Eur J Appl Physiol (2000) 83: 524±530 Ó Springer-Verlag 2000

ORIGINAL ARTICLE

E. Roderich Gossen á Digby G. Sale

Effect of postactivation potentiation on dynamic knee

extension performance

Accepted: 27 July 2000

Abstract Six men and four women performed, in sepa-

rate trials, maximal dynamic knee extensions with loads
Introduction
of 15%, 30%, 45% and 60% of maximal isometric knee
The force of muscle twitch contractions is increased after
extension peak torque (MVC). The dynamic extensions
a conditioning tetanus (posttetanic potentiation or PTP)
were done after postactivation potentiation (PAP) had
(O'Leary et al. 1997), a maximal voluntary contraction
been induced with a 10-s MVC, and in a control trial
(postactivation potentiation or PAP) (Hamada et al.
without PAP. PAP, measured as the increase in evoked
2000) or repeated sub-fusion stimuli (MacIntosh and
twitch torque, was 53 (4)% (SE) and 43 (3)% at the time
Kupsh 1987; MacIntosh and Willis 2000). For simplicity,
of the ®rst and second extensions with each load. PAP
all the aforementioned forms of activity-dependent pot-
failed to increase the attained peak velocity with any
entiation will be referred to as PAP. PAP is probably the
load; on the contrary, there was a trend for peak velocity
result of phosphorylation of the myosin regulatory light
to decrease in the ®rst extension, which occurred @15 s
chains via myosin light chain kinase (Sweeney et al. 1993),
after the 10-s MVC. The results suggest that fatigue
which theoretically increases the interaction between the
produced by the 10-s MVC suppressed any bene®t that
myosin cross-bridge and the thin ®lament (Levine et al.
could be derived from the induced PAP. A surface
1996; Persechini et al. 1985). In addition to increasing
electromyogram (EMG) recorded from one muscle of
peak force, PAP increases the twitch's rate of force de-
quadriceps femoris gave no indication of activation
velopment and decreases its time to peak force (Grange
failure in the ®rst knee extension; however, activation
et al. 1993; O'Leary et al. 1997; Sweeney et al. 1993).
impairment speci®c to the rate of force development
PAP also increases the force and rate of force devel-
cannot be ruled out. It is concluded that the strategy
opment of low-frequency (sub-fusion frequencies) te-
employed, namely of having knee extensions performed
tanic isometric contractions (Vandenboom et al. 1993),
soon after the 10-s MVC to maximize PAP at the time of
and under dynamic conditions (shortening contractions
performance, was unsuccessful because there had been
with various loads), again with sub-fusion frequencies,
insucient time for recovery from fatigue. It is possible
PAP increases work and power output (Grange et al.
that a longer recovery time, even at the cost of a di-
1995, 1998).
minished PAP, may have proved bene®cial.
In contrast, PAP does not increase the peak force (Po)
of high-frequency tetani; only the rate of force develop-
Key words Force±velocity relationship á Posttetanic
ment is enhanced (Vandenboom et al. 1993). Nor does
potentiation á Twitch contraction
PAP increase the maximum unresisted shortening velocity
(Vmax) of high-frequency tetani (Butler et al. 1983;
Persechini et al. 1985), or the maximum velocity of un-
resisted maximal voluntary concentric contractions
(Stuart et al. 1988). Although PAP does not increase Po or
Vmax, its positive e€ect on the isometric rate of force de-
E. R. Gossen á D. G. Sale (&) velopment raises the possibility that PAP might enhance
Department of Kinesiology, dynamic contractile performance with loads intermediate
McMaster University, Hamilton, between those representing Po and Vmax; that is, the load±
Ontario L8S 4K1, Canada
e-mail: saled@mcmaster.ca
velocity relation would be shifted upward and to the right.
Tel.: +1-905-5259140 ext. 23565 Thus, PAP could partly explain improvements in jump
Fax: +1-905-5236011 height (GuÈllich and Schmidtbleicher 1995; Young et al.

