Neuromuscular Adaptations to Unilateral vs. Bilateral Strength Training in Women

Abstract

Botton, CE, Radaelli, R, Wilhelm, EN, Rech, A, Brown, LE, and Pinto, RS. Neuromuscular adaptations to unilateral vs. bilateral strength training in women. J Strength Cond Res 30(7): 1924–1932, 2016—Considering the bilateral deficit, the sum of forces produced by each limb in a unilateral condition is generally greater than that produced by them in a bilateral condition. Therefore, it can be speculated that performing unilateral strength exercises may allow greater training workloads and subsequently greater neuromuscular adaptations when compared with bilateral training. Hence, the purpose of this study was to compare neuromuscular adaptations with unilateral vs. bilateral training in the knee extensor muscles. Forty-three recreationally active young women were allocated to a control, unilateral (UG) or bilateral (BG) training group, which performed 2 times strength training sessions a week for 12 weeks. Knee extension one repetition maximum (1RM), maximal isometric strength, muscle electrical activity, and muscle thickness were obtained before and after the study period. Muscle strength was measured in unilateral (right + left) and bilateral tests. Both UG and BG increased similarly their unilateral 1RM (33.3 ± 14.3% vs. 24.6 ± 11.9%, respectively), bilateral 1RM (20.3 ± 6.8% vs. 28.5 ± 12.3%, respectively), and isometric strength (14.7 ± 11.3% vs. 13.1 ± 12.5%, respectively). The UG demonstrated greater unilateral isometric strength increase than the BG (21.4 ± 10.5% vs. 10.3 ± 11.1%, respectively) and only the UG increased muscle electrical activity. Muscle thickness increased similarly for both training groups. Neither group exhibited pretesting 1RM bilateral deficit values, but at post-testing, UG showed a significant bilateral deficit (−6.5 ± 7.8%) whereas BG showed a significant bilateral facilitation (5.9 ± 9.0%). Thus, performing unilateral or bilateral exercises was not a decisive factor for improving morphological adaptations and bilateral muscle strength in untrained women. Unilateral training, however, potentiate unilateral specific strength gains.

Introduction

Strength exercises are often performed in 2 different fashions, with 1 limb at a time (unilateral) or both limbs simultaneously (bilateral). Previous studies have shown that it is possible to produce greater force in the sum of unilateral compared with bilateral (^2,8). The mechanisms explaining differences between these conditions are not completely understood. However, the most consistent explanation is that there is a neural limitation during bilateral exercise, blunting maximal force production (^16,23). This phenomenon is known as the bilateral deficit and is well documented by cross-sectional studies in different muscle groups (²¹), populations, (¹⁴) and test conditions (²).

Because of the bilateral deficit, it is possible to use greater loads when exercise is performed unilaterally, which may be a strategy to optimize strength gains. However, a limited number of studies have compared the chronic neuromuscular adaptations of unilateral and bilateral strength training (^5,11,21,22). Some previous studies (^5,11) have demonstrated that despite either unilateral or bilateral training, there are strength gains in the specific trained condition, and an increase in the untrained condition. In contrast, Taniguchi (²¹) showed that strength gains occurred only after bilateral training.

A cross-sectional study reported that subjects who trained at least 1 year unilaterally demonstrated bilateral deficit on the knee extensor muscles compared with subjects bilaterally trained (⁸). However, longitudinal studies (12–26 weeks) with untrained people have found that unilateral training had a slight nonsignificant increase in the bilateral deficit, possibly due to a similar percentage change in the bilateral and unilateral strength (^5,21). On the other hand, bilateral training resulted in significant decrease of bilateral deficit (^5,11,21) as a consequence of a greater magnitude of increase in bilateral compared with unilateral strength.

The divergent results among few studies found in the literature may be explained by different populations (males, females, or mixed groups), period of training (6–26 weeks), and type of resistance machine used during training (isoinertial or isokinetic). For example, Häkkinen et al. (⁵) investigated the effect of 12 weeks of isoinertial knee extension training on middle-aged and older men and women, while Janzen et al. (¹¹) performed 26 weeks of knee extension training in postmenopausal women. Differently of the aforementioned studies (^5,11), Taniguchi (²¹) performed 6 weeks of isokinetic knee extension training in young men and women. In addition, Kuruganti et al. (¹³) investigated the effect of isokinetic bilateral training on the bilateral deficit, but unilateral training was not studied. Moreover, previous studies (^21,22) have focused primarily on strength measurement while investigating differences between training conditions. Thus, there is a paucity of research regarding unilateral vs. bilateral training differences related to morphological and neural adaptations.

