (greater than
15 M at 100 Hz), with a common mode rejection ratio of 87 dB at 60 Hz and a bandwidth of
40-4000 pulses per second.
A custom-designed constant current stimulator and isolation unit was used to provide the
stimulus. The stimulator had a current range from 0 to 200 mA with current variations of less
than 5%, and a maximum voltage capability of 400 V. The stimulator was triggered by digital
pulses from a data-acquisition board (Metrabyte DAS 16F, Keithley Instruments Inc., Cleveland,
OH, USA) housed in a microcomputer under custom software control. The stimulation intensity
was supramaximal, which is approximately 50% greater than the intensity required to produce a
maximum compound muscle action potential. Stimulation inputs were delivered via a doublepronged stimulation electrode secured over the tibia nerve with the use of a gel pad that adhered
to the back of the leg.
Data collection procedures
The skin over the soleus muscle was scrubbed with alcohol, and an electrode was applied 2 cm
lateral to the muscle midline at one-third the distance from the lateral malleolus to the fibular
head. After visual inspection of the M-waves, the electrode was slightly repositioned as needed
in order to accentuate the biphasic waveform. Each subject had the right foot placed in the torque
measuring device and secured as previously described. The test position minimized the
contribution of the gastrocnemius to the plantar flexor torque.
The stimulus electrode was driven with pulse sequences generated by the custom-made constantcurrent stimulator using 250 sec pulse widths. The placement of the electrode was optimized to
produce the largest soleus M-wave peak-to-peak amplitude. Stimulation intensities were then
increased to approximately 1.5 times greater than required for a maximum M-wave to ensure
supramaximal stimulation throughout the protocol.
The force generated by the stimuli, the EMG, and the stimulation trigger signals were
simultaneously recorded and later analyzed using MATLAB 7.0 software (The Mathworks,
Natick, MA). All data were digitized at 1000 samples per second.
Following the work of Frey-Law and Shields (14), we began the stimulation protocol with a
series of 1 pps twitches to determine the optimal electrode placement and ensure supramaximal
levels of activation. We then gave five warm-up contractions, 7-pulse trains (20 pps trains, 330
msec on: 670 msec off) applied at 1 train/sec, to ensure the soleus muscle was in a more
potentiated state, but without inducing fatigue. Following the warm-up, we then administered a
pre-programmed number of stimulation trains with four inputs of each different input waveform.
The set of input waveforms consisted of the following: a twitch, a 5 pps input with no doublet (5
CT), 5 pps inputs with beginning (5 DT) and center doublets (5 DDT), a 10 pps input (10 CT),
followed by another 10 pps input with a beginning doublet (10 DT),an additional 10 pps input
with a dual doublet (10 DDT),a doublet ramp of 15 pulses (DR),and a 20 pps input (20
CT).Some of these pulse trains are shown in Figure 1.