1998) and arm push speed (GuÈllich and Schmidtbleicher
1995) achieved following maximal isometric contractions
or weight lifting. In these studies, however, the e€ects of
the presence or absence of PAP was not examined.
The purpose of the present study was to examine the
e€ects of PAP on dynamic knee extension performance
with imposed loads between the extremes represented by
Po and Vmax. We tested the hypothesis that PAP, by
increasing the rate of force development (Vandenboom
et al. 1993, 1995), would increase the acceleration and
thus peak velocity and power attained with these inter-
mediate loads.
Methods
Subjects
Ten moderately active subjects (six male, four female) between 22
and 35 years of age volunteered for the study after giving informed,
written consent. To be eligible for participation, all subjects had to
be free of lower leg injury. The study carried the approval of
McMaster University's Human Ethics Committee. Subjects were
instructed to refrain from any heavy lower leg exercises or ca€eine
consumption 24 h prior to testing.
Fig. 1 A schematic of the control (CON) and postactivation
potentiation (PAP) trials that made up each test session. A trial
began with the establishment of a baseline twitch (Pre). In the CON
trial, this was followed by a 10-s rest period, followed, at the times
indicated, by evoked twitches and dynamic knee extensions at one of
the four imposed loads. Twitches marked (*) indicate the extent to
which the knee extensions themselves induced PAP. After a 15-min
rest period, a PAP trial was done, which was identical to the CON
trial except that the 10-s rest pause after the baseline twitch was
replaced with a 10-s maximal voluntary isometric contraction (MVC)
(to induce PAP). When this session (CON+PAP trials) was
completed on one leg, another session (with the second of four
dynamic loads to be tested) was done with the other leg. In a second
visit to the laboratory, two additional sessions (one per leg) were
conducted to test the two remaining dynamic loads. On a third visit
two additional twitch control trials (one per leg) were conducted. The
twitch control trial was identical to the PAP trial, except that the
dynamic knee extensions were replaced with evoked twitches. See text
for details
525
General design
PAP and its e€ect on the load±velocity relation in the human
quadriceps muscle were studied. The experimental protocol is il-
lustrated in Fig. 1. The protocol was designed so that the extent of
PAP could be quanti®ed and monitored from evoked isometric
twitch responses, and the e€ect of PAP on the load±velocity rela-
tion could be assessed by the performance of dynamic knee ex-
tensions (``kicks'') with various loads. An experimental session
consisted of a control trial (CON), in which twitches were evoked
and two kicks were performed with a designated load, and then,
after a 15-min rest period, a PAP trial which was identical to the
CON trial, except for the inclusion of a 10-s maximal voluntary
isometric contraction (MVC) to induce PAP. Each subject repeated
the experimental session (CON+PAP trials) four times, one ses-
sion for each of four loads in random order (15%, 30%, 45% and
60% of the MVC peak torque determined in a previous familiar-
ization session). On a given day, one session was done with one leg,
and a second session with the other leg. In two separate twitch
control experiments, one for each leg, the protocol was identical to
that of a PAP trial (Fig. 1), except that the dynamic knee exten-
sions were replaced with evoked twitches. Finally, prior to the
CON + PAP sessions and the isometric control sessions, each
subject attended a familiarization session in which, for each leg,
electrical stimulation was administered, MVC peak torque deter-
mined, and dynamic knee extensions practised. Altogether, each
subject had to attend the laboratory on four occasions: familiar-
ization session (both legs), two CON + PAP sessions (two of four
loads per session, one load per leg), and one twitch control session
(both legs).
Dynamometer
A Biodex System 3 dynamometer (Biodex Medical Systems, Shir-
ley, NY, USA) was used to test the knee extensor muscles. The
dynamometer's ``isotonic'' mode was used to set loads for the dy-
namic knee extensions (concentric contractions), which began at a
joint angle of 90° and continued to full extension. The dynamom-
eter's isometric mode was used to measure the peak torque of
maximal voluntary isometric contractions (MVC) and the con-
tractile properties of evoked isometric twitch contractions (knee
joint angle ®xed at 90°).