Because unilateral training enables one to exercise with greater load per trained limb, it may be speculated that unilateral strength training would elicit greater skeletal muscle adaptations compared with bilateral exercises. Therefore, the purpose of this study was to investigate neuromuscular adaptations to unilateral vs. bilateral training, and their influence on the bilateral deficit in young women, because the few studies in this area were performed with middle-aged and elderly subjects and nonhomogeneous gender.

Methods

Experimental Approach to the Problem

To compare the effects of unilateral and bilateral training, subjects were required to visit the laboratory on 3 separate days before study (pre) and twice after 12 weeks (post) with 4 days of interval between visits. On the first visit, muscle thickness measurements were taken and subjects were familiarized with the dynamic and isometric strength tests. On the second day, anthropometric data were obtained and subjects performed a maximal isometric strength test with simultaneous collection of the electromyographic (EMG) signal. The maximal dynamic strength was measured on the third visit. Thereafter, subjects were randomized on a control group or training group, which exercised 2 times per week for 12 weeks. Post-testing took place 3 to 5 days after the last training session and was identical to pretesting. Intraclass correlation coefficients were calculated with data collected on day one and 2 for the isometric test; and day one and 3 for dynamic test.

Subjects

Forty-three young women (18–30 years) volunteered for this study. They were randomly assigned to either unilateral group (UG) (n = 14: 24.8 ± 1.4 years; 60.8 ± 6.4 kg; 163.0 ± 6.5 cm), bilateral group (BG) (n = 15: 24.3 ± 3.7 years; 57.0 ± 4.8 kg; 160.2 ± 5.8 cm), or control group (CG) (n = 14: 22.7 ± 2.8 years; 58.0 ± 5.7 kg; 163.6 ± 6.2 cm). Sample size was based on standard deviations and differences between means from Häkkinen et al. (⁵), with an alpha level of 0.05 and power of 80%, and resulted in a minimum of 12 subjects per group.

All subjects had not been involved in a systematic strength training program for at least 3 months before this study, and 6 participants had never practiced strength training before. Participants were not taking medications other than oral contraceptives and were instructed to avoid changes in their diet and recreational physical habits (e.g., sports, jogging, and walking) throughout the study period. Only 6 women did not take oral contraceptives (2 for each group). They were informed of possible risks and discomforts of participation and all gave their written informed consent before any testing. This study was conducted according to the Declaration of Helsinki and all procedures were approved by the Ethics Committee of the Federal University of Rio Grande do Sul.

Procedures

Maximal Dynamic Strength

Knee extension 1 repetition maximum (1RM) tests for bilateral and unilateral (right and left) conditions were assessed on the same test day on the same isoinertial knee extension machine used for training (Taurus, Porto Alegre, RS, Brazil). The same researchers, with identical equipment and subject positioning, conducted all pre- and post-testing. Subjects were carefully familiarized with the testing procedures and performed a warm-up of 10 repetitions of bilateral knee extensions with a light resistance. The bilateral extension was always tested first, followed by the unilateral extension test for the right and left leg in random order. Ten minutes of interval was given between the bilateral and unilateral tests. The cadence was fixed at 2 seconds for the concentric phase and 2 seconds for the eccentric phase with the aid of an analog metronome (Quartz, Los Angeles, CA, USA). Subjects started the test at 90° of knee flexion (0° = knee fully extended) and moved to full extension which was individualized for each subject and controlled by a delimiter device. The maximal load was found within 3 attempts for each testing condition, with 5-minute interval between trials. Intraclass correlation coefficients for bilateral and unilateral right and left were 0.96, 0.95, and 0.93, respectively.

Maximal Isometric Strength

Isometric knee extension test for bilateral and unilateral (right and left) conditions were performed on an isokinetic dynamometer (Cybex Norm, Ronkonkoma, NY, USA), calibrated according to the manufacturer's instructions. The unilateral tests were performed with the standard lever arm supplied by the manufacturer, whereas a custom made lever arm was used for bilateral testing (Figure 1). Subjects were seated with their hips flexed at 85° (0° = hip fully extended). For both tests, the dynamometer's axis of rotation was aligned with the lateral femoral condyle of the tested limb, while straps secured the torso and pelvis. Subjects performed 10 repetitions of concentric knee extension at 120° per second as a warm-up. Thereafter, they completed two 5-second knee extension maximal isometric actions at a 60° angle (0° = knee fully extended), with 3 minutes rest between attempts. All isometric tests on bilateral and unilateral (right and left) conditions were performed on the same day in randomized order with 30 minutes rest between bilateral and unilateral tests. Subjects were instructed and verbally encouraged to produce maximal torque throughout each trial. Peak torque values were determined by the dynamometer software (Humac norm 2009, Version 9.6.2). The greatest isometric peak torque value for each condition was used for further analyses. Intraclass correlation coefficients for bilateral and unilateral right and left were 0.78, 0.86, and 0.76, respectively.