governedbyEqs.13.Thismodeldescribesthetransientbehaviorofthetwostatevariables,
musclestimulation.Thestatevariable
theisometricforceproducedby
behavesasaMichaelisMentenprocess.(23)
Thevariablesaregovernedbythefreeparametersc,A,
timeconstantthatmodulatesCN.Theparameter
,1,and2.Theparametercisthe
22\* MERGEFORMAT ()
33\* MERGEFORMAT ()
44\* MERGEFORMAT ()
Equation 1 represents the time-varying change in developed force as a function of the
time-varying calcium-troponin binding throughout a contraction. Equation 2 represents the
global summation of nonlinear activation of individual muscle fibers in response to an input
train. Equation 3 represents the dynamics of calcium-troponin binding, captured in the unitless
variable
and modulated by the time constant c, which qualitatively describes the rate-
limiting step before the actin and myosin mechanically translate across each other and generate
force. (12)
Ding and colleagues (5) decomposed the contractile response to account for the distinct
physiological step of cross-bridge activation. To model cross-bridge activation, it was shown that
the force-prediction ability of the model is relatively insensitive to the specic curvature and
amplitude of the calcium and calcium-troponin complex transient. This implied that the first two
steps in the 1997 model (4) could be combined into one by the unitless variable CN. From the
following differential equation, the dynamics of CN are modulated by the time constant, c, which
describes qualitative the rate-limiting step before the actin and myosin mechanically translate
across each other and generate force. (5)
To account for the nonlinear summation of calcium transients in single muscle bers
stimulated with doublets, Ding et al (6) proposed the R-model in Eq. 3, which was based on the
earlier work of Duchateau and Hainaut (13) who investigated the force summation from human
adductor pollicis muscles triggered by paired stimuli at different interpulse intervals (IPIs)
ranging from 5 ms to 200 ms. The results showed that the forces generated by the doublet trains
were greater than the sum of two individual twitches. Furthermore, the force enhancement from
the second pulse was highest when the IPI was 5 ms and declined exponentially with increases of
the IPI. The enhanced force of the paired stimuli was suggested to be due to the enhanced release
of divalent calcium by the second pulse (13). In the R-model, a factor Ri was introduced to
account for the nonlinear summation. (6) Ding et al. modied the two-step model by adding a
factor where
is a scaling term that accounts for the differences in the degree of activation by
each pulse relative to the rst pulse of the train (6). The magnitude of the enhancement is
characterized by a scalar
Although the Ding model accurately predicted isometric force for gastrocnemius and
soleus muscles during brief trains of stimulation, it failed to predict force from long-train
stimulations. When the muscle was stimulated with long trains, the model overestimated the
higher frequency force output for both fast-contracting gastrocnemius and slow-contracting
soleus and underestimated the lower frequency force response for soleus muscle (5). One of the
limitations of this model is that it does not account for parameters that vary with time. It was
suggested that better curve tting and predictions of the data from long trains may require the
model to vary several parameters during the contraction (5), Modeling the forces from the long
train stimulation patterns was beyond the scope of this particular study Thus, however, the
changes to the model that are necessary to predict the force generated under long train
stimulation were not quantitatively explored. Furthermore, the original model cannot predict
forces during non-isometric contractions because the force-velocity and length-tension
relationships were not factored into the formulation. Finally, Ding and colleagues suggest that
the model be modied if it is to predict the response of any muscle during fatigue-inducing
repetitive activations. (5) However, they demonstrated that the parameter relationships found for
intact muscle did not apply to paralyzed muscle. (6)
Dingandcolleagues(5,6)hypothesizedthatthemodelwouldaccuratelypredictforce
responsetolongtrainstimulationforpatientswithspinalcordinjury.Thehypothesiswastested
overawiderangeofinputfrequenciesandstimulationpatternsinthenonfatiguedstateusing
datacollectedfromparalyzedhumanquadricepsfemorismuscles.Itwasfoundthatthe
predictionsfromtheforcemodelexplainedmorethan90%andofthevarianceofthecollected
forceprofiles.Hence,themodelwasshowntohavesuccessfullycalculatedrelationships
betweenforceandfrequencyfornonfatiguedmusclestothreevarietiesofinputtrains.
In the Ding models (5, 6), the parameter R0 is treated as a scalar parameter. Yet, it is
defined as a quantitative measure of the nonlinear summation of the Ca2+ from preceding pulses.
In addition to the empirical derivation, the construct of the definition is in itself mathematically
inconsistent. Hence, to transform the parameter to make it congruent with the physiological
reality of equilibrating chemical reactions that occur within a contacting muscle, we extended the
model by modeling R0 with a form of the Riccati-Bass differential equation dF/dt=(p-qF)(1-F).
(2). By including Eq. 5 in our augmented model, we were able to correct this inconsistency and
extend the model, thereby allowing for nonlinearity in the calcium concentration during extrinsic
stimulation.
55\* MERGEFORMAT ()
The time-dependent Riccati-Bass type differential equation (2) describes the manner in which
extra calcium in the sarcomere, elicited from extrinsic stimulation, behaves between pulses. The
differential equation must be negative to model the change in calcium-troponin binding from
subsequent pulses in a stimulus train as a decaying process.