The subject sat on the Biodex ``chair'', which provided thigh
and back support. A Velcro belt was secured over the subject's hips
to minimize movement in the dynamometer chair. The thigh was
secured with wide strap restraints and the ankle was secured to the
leg armature by a Velcro strap. The chair position was adjusted so
that the knee's axis of rotation (tibio-femoral joint) was aligned
with the dynamometer's armature axis of rotation with the knee at
90° (visually estimated).
526
Stimulating and recording electrodes
Prior to electrode attachment, skin surfaces were shaven, abraded
and swabbed with alcohol pads. Stimulation electrodes consisted of
an anode, a 6 cm ´ 4 cm lead plate coated with electrode gel
(Spectra 300) placed on the proximal quarter of the lateral thigh
using surgical tape (Blenderm), and a cathode, a 1.5 cm ´ 1.5 cm
lead plate with 1.5-cm-thick dense foam fastened to one side, which
was initially positioned immediately medial to the tendon of the
sartorius in the inguinal fold. To achieve ®nal positioning, the
subject was instructed to hold the electrode in place and apply
minimal pressure on the foam backing so as to mimic the pressure
that the tape would exert on the electrode. A 250-ls square-wave
pulse from a Grass S11 stimulator (Grass Instruments, Quincy,
Mass., USA) was used to evoke a minimal twitch response. All
twitches were evoked with the dynamometer in the o€ mode to
minimize signal noise. The subject was then instructed to move the
cathode medially down the inguinal fold by approximately 1 cm.
Using the same minimal stimulation intensity, a subsequent twitch
was evoked. This procedure was repeated until the electrode was
positioned directly over the femoral nerve, which was re¯ected by
the largest twitch response. The cathode was then taped into place.
The stimulus intensity was then increased until a maximal twitch
response was evoked.
To monitor agonist EMG activity, electrodes (Meditrace Ag/
AgCl ECG) were used in a monopolar con®guration. The active
electrode was placed on the vastus medialis approximately 8 cm
proximal from the knee. The reference electrode was positioned on
the patellar ligament immediately inferior to the apex of the patella,
while the ground was placed over the tibia midway between the
knee and ankle joints. The chosen electrode placement minimized
the e€ect of the stimulus artefact on the recording of the evoked
muscle action potential (M-wave).
Vastus medialis is one of four heads of quadriceps femoris. It
was chosen because M-waves could be recorded from this muscle
without large stimulus artefacts. While there is some evidence that
all heads of quadriceps have similar activation characteristics
during loaded and unloaded knee extension (Basmajian et al.
1972), which suggests that recording from only vastus medialis is
representative of the whole quadriceps, di€erential activation
among the heads of quadriceps cannot be ruled out.
Protocol
Familiarization session
The primary purpose of this session was to familiarize the subjects
with electrical stimulation of the knee extensors and performance
of maximal dynamic and isometric contractions on the dyna-
mometer. Subjects were required to perform three 10-s maximal
isometric knee extensions (MVCs) per leg. The dynamic loads for
subsequent test sessions were derived as a percentage of the best
MVC peak torque achieved with each leg. Following the MVCs,
each subject was required to do a minimum of three practice
maximal dynamic knee extensions at each of the four loads (15, 30,
45, 60% MVC).
Control (CON) and PAP sessions
Each of these sessions consisted of a CON trial (no 10-s MVC to
induce PAP) followed, after a 15-min rest period, by a PAP trial
(10-s MVC included to induce PAP). One session was done for each
of four randomly assigned dynamic loads [15%, 30%, 45% and
60% of the highest MVC torque obtained during the familiariza-
tion (habituation) session for the leg to be tested]. The time line
shown in Fig. 1 indicates when, during the course of a trial,
twitches were evoked and dynamic knee extensions performed.
Since isometric twitches were used to quantify the amount of PAP
during the dynamic actions, they were evoked as close as possible
to each dynamic knee extension. Due to constraints of the appa-
ratus, the time interval between the evoked twitch and dynamic
knee extensions could be no less than 10 s.
