Twelve subjects (6 female, 6 male, (mean ± standard deviation) age = 27.6  6.9 years, height = 175.4  11.6 cm, mass = 74.8  18.9 kg) voluntarily participated in this study. None of the subjects had previous strength training experience. None of the subjects were instructed as to whether the bilateral forces should be equivalent to, greater than or less than the unilateral forces. All subjects were familiarized with the testing apparatus in a familiarization session at least 24 hours prior to the test session. Subjects were asked to visit the laboratory three times, once for the familiarization session, and twice for a series of unilateral and bilateral trials. The two testing sessions were randomized. This research was approved by the University of New Brunswick Research Ethics Board.

A custom-built force chair was used to collect data. The subject was asked to sit in the chair and was strapped in via Velcro straps. The ankle was secured to a force transducer (Precision Transducers Ltd., Auckland) and the force data were sampled at 1000 Hz and displayed and analysed using data acquisition software (Windaq version 1.51, Dataq, Ohio, USA). Subjects were told to try to extend their knee as hard as they could by pushing against the ankle strap. The knee angle was set to approximately 85 degrees of knee flexion (0 degrees corresponded to full extension) and the hip angle was set to 90 degrees. The process of stimulating a nerve or muscle produces synchronous firings of units. This does not necessarily produce the same total force as when the muscle units are asynchronously activated by the CNS. Behm et al.(2001) and Van Diëen, et al. (2003) suggested that the use of superimposed multiple stimulation rather than twitch interpolation is more reliable. Therefore, rather than a single stimulus, paired stimuli (also known as a doublet) were used to enhance the resolution of the twitch. An isolated constant current stimulator (Digitimer stimulator model DS7A, Hertfordshire, England) was used to interject a twin rectangular pulse stimulus (doublet, square wave pulse, width, 200 µS; maximum voltage, 400 V) onto the quadriceps through carbon rubber electrodes (Empi, 2” X 4”) with gradually increasing current until there was no longer any increase in the twitch force. A lateral view of the test set-up is shown in Figure 1.

The familiarization session consisted of a screening process to ensure that the subject was healthy and fit. The subject was also asked to review and sign an informed consent form. The subject was then provided with an overview of the twitch interpolation technique, which involves stimulating the quadriceps muscle, to ensure they were comfortable with the procedure.

During the familiarization session, each subject’s supramaximal twitch response for unilateral left, unilateral right and bilateral conditions was determined by placing the stimulating carbon rubber electrodes over the quadriceps muscle and using the isolated constant current stimulator (Digitimer stimulator model DS7A, Hertfordshire, England) to interject a twin rectangular pulse stimulus (duration = 200 S, delay = 30 mS) during rest onto the quadriceps with gradually increasing current until there was no longer any increase in the twitch force. The stimulus, which no longer generates any further force from the quadriceps muscle, is called the “supramaximal twitch” and the subject was instructed that this current (or close to it) would be used for the remainder of the stimulation tests. The supramaximal twitch was then elicited at the beginning of each test session and used for stimulation for that session. The average supramaximal twitch current for unilateral right extensions was 340  119 mA, and for the left extensions, the supramaximal twitch current was 350  128 mA. For bilateral trials, the average supramaximal twitch current used to stimulate both muscles (total current for one lead to two legs) was 549  131 mA. While the current needed for the supramaximal bilateral twitch stimulation was greater than that required for either of the unilateral condition, it was not proportional (i.e. twice as high).

During all testing sessions subjects were allowed to watch a computer screen, which provided feedback regarding the force produced. In addition, subjects were verbally encouraged to ‘push as hard as they could’ throughout the contraction. The testing sessions consisted of either a series of unilateral contractions or a series of bilateral contractions. Half of the subjects performed the unilateral trials on the first test session day and the bilateral trials on the second test session day, and the remainder did the opposite. During the unilateral trials, half of the subjects began testing with their dominant leg first, and the remainder did the opposite. The dominant leg was determined by kicking a ball to the subject and asking the subject to return it. The subject was asked to do this 2 – 3 times. The leg used consistently was defined as the dominant leg. The majority of the subjects were right-legged. The order of the tests and the order of dominant versus non-dominant leg testing were randomized.