Therefore, using Eq. 5, nonlinearity to the second order may be examined for the
autocatalytic process of periodic calcium release in fresh muscle contraction, especially when
there is no force decay from fatigue. A more accurate estimate of the initial
to calculate Ri and, ultimately, rate-limiting CN during later stages of force generation at higher
stimulate on frequencies and in response to special pulse forms such as doublets. The numerical
values of the model coefficients were determined by parameter estimation. For our study, initial
parameter values were taken from Frey Law and Shields (14) as well as from Bass (2).
Model Validation
Table 2 contains the parameter sets for the Ding and Experimental muscle force models.
For each model, the subset of free parameters, as well as the subset of fixed parameters with
initial values, is given.
Our model of muscle force was validated using an input of 5 pulses per second with a
beginning doublet (5 DT). In addition, the model is shown to be robust using various inputs in
untrained and trained muscles in different subjects. The various inputs are:
Twitch
Doublet ramp
10 pps with beginning doublet (10 DT)
20 pps with no doublet (20 CT)
Table 2 Initial values of parameter sets for force models under unconstrained optimization. Free
parameters were estimated by Levenburg-Marquardt unconstrained optimization.
Model
Ding
Experimental
Fixed parameters
with initial values
c=20 ms
R0=6.9
c=20 ms
Free parameters
A, km, 1, 2
km, 1, 2, 1, 2
A is scaled
algorithm to estimate model parameters. Model agreement with the data was calculated by
minimizing the residual difference between the force predicted from the results of the nonlinear
ordinary differential equation grey-box model and the force measured from test subjects (with
unknown parameters (idnlgrey). In this study, agreement is defined as how well a particular
model traces a force plot taken from a paralyzed human muscle. To estimate agreement, we used
a third-order estimator with three initial states and, for unconstrained optimization, six free
parameters.
In addition To quantify the difference, we used paired t-tests and two-way analysis of variance
without replication in this study. We used paired t-tests to compare average mean square error and
average correlation coefficient in Ding and Experimental model structures. Two-way We used the
two-way ANOVA without replication investigated sources of error between training status and differences
between subjects as well as differences in model predictive ability across different inputs. All statistics
have been done in Excel 2007 (Microsoft, Redmond, WA). The level of significance was set at 0.05.
Results
Validation and testing of our model
Our model of muscle force was validated using an input of 5 pulses per second with a beginning
doublet (5 DT). In addition, the model is shown to be robust using various inputs in untrained
and trained muscles in different subjects. Unconstrained optimization on the six free parameters
is used throughout this study to calculate agreement between model and data.
Human Data
3
2.5
Model/Subject/Trainin
2
g
1.5
Experimental 18
trained
1
Mean
Squared
Error
0.02486
4
Mean
Absolute
Error
0.00984
4
Correlatio
n
Coefficient
0.764018
95% CI
r2
p-value
(0.6612320.866804
)
0.58372
4
<0.000
1
0.5
0
Figure
2.Validation
of
Experimental
force
model
with
5 pps
train with
0
100
200
300
400
500
600
700
800
900
Time (ms)
beginning doublet in trained
muscle in Subject 18.The model had
To investigate
robustness
using
realistic
initial parameter values (c=20, 1=28.3, 2=62.7,
76.4%
agreement
with
the data.
km=0.06, 1=0.38, 2=0.5), our force model was run with various inputs.
The model was run first with a twitch input in the trained muscle of Subject 18.
Human Data
Model/Subject/Trainin
g
Experimental 18
trained
Mean
Squared
Error
0.02101
9
Mean
Absolute
Error
0.00556
3
Correlatio
n
Coefficient
0.923751
95% CI
r2
p-value
(0.9195780.927925
)
0.85331
6
<0.000
1
The model was then run with a doublet ramp input in the trained limb of Subject 17.