Isometric control session
An isometric twitch control trial was included because of the
possibility of signi®cant degradation of potentiation during the
period between the end of the MVC and the beginning of the ®rst
dynamic knee extension, and the added confounder of additional
time delays between switching from isometric to isotonic modes
on the dynamometer during each measurement interval. The
twitch control trials were identical to the PAP trials except that
isometric twitches were evoked when dynamic knee extensions
would have occurred in the CON and PAP trials. This provided
an estimate of the amount of PAP present just as a dynamic knee
extension was initiated. In addition, comparisons between
twitches evoked prior to and after each contraction during the
four load sessions and those of the twitch control session served
to verify the consistency of potentiation within subjects across test
days.
Measurements
All isometric and dynamic analogue signals were fed directly from
the dynamometer head into a 12-bit A/D converter, then into a
computer sampling at 2000 Hz per channel using CODAS software
(DATAQ electronics, Akron, Ohio, USA). Measures included
torque and armature angle. The EMG was collected using a
bandwidth ranging from 4 to 2000 Hz (Grass P5 series ampli®er,
Grass Instruments, Quincy, Mass., USA) and afterwards digitally
®ltered during processing.
Twitch response measurements included peak twitch torque (Pt)
and the amplitude of the associated muscle compound action po-
tential (M-wave). Dynamic performance measures included peak
torque, peak power, work attained up to the point of peak power
(hereafter referred to as work to peak power) and peak velocity.
Velocity was calculated by di€erentiating angular displacement
using custom residual analysis software. Velocity was then multi-
plied by torque to derive power. The power±time curve was then
integrated to obtain work.
Both voluntary and evoked (M-wave) EMGs were monitored.
The voluntary EMG was intended to determine whether any per-
formance enhancements were in part due to increased motor unit
activation. M-wave measurements monitored any changes in the
muscle compound action potential, and also served to normalize
voluntary EMG. Prior to analysis, all voluntary EMGs were
highpass ®ltered (cuto€ frequency ˆ 10 Hz) with custom software
to remove movement artefacts. To obtain a measure of activation
across an entire movement, an average EMG (AEMG) was cal-
culated by rectifying the signal during each knee extension and
averaging all data points during movement; values were then nor-
malized to M-wave amplitude and expressed as a percentage of the
latter.
Statistics
A three-factor mixed analysis of variance (ANOVA) on all dy-
namic measures (within, 4 loads ´ 2 conditions [CON/PAP] ´ 2
kicks per condition) was performed. Twitch response measure-
ments were analysed by a three-factor ANOVA (within, 4 loads ´ 2
conditions ´ 4 times) to assess the degree of isometric twitch pot-
entiation (PAP) across trials. ANOVA was also used to compare
the twitch responses across the CON, PAP and twitch control
conditions. When signi®cant main e€ects or interactions were
found with ANOVA, the Tukey post hoc procedure tested for
di€erences among mean values. Statistical signi®cance was set at
P £ 0.05. Descriptive statistics include mean and standard error of
the mean (SE).

Results
Twitch contractile properties
Baseline or ``resting'' twitch peak torque values (Pt) were
determined in ten experiments: the control (CON) dy-
namic performance tests with 15, 30, 45 and 60% MVC
loads, the PAP dynamic performance tests at the same
loads and the right and left leg twitch control experi-
ments (i.e. without dynamic performance). The mean Pt
values ranged from 29.1 (2.4) to 32.9 (3.3) N á m, with
no signi®cant di€erences among the mean values. Of
particular note was the lack of signi®cant di€erences
between the CON and PAP trials, indicating that the 15-
min rest interval between these paired experiments was
sucient for restoration of baseline (twitch) conditions;
that is, any PAP induced during the CON trial had
subsided before the beginning of the PAP trial. Simi-
larly, there were no signi®cant di€erences in baseline M-
wave amplitude across the ten experimental conditions.
Values ranged from 17.4 (1.3) to 19.4 (1.3) mV. All
evoked measures showed acceptable reproducibility.
Method errors calculated across the two test days were
7.7% for Pt, 4.8% for Pt immediately following the 10-s
MVC (i.e. PAP) and 7.1% for M-wave amplitude. The
method error for the peak torque of the 10-s MVCs done
on separate days was larger (16.3%).