Myoelectric signal (MES) data were collected from the agonist (quadriceps) and antagonist (hamstrings) muscles of the active leg throughout the contraction. Disposable silver-silver-chloride surface electrodes (Duotrode, bipolar configuration) were placed on the vastus lateralis of the right and left legs to record surface myoelectric signals during the MVCs. The location of the electrodes was indicated with a semi-permanent marker to ensure that electrode placement was maintained throughout the testing sessions. The electrodes were connected to a high performance electromyography/processor system (Moroz Biomeasurement Systems 4P-AX, Hamilton, ON, Canada). The data were sampled at a rate of 1000 samples per second. The signal was low and high pass filtered at 15 and 700 Hz, amplified, full wave rectified and smoothed (20 ms time constant). Surface MES data were used for analysis of voluntary contractions. During stimulated contractions, the twitch force was used to measure activation.

The experimental set-up (bilateral trials) used is shown in Figure 2. Notice that both legs were secured at the ankle and attached to a force transducer. The stimulating electrodes were placed at the approximate ends of the quadriceps muscle. The ground electrode was placed at the knee and the recording electrodes were placed over the vastus lateralis muscle of the quadriceps and the biceps femoris of the hamstrings. Feedback was provided through a computer screen indicating the amount of force being produced.

During the test sessions strength was measured in two ways; maximum voluntary strength and twitch force. Maximum voluntary strength was measured from the subject exerting as much force as possible using the force transducer. Subjects were instructed to push with the active leg (left, right or both legs) as “hard as you can.”

Twitch force was measured by manually delivering two short, supramaximal twitches (doublet) to the subject’s quadriceps muscle of the active leg during the voluntary effort.

Once the voluntary strength and twitch interpolation tests were complete, subjects were administered two fatigue tests, one without stimulation and the second with stimulation. These tests consisted of maintaining a maximum voluntary isometric contraction of the active leg (left, right, or bilateral) for 30 seconds or until exhaustion. Again, surface electrodes were used to measure the electrical activity of the quadriceps and hamstrings. The subject was then allowed to rest for 30 minutes. Finally, the subjects were administered a second fatigue test during which a supramaximal twitch was delivered to the quadriceps at 5, 15, 25 seconds and then once again immediately after the contraction was completed.

Measures of Bilateral Limb Ratio Using Voluntary Isometric Force

The amount of force deficit or Bilateral Limb Ratio (BLR) was calculated similarly to previous investigations (Ohtsuki 1983)

Figure 1

Measures of Bilateral Limb Ratio Using Myoelectric Signal Activity

To examine the myoelectric signal in the time domain, the raw signals were processed through a root mean square (RMS) calculation over a one-second window, during maximum voluntary contraction. The mean amplitude (A, measured in mV) and the standard deviation about this measure were obtained for each muscle. The RMS and BLR_MES were calculated as follows:

Figure 2

Where A = RMS value of the MES (mV)

‘m(t)’ = raw MES signal at time t

T = window length (1 second)

where BR denotes the bilateral right measure, BL denotes the bilateral left measure, UR denotes the unilateral right measure, and UL denotes the unilateral left measure.

Unilateral and Bilateral Twitch Interpolation Differences

The purpose of the twitch interpolation test was to determine if the nervous system was fully activating the muscle, by comparing the force exerted during a maximum voluntary contraction with the force produced when the muscle was artificially stimulated using current preset to the supramaximal twitch level. Figure 3 illustrates the superimposed twitch and the control twitch for one subject. The two superimposed twitches were elicited at approximately 1.7 and 4.2 seconds (during exertion) with a control twitch elicited at approximately 7.2 seconds (during relaxation). The force amplitudes of the superimposed and control twitches were then used to calculate voluntary activation (VA). The amplitude of the first twitch was approximately 31 N, and amplitude of the second twitch was approximately 28 N. These two twitches were averaged to give the amplitude of the superimposed twitch (29.5N). The amplitude of the control twitch was approximately 226 N.

Figure 3

Using twitch interpolation, voluntary activation (VA) was calculated as in previous studies (Herbert & Gandevia 1996, Jakobi & Chilibeck 2001) as:

Figure 4

For the data displayed in Figure 3 the VA was calculated to be:

Figure 5

The maximum voluntary and twitch forces of each leg under unilateral and bilateral conditions were compared to determine whether neural inhibition was present on either side. In addition, a bilateral deficit was quantified using voluntary activation data as follows (Khodiguian et al. 2003):

Figure 6

A large bilateral deficit would then be indicated by a BLD_VA below 100% and a small bilateral deficit would be indicated by a BLD_VA close to 100%.