Figure 1.: Modified lever arm to perform bilateral testing on the dynamometer.

Muscle Electrical Activity

Electromyographic (EMG) activity was recorded from the vastus lateralis (VL) and rectus femoris (RF) muscles of the right and left limbs during maximal isometric strength testing. Before electrode placement, skin preparation was performed, including shaving excess hair and cleaning the skin with isopropyl alcohol (to reduce impedance below 2,000 kΩ). Bipolar configuration electrodes (20-mm interelectrode distance) were positioned along the estimated direction of the muscle fibers on the muscle belly according to SENIAM (www.seniam.org). Electrode position was carefully mapped using a transparent sheet to ensure replication positioning at post-testing.

Electromyographic signals were recorded using an electromyographic system (Miotool, Miotec-Equipamentos Biomédicos), amplified by a factor of 100 and digitized at a sampling frequency of 2,000 Hz. The EMG signal was Butterworth filtered using cut-off frequencies of 20 Hz and 500 Hz for lower and upper band-pass respectively. After filtering, the EMG signal from the highest isometric torque (Nm) attempt was selected across a one-second torque–time curve plateau, and the root mean square (RMS) value for each muscle was calculated. For total muscle electrical activation (EMG_T), the RMS values of each muscle of each limb were summed.

Muscle Thickness

Quadriceps femoris muscle thickness was obtained using a B-Mode ultrasonographic apparatus (Nemio XG, Toshiba, Japan), with a 7.5 MHz linear array probe (38-mm length). Before muscle thickness evaluation, each subject rested for 15 minutes in a supine position with their legs extended and relaxed to allow fluid shifts to occur (¹). The probe was coated with a water-soluble transmission gel to provide acoustic contact without depressing the dermal surface. Great care was taken to apply minimal pressure during scanning to avoid compression of the muscles. Muscle thickness of the RF, VL, vastus medialis (VM), and vastus intermedius (VI) were measured at the same sites described in previous studies (^4,12,15,19). All images were digitized and later analyzed in Image-J software (National Institutes of Health, version 1.37; USA). Subcutaneous adipose tissue-muscle interface and the muscle-bone interface were identified in each image, and the distance between them was accepted as muscle thickness. Overall quadriceps femoris muscle thickness (MT QUA_sum) was calculated from the sum of the four muscles (RF + VL + VM + VI) (^3,17,19) for each limb followed by the sum of the right and left. Post-testing measures were performed 3–5 days after the last training session to avoid any potential exercise-induced swelling. The same researcher performed all measurements before and after training. These measurements have demonstrated high reproducibility in our laboratory and in previous studies (^17,19).

Bilateral Index

Bilateral index (BI) was calculated using an equation proposed previously (⁸), with peak torque values (BI_PT), 1RM values (BI_1RM) and RMS values (BI_EMG) obtained in the bilateral and unilateral (right and left) conditions. The equation is:

Positive values indicate that bilateral condition was greater than the sum of the unilateral conditions (bilateral facilitation), and negative values indicate that bilateral condition was less than the sum of the unilateral conditions (bilateral deficit).

Strength Training Program

Training occurred across 12 weeks, consisting of 2 sessions per week on nonconsecutive days (total of 24 training sessions). Rest between sessions normally ranged from 48 to 72 hours within a week. The UG performed the knee extension exercise with 1 leg at a time whereas BG performed with 2 legs simultaneously. All training sessions were monitored and supervised by at least 1 experienced investigator. Apart from the knee extension, subjects performed the following exercises: bilateral knee flexion, bench press, lateral pull-down, hip abduction, hip adduction, crunch, biceps flexion, and triceps extension. These exercises were performed in the same way by both groups and were included to keep participants interested and motivated. Participants started all training sessions by the knee extension exercise.