Figure 2 shows the twitch responses (normalized to
baseline values) throughout the course of the CON, PAP
(both collapsed across the four loads) and twitch control
experiments (collapsed across right and left legs). The
e€ectiveness of the 10-s MVC in inducing PAP and
sustaining it during the observation period is demon-
strated in the twitch control experiment; speci®cally,
PAP was present at the times when the ®rst [52.6 (3.7)%]
and second [43.2 (2.9)%] dynamic knee extensions
Fig. 2 Twitch peak torque (Pt) normalized to baseline values (PRE)
during CON and PAP trials collapsed across the four dynamic loads,
and twitch control trials collapsed across left and right legs
(observations per mean for CON and PAP ˆ 40; twitch con-
trol ˆ 20). n Time points when dynamic knee extensions (``kicks'')
occurred during CON and PAP; + greater than PRE, P < 0.01;
*
di€erent from all other values at same time point, P < 0.05
527
(``kicks'') were performed. In the CON experiment, that
is without the 10-s MVC, it was evident that the kicks
themselves induced PAP. In the PAP trial, the e€ects of
the 10-s MVC and kicks on PAP were additive; that is,
PAP assessed at 30 s and especially at 50 s postMVC
was greater with than without (twitch control trial) the
added kicks (see Fig. 2). Although not presented, the
potentiated twitches had a shortened time course and
increased peak rate of torque development, as shown
previously in this muscle group (Hamada et al. 2000).
M-wave amplitude did not change signi®cantly during
any of the trials (data not shown).
Dynamic knee extension performance
PAP did not increase the peak velocity achieved across
the four dynamic loads (Fig. 3). On the contrary, in the
PAP trial, the peak velocity achieved with kick no. 1
tended to be depressed compared to kick no. 1 in the
CON trial and compared to kick no. 2 in both trials.
Post hoc tests to determine signi®cant di€erences among
the mean values shown in Fig. 3 were not done because
there were no signi®cant interactions. However, the
tendency for peak velocity to be depressed in kick no. 1
of the PAP trial resulted in two signi®cant main e€ects.
There was a main e€ect for condition (P < 0.03): col-
lapsed across kicks and loads, peak velocity was greater
in CON [341.6 (10.6) °/s] than PAP [326.7 (10.6) °/s].
There was also a main e€ect for kick order (P < 0.02):
collapsed across loads and conditions, peak velocity was
greater in kick no. 2 [340.2 (10.6) °/s] than kick no. 1
[328.1 (10.7) °/s].
Fig. 3 Load±velocity relation obtained with the ®rst and second knee
extensions in the control (CON1, CON2) and PAP trials (PAP1,
PAP2). Values are means (10 observations for each mean) and SE
(some error bars omitted for clarity). Since there were no signi®cant
interactions, comparisons among mean values were not made (no post
hoc tests)
528
The peak torque attained during the dynamic knee
extensions (kicks) was also una€ected by PAP; nor did
peak torque change from kick no. 1 to 2. Only the load
main e€ect was signi®cant (P < 0.0001); that is, the
peak torque increased in kicks against greater loads. For
example, in kick no. 1 of the PAP trials, peak torque
increased from 67.6 (6.1) N á m at a load of 15% MVC
to 138.9 (13.0) N á m at 60% MVC.
PAP did not enhance peak power or work to peak
power (i.e. no signi®cant interactions). There were two
notable main e€ects, however (Table 1). In the CON
condition, work to peak power was greater in kick no.
2 than no. 1; peak power and time to peak power
were una€ected. In contrast, in the PAP condition,
peak power was greater in kick no. 2 than no. 1; work
to peak power and time to peak power were unaf-
fected.
AEMG from the vastus medialis, expressed as a
percentage of M-wave amplitude, did not di€er between
the CON and PAP trials [CON, 4.9 (0.2)%; PAP, 4.5
(0.3)%, P ˆ 0.26].