4
3.5
3
Force
response
(N)
2.5
2
1.5
1
0.5
0
0
Model/Subject/Training
Experimental 17
trained
1.8
200
400
600
Time (ms)
800
Mean
Mean
Correlatio
Squared
Absolute
n
Error
Error
Coefficient
0.2627 0.071286 0.66334
83
53response. (sim)
7
Force
95% CI
r2
p-value
(0.64215
3
0.70022
1)
0.4400
29
<0.0001
1.6
e response (N)
Figure
4.Doublet ramp in trained muscle in Subject 17. The model had
1.4
66.3% agreement with the data.
1.2
1
0.6
0.4
0.2
0
0
20
40
60
80
100 120
Time (ms)
140
160
180
200
The model was then run with a 10 DT input in the untrained limb of Subject 18.
Force response. (sim)
Human Data
Model fit: 49.28%
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
200
Model/Training/Input
Experimental
untrained 10 DT
400
Mean
Squared
Error
0.21256
9
600
800
Time (ms)
Mean
Absolute
Error
0.103256
1000
Correlatio
n
Coefficient
0.49276
1200
95% CI
r2
p-value
(0.39156
0.59781)
0.24281
<0.000
1
The model was then run with a 20 CT input in the untrained muscle of Subject 18.
Force response. (sim)
Human Data
Model fit: 84.38%
4
Force
3
response
(N)
0
0
100
200
300
400
500
Time (ms)
600
700
Model/Training/Inpu
t
Mean Squared
Error
Mean Absolute
Error
Correlation
Coefficient
95% CI
r2
p-value
Experimental
untrained 20 DT
0.101846
0.0502358
0.843761
(0.7523
8
0.93714
)
0.71193
3
<0.000
1
r2
p-value
(0.8068250.889838
)
0.71966
5
<0.000
1
0.868417
(0.814411
0.922424)
0.75414
9
<0.000
1
0.239113
0.370379
(0.320528
0.418183)
0.13718
<0.000
1
0.34596
6
0.208889
0.531295
(0.489438
0.570709)
0.28227
4
<0.000
1
Ding
untrained 10 DT
Experimental
untrained 10 DT
Ding
untrained 10 DDT
0.28267
0.138498
0.715163
0.511458
0.21256
9
0.441139
0.103256
0.849276
(0.675599
0.750619)
(0.826456 0.86931)
0.305584
0.450897
(0.408164
0.491659)
0.72126
9
0.20330
8
<0.000
1
<0.000
1
<0.000
1
Experimental
untrained 10 DDT
0.36855
9
0.202462
0.498218
(0.457791
0.536589)
0.24822
1
<0.000
1
Ding
untrained 20 CT
0.28267
0.138498
0.715163
(0.675599
0.750619)
0.511458
<0.000
1
Experimental
untrained 20 CT
0.21267
3
0.103256
0.849276
(0.826456 0.86931)
0.72226
1
<0.000
1
Ding
untrained 20 DT
0.18044
6
0.0880614
0.867647
(0.848684
0.884382)
0.752811
<0.000
1
Experimental
0.10184
0.0502358
0.96142
(0.955573
0.92432
<0.000
Model/Training/Input
Ding
trained 10 CT
Mean
Squared
Error
0.01687
2
Mean
Absolute
Error
0.01583
Correlatio
n
Coefficient
0.848331
Experimental
trained 10 CT
0.12140
5
0.015871
Ding
untrained 10 CT
0.38907
2
Experimental
untrained 10 CT
untrained 20 DT
Ding
trained 20 DDT
0.27153
9
0.155377
Experimental
trained 20 DDT
0.20427
2
0.120135
0.966511)
0.757612
(0.722992
0.788436)
0.57397
6
<0.000
1
0.874776
(0.855515
0.891619)
0.76522
2
<0.000
1
Table 4. Comparison of average mean square error and average correlation coefficient in Ding
and Experimental model structures under unconstrained optimization with input frequencies
greater than 5 pulses per second.