Measures of Unilateral and Bilateral Muscle Fatigue - Voluntary Force Contractions

Fatigue was indexed using the mean frequency (MF) of the power spectrum of the MES. The MF was calculated using the frequency spectrum for the vastus lateralis muscles from the right and left legs. A power spectral analysis was performed on a two second window for each muscle. A fast Fourier transformation (Hanning window processing) was performed using 512 ms segments, overlapping each other by half their length (256 ms). The MF was calculated every two seconds over the entire contraction (approximately 30 seconds) as follows:

Figure 7

Where f(f) is the power spectrum of the signal.

Measures of Unilateral and Bilateral Muscle Fatigue - Twitch Force Contractions

Only three of the twelve subjects showed any bilateral deficit (BLR_{Force Subject 7} = 85.84 %, BLR_{Force Subject 11} = 94.89 % and BLR_{ForceSubject 12} = 49.61%). The mean BLR_Force calculated using force, as described by Equation 1, was 104.30 ± 20.42 indicating that, on average, there was no bilateral deficit detected during isometric knee extensions. In addition, a one-sample t-test showed no significant difference from 100%, indicating no bilateral facilitation. The BLR_Force found in this experiment was similar to other studies of isometric knee extensions (Häkkinen et al. 1999, McLean et al. 1998, Schantz et al. 1989).

The BLR_MES measured during the 5-second contractions did not indicate the presence of bilateral deficit (mean BLR_MES: 116.8 ± 38.9%). Similar to the force data, the means were not significantly different from 100 %, indicating no bilateral facilitation.

The current needed to elicit a supramaximal twitch differed between unilateral and bilateral contractions. The mean (± standard deviation) currents for each condition are shown in Table 1. As can be seen the total current needed for the bilateral condition was less than the sum of the unilateral currents. A paired t-test showed that the total unilateral supramaximal twitch current was significantly greater than the bilateral current (p=0.003).

Similar to the BLR_Force and BLR_MES, a BLR was calculated using the total bilateral current over the total unilateral current needed to elicit a supramaximal twitch (BLR_{Supramax. Current}). A BLD was detected in ten of the twelve subjects and the majority of subjects (8) exhibited deficits between 13 and 30%. The mean BLR_{Supramax. Current} was 82.48 ± 13.30% indicating that there is a difference in the twitch response between bilateral and unilateral conditions prior to the contraction.

Voluntary activation measures were all similar (Table 2), with no statistically significant differences between left-right, bilateral-unilateral extensions.

The BLD_VA calculated using voluntary activation data showed that there was no significant difference in bilateral and unilateral voluntary activation levels (BLD_VA = 103.29 ± 4.48 %) during the 5-second contractions. This indicates that, on average, activation levels did not differ between bilateral and unilateral conditions.

While no bilateral deficit was detected in the force, MES or activation data during the 5-second isometric contractions, it was suspected that there might be differences in the fatigue rates between bilateral and unilateral contractions. The fatigue was indexed with the mean frequency calculated every two seconds over the entire fatigue contractions. The amount of decay of the fatigue was the percent drop from the first 2-second segment to the last two-second segment.

Table 3 shows the average mean frequency for the five-second isometric contractions as well as at the beginning and at the end of the fatigue contraction. Columns 1 and 2 of Table 3 show that the average mean frequency is lower for the short, 5-second contractions than the early portion of the 30-second contraction. The average mean frequency (across subjects) during the first two seconds of the fatigue contraction was 62.95 ± 7.58 Hz for the left leg and 63.74 ± 4.46 Hz for the right leg. At the end of the 30-second contraction the average mean frequency for the left leg was 48.89 ± 5.43 Hz and 49.40 ± 5.05 Hz for the right leg. The mean frequency dropped by 21.4 % for the unilateral left knee extension and 22.3 % for the unilateral right knee extension. This drop was statistically significant, with all p-values approximately equal to 0.

There was no statistically significant difference in the average amount of decay in the mean frequency between bilateral and unilateral contractions. The mean frequency dropped by 20% to 25% for all contractions during fatigue contraction.