Discussion
PAP does not increase maximal tetanic isometric peak
force (Po) (Vandenboom et al. 1993) or maximal un-
loaded shortening velocity (Vmax) (Butler et al. 1983;
Grange et al. 1993; Stuart et al. 1988). On the other
hand, PAP does increase the isometric rate of force de-
velopment (Vandenboom et al. 1993, 1995, 1997). This
latter ®nding suggests that PAP might enhance perfor-
mance (i.e. increase peak velocity and power) with loads
between the extremes represented by Po and Vmax. By
increasing the acceleration and hence the velocity at-
tained with a given load, the increased rate of force
development induced by PAP would raise the portion of
the load±velocity relation between Po and Vmax upward
and to the right. The present study tested the hypothesis
of a PAP-induced upward and rightward shift of the
load±velocity relation. The results failed to support the
hypothesis; when a series of four loads between Po and
Vmax was tested, there were no increases in peak velocity
or power during periods of twitch PAP, and thus there
were no signi®cant changes in the load±velocity relation.
The loads selected (15±60% MVC) were considered to
Table 1 Peak power, work attained at the time of peak power, and time to peak power of the dynamic knee extensions (``kicks''). Values
(collapsed across the four loads) are mean (SE)
Control trials
Kick no. 1
Mean SE
Peak power (W) 456.9 23.5
Work to peak power (J) 42.7 2.7
Time to peak power (ms) 195 7 201
*P < 0.05 for di€erence between kicks
be representative of loads used in performance (e.g.
throwing, kicking, etc.).
Our results do not agree with previous studies of the
e€ect of ``conditioning'' contractions on dynamic per-
formance. GuÈllich and Schmidtbleicher (1995) found, in
a group of power event athletes, that vertical jump
height increased after three (presumably isometric)
maximal voluntary contractions (MVCs). In addition,
the movement time of an ``arm push'' test was reduced.
A second study (Young et al. 1998) found, in a group of
ten men with at least 1 year's experience in the barbell
squat exercise, that a preceding set of ®ve repetitions of
the barbell squat exercise with the heaviest possible
weight (i.e. 5 RM), increased the average height of a
series of ®ve vertical jumps with a 19-kg load. There was
a 4-min rest interval between the end of the set of squats
and the beginning of the series of vertical jumps. There is
a third brief (abstract) report indicating mixed results of
the e€ects of prior activity on horizontal countermove
jump performance (Radcli€e and Radcli€e 1996). In a
group of 24 male college athletes, 4 sets of 4 repetitions
in the power snatch exercise, with 75±85% of the 4 RM,
increased the distance attained in three countermove
jumps initiated within 3 min of completion of the fourth
set of the prior exercise. In contrast, 4 sets of 4 repeti-
tions of the squat exercise with 75±85% 4 RM, 4 sets of
4 loaded jumps with 15±20% body weight added, and 4
sets of 4 unloaded jumps failed to improve subsequent
countermove jump performance. In a group of 11 female
athletes, none of the prior activities improved perfor-
mance. None of these studies (GuÈllich and Schmidtb-
leicher 1995; Radcli€e and Radcli€e 1996; Young et al.
1998) tested for the presence of PAP following the
conditioning activities, but it was probably present.
The discrepancy between the present and previous
studies, regarding the bene®t of conditioning activities,
is probably related to the conditioning activity's dual
e€ects of inducing both PAP and fatigue (Krarup 1981).
In the present study there were two indications that the
10-s MVC induced fatigue as well as PAP. First, torque
declined on average by 16.3% (P < 0.001) during the
10-s MVC. Secondly, the peak velocity of the ®rst knee
extension was depressed following the MVC. In selecting
a strategy to take advantage of PAP, there is the di-
lemma that a brief interval between the conditioning
contraction inducing PAP and performance maximizes
PAP trials
Kick no. 2 Kick no. 1 Kick no. 2
Mean SE Mean SE Mean SE
467.6 22.9 425.7 19.5 444.5* 20.8
46.4* 2.8 42.7 2.4 43.5 2.3
6 198 7 200 9

both the PAP and fatigue e€ects. A longer recovery in-
terval reduces both fatigue and PAP. Compared to our
recovery interval of 15 s, previous studies that showed a
positive e€ect of conditioning contractions had longer
recovery intervals of 3±4 min (Radcli€e and Radcli€e
1996; Young et al. 1998). Had we used a recovery in-
terval of 4 min, PAP would probably have diminished to
@15±20% (Hamada et al. 2000; Houston and Grange
1990) instead of the @50% found at 15 s in the present
study; nevertheless, the longer recovery time may have
more than compensated for the diminished PAP.