Model
Structure
Average Mean
Squared Error
(95% CI)
Ding
0.266344
0.27228
0.223899
0.199028
Experimental
Average
Correlation
Coefficient
(95% CI)
0.556248
0.579553
0.776097
0.358257
From the 95% confidence intervals of the means for both metrics from each model, it is
seen that our model structure has less average mean squared error and greater correlation than
the Ding model structure of muscle force across different inputs and subjects. Furthermore, a
two-sided paired t-test with six degrees of freedom for the average mean squared error at =0.05
yields t=-5,19888. That is significantly less than the critical value t=2.446912. For the
correlation coefficient, a two-sided paired t-test with six degrees of freedom at =0.05 yields t=2.67107. That is considerably less than the critical value t=2.446912.
Table 5.Average prediction measures for doublet ramp inputs across training states from both
models in unconstrained optimization in all Subjects.
Model/Subject/Training
Ding 17
trained
Mean
Squared
Error
0.2594
69
Mean
Absolute
Error
0.071415
1
Correlatio
n
Coefficient
0.70922
3
Experimental 17
trained
0.2627
83
0.071286
53
0.66334
7
Ding 18
trained
0.5029
97
0.123568
0.83168
6
Experimental18
trained
0.1608
62
0.051936
3
0.84793
2
Ding 18
untrained
0.3180
82
0.198839
0.56072
9
Experimental 18
untrained
0.1920
51
0.066694
1
0.78784
9
Ding 27
trained (w/ noise)
0.5137
29
0.280128
0.03512
27
Experimental 27
trained (w/ noise)
0.5097
8
0.280446
0.02879
45
95% CI
r2
p-value
(0.68190
5
0.73456
5)
(0.64215
3
0.70022
1)
(0.81462
5
0.84730
9)
(0.83234
4
0.86218)
0.5029
97
<0.0001
0.4400
29
<0.0001
0.6917
02
<0.0001
0.7189
89
<0.0001
(0.52335
1
0.59595
4)
(0.76691
1
0.80711
2)
(0.01781
3
0.08786
27)
(0.02414
0.3144
17
<0.0001
0.6207
07
<0.0001
0.0012
33
0.1933
77
0.0008
29
0.2863
33
Ding 28
trained
0.2367
06
0.161284
0.82262
2
Experimental 28
trained
0.2441
11
0.160966
0.79552
8
Ding 28
untrained
0.1351
37
0.048723
8
0.90484
5
Experimental 28
untrained
0.1332
02
0.047086
1
0.91725
9
5
0.08157
34)
(0.80474
0.83901
2)
(0.77525
2
0.81416
6)
(0.89498
4
0.91382
3)
(0.90862
2
0.92511
2)
0.6767
06
<0.0001
0.76522
2
<0.0001
0.8187
45
<0.0001
0.8413
65
<0.0001
Table 6. Comparison of average mean square error and average correlation coefficient in Ding
and Experimental model structures under unconstrained optimization for doublet ramp inputs.
Model
Structure
Average Mean
Squared Error
(95% CI)
Ding
0.327687
0.15195
0.250465
0.136109
Experimental
Average
Correlation
Coefficient
(95% CI)
0.644038
0.321587
0.673452
0.326668
From the 95% confidence intervals of the means for both metrics from each model, it is
seen that our model structure has less average mean squared error and greater correlation than
the Ding model structure of muscle force across different subjects as well as training status when
using the doublet ramp input stimulus. Furthermore, a two-sided paired t-test with five degrees
of freedom for the average mean squared error at =0.05 yields t=-2.24589. That is significantly
less than the critical value t=2.57082. For the correlation coefficient, a two-sided paired t-test
with five degrees of freedom at =0.05 yields t=-1.79335. That is considerably less than the
critical value t=2.57082. Hence, the hypothesis that the summation of force from individual
muscle fibers is nonlinear and time-varying throughout a contraction could not be rejected.
Table 7. Two factor analysis of variance without replication for mean squared error between
Ding and Experimental force models for inputs greater than 5 pulses per second using
unconstrained optimization.