Training intensity was controlled using the repetition maximum (RM) method as in previous studies (^6,18,19), thus the heaviest possible weight was used for the designated number of repetitions for each condition. The intensity was the same for both training groups. Training for weeks 1–3 was 2 sets of 12–15RM; weeks 4–6 was 3 sets of 9–12RM; weeks 7–9 was 3 sets of 7–10RM; and weeks 10–12 was 4 sets of 5–8RM. Interset rest interval was 1 minute for weeks 1–3, 2 minutes for weeks 4–9 and 3 minutes for weeks 10–12. When subjects were able to perform more than the desired number of repetitions, the load was increased for the next session in increments from 1.0 to 5.0 kg. All subjects attended more than 80% of the training sessions.

Training Loads

To compare differences in knee extension training loads between groups, the load used during the last training session of each mesocycle was considered for analysis. Loads were expressed as an absolute value (kg). Loads of each limb for UG were summed for comparison.

Statistical Analyses

All data are presented as means ± SD. Normality, homogeneity, and sphericity for outcome measures were tested using the Shapiro–Wilk, Levene, and Mauchly tests, respectively. Two-way analysis of variance (ANOVA) with repeated measures was used to examine interactions between time x group for 1RM, isometric peak torque, EMG_T, and MT QUA_sum values. Whenever a significant interaction was observed, a paired t-test was used to determine within-group differences, and a one-way ANOVA to determine between-group differences for unilateral and bilateral conditions. Analysis of variance also examined interactions between group × condition for percent change values for 1RM, isometric peak torque, and EMG_T measures. Whenever a significant interaction was observed, a paired t-test was used to determine within-group differences, and a one-way ANOVA to determine between-group differences. A one-sample t-test was used to determine if bilateral index was significantly different from zero at baseline and post-testing for BI_1RM, BI_PT and BI_EMG. A 2-way (time × group) ANOVA with repeated measures was used to determine differences in the time course of workloads. Tukey post hoc tests were used when necessary, to verify differences between groups. Test-retest reliability of force measurements (dynamic and isometric) were calculated at baseline for typical error = standard-deviation of the difference between day 1 and day 2 measurements/√ 2; and coefficient of variation (%) = (typical error of the difference between day 1 and day 2/means of day 1 and day 2) * 100 (⁷). Significance level was set a priori at p ≤ 0.05. All statistical procedures were performed using the Statistical Package for the Social Sciences (SPSS) version 18.0 software (IBM SPSS, Inc., Chicago, IL, USA).

Results

There were no significant (p ≥ 0.05) differences between groups at baseline for 1 RM, isometric peak torque, EMG_T, MT QUA_sum, BI_1RM, BI_PT, or BI_EMG.

Maximum Strength (Dynamic and Isometric)

There were significant (p ≤ 0.05) time × group interactions for 1RM and isometric peak torque where UG and BG significantly increased after training but CG showed no change. Both UG and BG showed greater bilateral 1RM values than CG at post. However, only UG showed greater values than CG in the unilateral 1RM test at post. For isometric peak torque only UG showed greater values than CG for both test conditions at post (Table 1).

Table 1.: Maximum dynamic and isometric strength values pre- and post-training, in unilateral and bilateral tests between groups (means ± SD).*

The typical error for 1RM measures was 1.44 kg (3.92%), 1.03 kg (5.70%), and 0.90 kg (5.16%) for bilateral, unilateral right and left tests, respectively. Typical error for peak torque measures was 38.50 Nm (10.03%), 15.15 Nm (7.12%), and 20.49 Nm (9.48%) for bilateral, unilateral right and left tests, respectively.

For 1RM percentage changes, there was a significant (p ≤ 0.05) interaction for group × condition. The UG and BG increased unilateral 1RM (33.3 ± 14.3% vs. 24.6 ± 11.9%, respectively) and bilateral 1RM (20.3 ± 6.8% vs. 28.5 ± 12.3%, respectively) similarly and both groups showed significantly (p ≤ 0.05) greater values than the CG in unilateral (0.1 ± 5.5%) and bilateral (0.8 ± 5.0%). Moreover, UG showed a greater (p ≤ 0.05) increase in unilateral than in bilateral whereas the BG was not different. In contrast, for peak torque percentage values, there was only a significant (p ≤ 0.05) main effect for group. For the unilateral test, UG increased peak torque values 21.4 ± 10.5% which was significantly greater (p ≤ 0.05) than the BG (14.7 ± 11.3%). In the bilateral test, groups increased similarly (p ≥ 0.05) (UG 14.7 ± 11.3% vs. BG 13.1 ± 12.5%). Both BG and UG showed significantly (p ≤ 0.05) greater values than the CG in both tests.