The 30-second maximum voluntary contraction fatigue test was also conducted with simultaneous muscle stimulation. This was done in order to determine if there were any differences in the amount of voluntary activation between bilateral and unilateral contractions, particularly as a muscle starts to fatigue. As can be seen in Table 4 the average voluntary activation level was found to be approximately 80 % throughout the contraction for both the left and right legs. During the five second contractions, the voluntary activation levels were ~ 91% for the unilateral trials (Table 3). Paired t-tests indicated that the voluntary activation values reported in Table 2 (short contractions) are significantly different from those reported in the first column of Table 4 (early in the fatigue) for all conditions (unilateral and bilateral, p<0.05). This indicates that there was no significant loss in activation during the shorter unilateral contractions whereas there was a marked loss five seconds into the fatigue test. Interestingly, the voluntary activation did not decrease significantly from five to thirty seconds during the fatigue test This suggests that the subjects were either highly motivated to maintain the level of the contraction, or while there was an inability to maximally activate the quadriceps during the contraction, the amount of activation remained relatively constant.

Similar to the unilateral fatigue test, voluntary activation levels were less than those reported in Table 2 during the initial five seconds of the bilateral fatigue contractions, and only slightly during the last twenty-five second (Table 5). The bilateral right leg showed the greatest decrease from 83.1% to 74.1%, however when a paired t-test was conducted, this difference was not found to be statistically significant.

The voluntary activation levels during the five-second bilateral contractions was ~94% (Table 2). Similar to the unilateral contractions, there was a significant overall loss of voluntary activation during the fatigue contractions compared to the shorter five-second contractions. Overall voluntary activation levels after the first five seconds of the fatigue contractions were approximately 10% lower than those found during the five-second contractions.

Voluntary activation levels were compared from the first five seconds of the contraction to the last five seconds of the fatigue contraction to determine if there was a significant drop due to fatigue. It was found that while there were slight decreases in the voluntary activation levels measured, the decreases were not significant. It is possible that while the fatigue protocol was sufficient to elicit a marked reduction in mean frequency, it may not have been long enough to detect a drop in voluntary activation.

Finally, the BLR was calculated using both force and MES data over the duration of the fatigue contraction (30-seconds) and no bilateral deficit was detected (Table 5). This was similar to the lack of bilateral deficit in the shorter (5-seconds) contractions.

Figure 8

Table 1. Mean current needed to elicit supramaximal twitch at rest during unilateral and bilateral conditions.

* - Indicates that the current needed to elicit a supramaximal twitch was less (p<0.05) for bilateral contractions than the total current needed unilaterally (left + right).

Figure 9

Table 2. Mean voluntary activation levels (percent of maximal 100%) during 5-second unilateral and bilateral stimulated contractions.

Figure 10

Table 3. Mean frequency (Hz) during the 5-second and during the first 5 seconds (early) and last 5 seconds (late) portions of the 30-second fatigue contraction. Differences between the late and early portions of the fatigue contraction are indicated in column 4 along with the decrease in mean frequency in column 5.

* - Indicates that the differences between the mean frequencies of the short contraction and the early portion of the long contraction are statistically significant (p<0.05).

^†- Indicates that the decrease in mean frequency was statistically significant (p<0.05). Mean frequencies were calculated during the first two seconds of the contraction (early) and at the end of the thirty-second contraction (late).

Figure 11

Table 4: Mean voluntary activation levels (percent of maximal 100%) at 5, 15, and 25 seconds of the 30-second fatigue contractions. The differences in activation after the unilateral and bilateral fatigue contractions are shown in column 5.

Figure 12

Table 5. BLR and BLR for short (5-second) and long (30-second) contractions. Neither type of contraction exhibited BLD.

Figure 13

Figure 1: Test Apparatus (lateral view). A constant current stimulator was used to stimulate the muscles. MES and force data were collected and processed.

Figure 14

Figure 2: Experimental Set-up: bilateral test session set-up. The stimulating and recording surface electrodes are placed identically as shown in Figure 2. During bilateral contractions, both ankles are secured firmly to the force transducers.

Figure 15

Figure 3: Example of superimposed twitch during 5-second MVC. The two superimposed twitches are at approximately at 1.7 seconds and 4.2 seconds. The control twitch was elicited immediately after the contraction at approximately 7.2 seconds. The amplitude of the control twitch in this example was 226 N.

We would like to thank the subjects for their participation. The study was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Medical Research Fund of New Brunswick (MRFNB).