Source
of
Variatio
n
Model
SS
Error
0.16468
0.00630
6
0.01297
9
Total
0.18396
4
Training
df
6
1
6
MS
0.02744
7
0.00630
6
0.00216
3
F
12.6879
7
2.91496
7
P-value
0.00348
4
0.13863
F crit
4.28386
6
5.98737
8
13
Table 8.Two factor analysis of variance without replication for mean squared error between Ding
and Experimental force models for doublet ramp inputs using unconstrained optimization.
Source
of
Variatio
n
Subject
Training
SS
0.15944
9
Error
0.01789
0.04862
3
Total
0.22596
2
df
MS
5
0.03189
0.01789
0.00972
5
F
3.27929
1
1.83962
6
P-value
0.10918
9
0.23302
5
F crit
5.05032
9
6.60789
1
11
Conclusion
This study provides the first systematic evaluation of a model that includes calcium
dynamics for predicting force in electrically stimulated, chronically paralyzed human muscle.
The major finding of this study is that, by including a Riccati-Bass diffusion function (2) for R0in
modeling force, the structure of our model predicts force with greater agreement and with less
mean squared error than the Ding model using unconstrained optimization. This is highly
significant significantly. Hence, the specific aim of this study has been successfully answered.
Our force model structure is the first to incorporate Riccati-Bass diffusion functions (2) to
characterize the calcium dynamics in force summation from individual muscle fibers. These
functions are used for making predictions of the continuation of ongoing diffusion processes. In
this study, the problem is one of parameter estimation for specific diffusion function coefficients
at the beginning of actual physiologic processes and then making predictions about the future
behavior of the processes by optimizing the estimated functions at different points in time. (16)
Furthermore, the two influences on the calcium diffusion process are distinguished as probability
of binding and probability of dissociation. (15) These phenomena are reflected by 1 and 2 in
our model. The external influence comes from the extrinsic stimulation, which releases more
free calcium, applied to the muscle that causes calcium to move. Internal influence is a function
of the relevant structures and ions in the sarcoplasmic reticulum, namely calcium channels and
free calcium, which are thought to behave according to specific aggregate probabilities in
different states.
However, upon model optimization, the respective strengths of these two influences are
unequal, with the second parameter always being greater than the first. In fresh muscle, the
affinity for calcium and troponin to dissociate after the first pulse during a stimulated contraction
was always greater than the affinity for calcium and troponin to bind after the first pulse during a
stimulated contraction. It has been shown that the quality of a functional state is a vital factor
that underlies such differences. (16) In fact, an implicit assumption of many muscle farce-
fatigue models is the assumption that diffusion occurs homogeneously, or that the tendency
towards binding or activation remains constant throughout a contraction. (15) The diffusion
processes are functions of time-varying inputs, and hence, the models are exponentially
modulated. However, the processes must also be considered with respect to structural
heterogeneity of a muscle membrane. Hence, the question of whether the assumed mathematical
distribution of free divalent calcium actually modulates the modeled force must be posed.
ThevariableR0hasthefollowingphysiologicalrationale.Repetitivestimulationof
muscleselicitscontractionsthatsummatenonlinearly,especiallywhentensionsaturatesatthe
highfrequencytetaniclevel.Theforcefrequencycurveunderisometricconditionsisgenerally
sigmoidal.Thisindicatesanadditionalnonlinearityatlowfrequencies.Ithasalsobeenfound
thattheforcefromtwocloselyspacedstimulicouldbesignificantlylargerandmoreprolonged
comparedtoatwitch.Atlowfrequenciesthatgeneratetwitcheswithoutfusion,staircase
phenomenainbothmathematicaldirectionshavebeenobserved.Thisimpliesafacilitationor
depressionoftheforceprofilesfromsuccessiveinputs,aswellasposttetanicpotentiationof
twitches.Thelengthtensionandforcevelocitycurvesofmuscleshowotherwellknown
nonlinearities.Finally,smallmagnitudenonlinearitieshavebeenrecordedthatareassociated
withthebendingofmyofilamentbondsuntilfracture.(22)
SteinandParmiggiani(22)foundtwophasesofnonlinearsummationunderawiderange
ofconditionsinfastandslowtwitchmammalianmuscles.Earlydepressionorlessthanlinear
summationoccurswhenasecondcontractionissuperimposedontherisingphaseofatwitch,
andaphaseoffacilitationormorethanlinearsummationisseenlaterinthetimecourseofthe
twitch.Thetwophasescouldbedifferentaspectsofthesamebasicphenomenonsincetheearly
depressionbecomesmoreprominentwithrepetitivestimulationasthelaterfacilitationbecomes
lesspronounced.Yet,theybelievedthatthetwotypesofnonlinearitiesarisefromdistinct
mechanismsforseveralreasons.First,laterfacilitationwasstillpresentwithtimeintervals
greaterthanthecontractiontimeofthemusclesthatwerestudied.Meanwhile,earlydepression
wastotallyeliminated.