Muscle Electrical Activity

There was a significant (p ≤ 0.05) group × time interaction for muscle electrical activity in unilateral only. The UG showed a significant (p ≤ 0.05) increase in maximal EMG activity values at post training in unilateral (Table 2), showing greater (p ≤ 0.05) percentage change (39.9 ± 18.3%) in comparison to the BG (12.0 ± 21.7%) and CG (5.1 ± 21.4%). For bilateral, there was no significant (p ≥ 0.05) difference between UG (25.4 ± 18.0%), BG (15.7 ± 31.8%), and CG (6.9 ± 20.7%). The BG and CG showed no significant increase (p ≥ 0.05) on muscle electrical activity at post training for unilateral and bilateral conditions.

Table 2.: Total electromyographic activation values, pre- and post-training, in unilateral and bilateral test condition (means ± SD).*

Muscle Thickness

A significant (p ≤ 0.05) group × time interaction was observed for muscle thickness. The training groups significantly (p ≤ 0.05) increased MT QUA_sum at post, while the CG significantly (p ≤ 0.05) decreased it (Table 3). There was no difference on muscle thickness between training groups (p ≥ 0.05) at post training.

Table 3.: Sum of muscle thickness values of quadriceps muscles, pre- and post-training (means ± SD).*

Bilateral Index

At baseline, there were no significant (p ≥ 0.05) BI_1RM values for any group. At post training BI_1RM values of the training groups were significantly different from zero (p ≤ 0.05), indicating bilateral deficit and bilateral facilitation for UG and BG, respectively (Table 4). For BI_PT, all groups showed values significantly lower than zero (p ≤ 0.05), and post training values did not show any significant changes. For BI_EMG values, there was a significant difference from zero (p ≤ 0.05) at baseline and post training for both training groups (Table 4). There was large variability of BI_1RM between subjects within the same group (Figure 2).

Table 4.: Percentage values of bilateral index at pre- and post-training (means ± SD).*

Figure 2.: Individual percentage bilateral index with 1 repetition maximum values (BI_1RM) before and after training. UG = unilateral group; BG = bilateral group; CG = control group.

Training Loads

A significant time × group interaction (p ≤ 0.05) for training load was observed. Both training groups significantly increased (p ≤ 0.05) at every mesocycle, but the UG trained with a significantly greater (p ≤ 0.05) loads in the last mesocyle (Figure 3).

Figure 3.: Absolute loads used during the training period (mean ± SD). UG: unilateral group; BG: bilateral group; *Significantly greater than the previous mesocycle (p ≤ 0.05); δ Significantly greater than BG workload (p ≤ 0.05).

Discussion

The aim of this study was to compare the effects that unilateral and bilateral strength training have on muscle strength, neural, and morphological adaptations of the knee extensors in young women after 12 weeks of training. This study investigated the responses in a homogenous population, and is the first to consider only young women for training. In general, we found that the 2 forms of training resulted in increases in quadriceps muscle strength and thickness after training. Strength increases were not restricted to the specific training condition, but interestingly and for the first time, the lateral specificity was shown only for UG in both strength and muscle activation changes. The training loads were similar between groups for almost all training periods, but in the last mesocycle, the UG trained with greater workload than the BG. Moreover, training specificity seems to influence the bilateral index differently when considering isometric or dynamic strength. The BI_1RM seems to be reduced by bilateral training and raised by unilateral training, whereas BI_PT values were not altered as a function of the intervention.

The outcomes of this study relative to maximum strength gains showed that both UG and BG demonstrated significant increases in trained and untrained conditions. In relation to dynamic strength, our results are in agreement with others (^5,11,13) that have found increases in both bilateral and unilateral strength, after knee extension training performed bilaterally or unilaterally. Even when only bilateral isokinetic training was performed, Kuruganti et al. (¹³) found knee extension strength increases in both conditions for men and women. However, our results do not support some previous findings (²¹), that demonstrated increases in isokinetic knee extensor strength in men and women only for BG in the specific trained condition. These authors conclude that one of the reasons for this may be the lower initial physical activity level of subjects included in the BG, thus demonstrating greater potential for adaptation. Participants in both training groups of this study had not being involved with strength training for a similar time (10.2 ± 6 months for UG and 13 ± 11.6 months for BG) between training groups and 2 participants in each group had never undertook strength training routines.