The fatigue caused by the 10-s MVC, which either
impaired dynamic performance or prevented it from
being enhanced by PAP, could have been the result of
central or peripheral mechanisms. In a sustained MVC,
maximum sustainable motor unit discharge rates de-
crease within seconds, and some high-threshold, pre-
sumably fast, motor units cease discharging (Grimby
et al. 1981). Since a high rate of force development
depends on high discharge rates and activation of fast
motor units (Zehr and Sale 1994), dynamic perfor-
mance depending on speed would probably be im-
paired. Further evidence implicating central
mechanisms is that the ability to perform voluntary
contractions with a high rate of force development
deteriorates rapidly when the rate of force development
of evoked contractions is not diminished (Miller et al.
1993). Our observed lack of change in AEMG in the
vastus medialis does not indicate the presence of central
fatigue, but surface EMG may not be sensitive enough
to re¯ect changes in peak motor unit ®ring rates. Fur-
thermore, altered activation of the other heads of
quadriceps cannot be ruled out.
Establishing a strategy for deriving the greatest ben-
e®t from PAP has two key parts: (1) setting the intensity
and duration of the conditioning activity, and (2) setting
the recovery interval between the end of the conditioning
activity and the start of the performance to be improved.
The ®rst part a€ects the relative amounts of PAP and
fatigue produced, and the second part a€ects the net
e€ect of PAP decay and recovery from fatigue. Finding
the best strategy requires a series of ``trial and error''
experiments in which the two key parts of the strategy
are manipulated in various ways. As mentioned earlier,
the present experiment may have failed to support the
hypothesis because of an insucient recovery interval. A
longer recovery interval before the ®rst knee extension,
or several additional knee extensions over a period of 4±
5 min after the 10-s MVC, may have shown the hy-
pothesized results. The latter approach would add the
complication that the knee extensions themselves can
cause both PAP and fatigue. In the present study, the
knee extensions induced PAP (Fig. 2). Similarly, the
increased peak torque over the ®rst knee extensions of
an isokinetic fatigue test (Tesch and Wright 1983;
Thorstensson and Karlsson 1976), and the increased
jump height over a series of jumps (GuÈllich and
Schmidtbleicher 1995) may have re¯ected a progres-
sively developing PAP.
529
In conclusion, the present study has shown that a
10-s MVC, known to cause a high degree of PAP, fails
to enhance dynamic performance which begins @15 s
after the MVC, when PAP is still at a high level.
However, we tested a combination of type, intensity
and duration of conditioning activity, and recovery
interval between conditioning activity and performance
that is but one of many possible strategies intended to
enhance performance. The various strategies may en-
hance, impair, or have no e€ect on performance, de-
pending on the interaction induced between PAP and
fatigue. A series of experiments will be needed to
identify the best strategy.
Acknowledgements John Moroz provided technical assistance. The
study carried the approval of McMaster University's Human
Ethics Committee, and by extension was in accordance with cur-
rent laws in Canada regarding research with human subjects.
References
Basmajian JV, Harden TP, Regenos EM (1972) Integrated actions
of the four heads of quadriceps femoris: an electromyographic
study. Anat Rec 172: 15±20
Butler TM, Siegman MJ, Mooers SU, Barsotti RJ (1983) Myosin
light chain phosphorylation does not modulate cross-
bridge cycling rate in mouse skeletal muscle. Science 220:
1167±1169
Grange RW, Vandenboom R, Houston ME (1993) Physiological
signi®cance of myosin phosphorylation in skeletal muscle. Can
J Appl Physiol 18: 229±242
Grange RW, Cory CR, Vandenboom R, Houston ME (1995)
Myosin phosphorylation augments force±displacement and
force±velocity relationships of mouse fast muscle. Am J Physiol
269: C713±C724
Grange RW, Vandenboom R, Xeni J, Houston ME (1998) Pot-
entiation of in vitro concentric work in mouse fast muscle.