Withrespecttowholemuscleactivation(Ri),SteinandParmaggiani(22)additionally
foundthattheincreasingdepressionwithadditionalpulsesandvariationofthedepressionwith
thetwitchtotetanusratiounderavarietyoftestconditionsweretotallyconsistentwiththe
initialdepressionduetoasaturatingprocessinamuscle.Fromexperimentsonamphibian
muscleatlowtemperatures,itwashypothesizedthatamuscleismaximallyactivatedfora
prolongedtimeperiod.Thisisdenotedastheplateauoftheactivestate.Thisplateauoccludes
forcegenerationsincealaterinputcouldnotgenerateanyextraforceunlesstheactivestate
decreasedbelowitsplateaulevel.Nevertheless,itisgenerallyacceptedthatmammalianmuscles
atnormaltemperaturesarenotatmaximalactivationfromasingleinput.However,ifasecond
pulseissuperimposedontheriseofatwitch,alessthanlinearsummationofgeneratedforce
initiallyoccurs.Thisisknownasdepression.(22)Asthetwitchpeaks,thisearlydepression
transitionstoasubsequentmorethanlinearsummation.Thisisknownasfacilitation.Several
possiblemechanismshavebeenproposedforthefacilitationofforcegenerationfromsubsequent
inputs.(21)
ParmaggianiandStein(21)foundthatthefacilitationofthetensionproducedbya
subsequentpulsedecayedinexponentiallyasafunctionoftheinterpulseinterval,regardlessof
howfacilitationwasmeasured.Incomparingfastandslowtwitchmuscles,thetimecourseof
thefacilitationdependedonmusclecontractiontime.Thus,itwasconcludedthatthesame
processesthatmodulatetwitchcontractiontimealsolikelymodulatetensionfacilitationduetoa
subsequentpulse.However,tensionfacilitationisremarkablydifferentforlaterpulsesinatrain.
Whilethefacilitationduetoasecondpulsedecreasesexponentiallywiththeinterpulseinterval,
theoptimalintervalforthirdpulsefacilitationisgreaterthan100msinslowtwitchmuscles.
Comparingfastandslowmuscles,theoptimalinterpulseintervalwasfoundtobecloselyrelated
tothedurationofcontraction,andwasshowntorangefrom1.26to1.5timesthedurationofthe
twitchinthefastandslowmusclesinvestigated.Hence,thissuggestedagainarelationship
betweenthemechanismsthatmodulatedurationofcontractionandthosethatgoverntension
facilitation.Nevertheless,ParmiggianiandStein(21)envisionedaprogressivesaturation,instead
ofsuddenocclusion,asadditionaldivalentcalciumreleasedbysuccessivestimulioccupymore
andmoreofthesitesontroponin.
The biophysical interpretation of the mathematics of calcium diffusion
in muscle (
there will be a specific instant in time when the state converges to a limit of
stability. The amount transformed from the aggregate CaTr in a specific
time unit depends on the characteristic of the species-specific limit, and on
the number of perturbations in the state of the aggregate. (17, 19, 23) It is
known that in a system in which consecutive reactions occur in the presence
of a product of autocatalysis, essential conditions exist that allow oscillating
processes with damping, especially in the case of two consecutive reaction of
better enable function will decrease health care costs and increase the societal contributions
made by these individuals. In the modern economy, every bit of savings is exponentially helpful.