J Appl Physiol 84: 236±243
Grimby L, Hannerz J, Hedman B (1981) The fatigue and voluntary
discharge properties of single motor units in man. J Physiol
(Lond) 316: 545±554
GuÈllich A, Schmidtbleicher D (1995) Short-term potentiation of
power performance induced by maximal voluntary contrac-
tions. XVth Congress of the International Society of Biome-
chanics, pp 348±349
Hamada T, Sale DG, MacDougall JD, Tarnopolsky MA (2000)
Postactivation potentiation, ®bre type, and twitch contraction
time in human knee extensor muscles. J Appl Physiol 88: 2131±
2137
Houston ME, Grange RW (1990) Myosin phosphorylation, twitch
potentiation, and fatigue in human skeletal muscle. Can J
Physiol Pharmacol 68: 908±913
Krarup C (1981) Enhancement and diminution of mechanical
tension evoked by staircase and by tetanus in rat muscle.
J Physiol (Lond) 311: 355±372
Levine RJC, Kensler RW, Yang Z, Stull JT, Sweeney HL (1996)
Myosin light chain phosphorylation a€ects the structure
of rabbit skeletal muscle thick ®laments. Biophys J 71:
898±907
MacIntosh BR, Kupsch CC (1987) Staircase, fatigue and ca€eine in
skeletal muscle in situ. Muscle Nerve 10: 717±722
MacIntosh BR, Willis JC (2000) Force±frequency relationship and
potentiation in mammalian skeletal muscle. J Appl Physiol 88:
2088±2096
Miller RG, Moussavi RS, Green AT, Carson PJ, Weiner MW
(1993) The fatigue of rapid repetitive movements. Neurology
43: 755±761
530
O'Leary DD, Hope K, Sale DG (1997) Posttetanic potentiation of
human dorsi¯exors. J Appl Physiol 83: 2131±2138
Persechini A, Stull JT, Cooke R (1985) The e€ect of myosin
phosphorylation on the contractile properties of skinned rabbit
skeletal muscle ®bres. J Biol Chem 260: 7951±7954
Radcli€e JC, Radcli€e L (1996) E€ects of di€erent warm-up pro-
tocols on peak power output during a single response jump
task. Med Sci Sports Exerc 28 [Suppl. 5]: S189
Stuart DS, Lingley MD, Grange RW, Houston ME (1988) Myosin
light chain phosphorylation and contractile performance of
human skeletal muscle. Can J Physiol Pharmacol 66: 49±54
Sweeney HL, Bowman FB, Stull JT (1993) Myosin light chain
phosphorylation in vertebrate striated muscle: regulation and
function. Am J Physiol 264: C1085±C1095
Tesch PA, Wright JE (1983) Recovery from short term intense
exercise: its relation to capillary supply and blood lactate con-
centration. Eur J Appl Physiol 52: 98±103
Thorstensson A, Karlsson J (1976) Fatigability and ®bre composi-
tion of human skeletal muscle. Acta Physiol Scand 98: 318±322
Vandenboom R, Grange RW, Houston ME (1993) Threshold for
force potentiation associated with skeletal myosin phosphory-
lation. Am J Physiol 265: C1456±C1462
Vandenboom R, Grange RW, Houston ME (1995) Myosin phos-
phorylation enhances rate of force development in fast-twitch
skeletal muscle. Am J Physiol 268: C596±C603
Vandenboom R, Xeni J, Bestic NM, Houston ME (1997) Increased
force development rates of fatigued mouse skeletal muscle are
graded to myosin light chain phosphate content. Am J Physiol
272: R1980±R1984
Young WB, Jenner A, Griths K (1998) Acute enhancement of
power performance from heavy load squats. J Strength Cond
Res 12: 82±84
Zehr EP, Sale DG (1994) Ballistic movement: muscle activation
and neuromuscular adaptation. Can J Appl Physiol 19: 363±378