ACKNOWLEDGEMENT
The project described was supported by Award Number R01NR010285 from the National
Institute of Nursing Research. The content is solely the responsibility of the authors and does not
necessarily represent the official views of the National Institute of Nursing Research or the
National Institutes of Health.
APPENDIX: List of parameters for muscle force models.
Symbol
Unit
CN
Definition
Value
Varies
Mechanical force
Varies
ti
ms
Varies
Varies
N
tp
ms
Varies
tq
ms
Varies
ms
20
N/ms
Scaling factor
Varies
ms
28.3
ms
62.7
0.06
0.38
km
ms-1
ms-1
0.5
REFERENCES
1.
2.
4.
5.
Ding J, Wexler AS, Binder-Macleod SA. A mathematical model that predicts skeletal
muscle force. IEEE Trans Biomed Eng, 1997; 44(5): 337-48.
6.
Ding J, Binder-Macleod SA, and Wexler AS. Two-step, predictive, isometric force model
tested on data from human and rat muscles. J ApplPhysiol, 1998; 85: 2176-89.
7.
Ding J, Wexler AS, and Binder-Macleod SA. Development of a mathematical model that
predicts optimal muscle activation patterns by using brief trains. J ApplPhysiol, 2000; 88:
917-25.
8.
Ding J, Wexler AS, Binder-Macleod SA. A predictive model of fatigue in human skeletal
muscles. J ApplPhysiol, 2000; 89:1322-32.
9.
Ding J, Wexler AS, Binder-Macleod SA. A predictive fatigue model--I: Predicting the
effect of stimulation frequency and pattern on fatigue. IEEE Trans Neural SystRehabil
Eng. 2002 Mar;10(1):48-58. Erratum in: IEEE Trans Neural SystRehabil Eng. 2003
Mar;11(1):86.
10.
Ding J, Wexler AS, Binder-Macleod SA. A predictive fatigue model--II: Predicting the
effect of resting times on fatigue. IEEE Trans Neural SystRehabil Eng. 2002
Mar;10(1):59-67.
11.
Ding J, Wexler AS, Binder-Macleod SA. A mathematical model that predicts the forcefrequency relationship of human skeletal muscle. Muscle Nerve. 2002 Oct;26(4):477-85.
12.
Ding J, Wexler AS, Binder-Macleod SA. Mathematical models for fatigue minimization
during functional electrical stimulation. J ElectromyogrKinesiol. 2003 Dec;13(6):575-88.
13.
Ding J, Wexler AS, Lee SC, Johnston TE, Scott WB, Binder-Macleod SA.
Mathematical model that predicts isometric muscle forces for individuals with spinal cord
injuries. Muscle Nerve, 2005 Jun; 31(6): 702-12.
14.
15.
Frey-Law LA, Shields RK. Predicting human chronically paralyzed muscle force: A
comparison of three mathematical models. J ApplPhysiol, 2006;100: 1027-34.
16.
17.
18.
Lotka AJ. Studies on the mode of growth of material aggregates. Am J Sci, 1907; 24(141):
199-217.
19.
Lotka AJ. Contribution to the theory of periodic reactions.JPhysChem, 40); 14: 271-4.
20.
Lotka AJ. Elements of physical biology. New York: Dover Publications, 1925.
21. .Lotka AJ. Elements of mathematical biology. New York: Dover Publications, 1953
22. Parmiggiani F, Stein RB. Nonlinear summation of contractions in cat muscles. II. Later
facilitation and stiffness changes. J Gen Physiol. 1981 Sep;78(3):295-311.
23. Stein RB, Parmiggiani F. Nonlinear summation of contractions in cat muscles. I. Early
depression. J Gen Physiol. 1981 Sep;78(3):277-93.
24.