Although the 2 groups increased strength in both test conditions, Häkkinen et al. (⁵) and Janzen et al. (¹¹) showed a greater percentage change for each group in their specific trained condition. Our results partially corroborate with these findings, since only UG showed greater 1RM gains in the specific trained condition. The BG strength changes were similar between bilateral and unilateral tests, suggesting no lateral specificity. It is possible that differences in study populations, level of physical activity of the subjects, selection criteria and randomization of the sample may have influenced these study differences (^5,11). To our knowledge this study is the first that studied only young women. Häkkinen et al. (⁵) investigated adults and older, men and women, while Janzen et al. (¹¹) studied post-menopausal women. From the results found in our work, we could suggest that there is no restriction in strength gains only in the trained condition, but that positive lateral specificity seems to occur with unilateral training. Unlike the results found for 1RM, maximum isometric strength gains in the unilateral test condition were greater for UG than BG. All groups began training with greater peak torque values in unilateral than bilateral, different than 1RM pre values. Therefore, UG training maximized unilateral strength gains without hampering bilateral adaptations.

The specificity of isometric strength gains for UG can be explained by an increase in quadriceps muscle activation. The percentage of increase was greater for UG in the unilateral test than the BG, which was similar to that observed in strength changes. Only UG showed significant increases in muscle activation in their specific training condition. Muscle activation was not measured during maximum dynamic tests which was the specific training condition. Therefore, we cannot rule out neural mechanisms involved with maximal isoinertial strength, although it can be speculated that a transfer of strength and muscular activation gains may have occurred in the nonspecific test (isometric) condition.

Regarding morphological responses, few studies have compared the effects of unilateral and bilateral training on skeletal muscle adaptations (^5,11,13). Some previous investigations have hypothesized that unilateral training might allow the use of heavier loads, resulting in greater gains than bilateral training (¹¹). In our current study, both BG and UG significantly increased muscle thickness similarly which corroborates previous studies (^5,11) examining knee extensor muscles. Furthermore, the absolute training loads were greater in UG than BG only in the last mesocycle; so it is possible to suggest that for longer periods of training, the difference between loads will be more apparent and could optimize gains from unilateral training. In a 26-week study by Janzen et al. (¹¹), UG trained with heavier loads than BG at 3 time points (pre, mid, and post-training). However, the authors performed their analysis with percent loads relative to baseline 1RM values. Moreover, the UG began training with greater loads.

The bilateral index in the present study was measured for both 1RM and isometric peak torque. For BI_PT, both training groups presented a bilateral deficit at baseline, however neither group showed bilateral deficit or facilitation before training in BI_1RM. According to Jakobi e Chilibeck (¹⁰), isometric muscle actions can best represent the bilateral deficit, because they are more stable than dynamic actions. In addition, it is possible that during the isoinertial 1RM test, the difference between loads for unilateral and bilateral is not pronounced. Accordingly, Janzen et al. (¹¹) and Häkkinen et al. (⁵) did not find significant knee extension 1RM bilateral deficit at pretraining using a isoinertial equipment. These results highlight the importance of considering the specific test designed to measure bilateral deficit. Hence, it seems that bilateral deficit occurs in maximal isometric and isokinetic tests (²), but may be limited when derived from isoinertial resistance equipment (^5,11), suggesting an action-specific behavior.

Performing bilateral strength exercises has been suggested to reduce the BI (^5,11,21,22) because of the greater magnitude of increase in bilateral strength, whereas unilateral training may increase the bilateral deficit according to a cross-sectional study (⁸). Our findings regarding BI_1RM are in accordance with this specificity concept, because the BG showed bilateral facilitation, whereas bilateral deficit was evident after training in UG. However, there were no changes in the bilateral deficit with isometric peak torque values. Since BI_PT was large for all groups at pre, it is possible that the 12 weeks of training was not enough to generate measurable BI_PT changes. In addition, the isometric test differs from specific training condition.

Häkkinen et al. (⁵) reported that bilateral training significantly increased the ratio of bilateral and unilateral force by 7% after 12 weeks of strength training in the knee extensors of adults and elderly men and women, while unilateral training showed a not significant 2% change in 1RM. In the present study, the percent change of the UG was approximately 10%, whereas BG showed an increase of approximately 3%. These divergent results between studies could be related to variability in the BI values ​​among individuals of the same group, which can be observed in Figure 2. Moreover in a previous study by Hakkinen et al. (⁵), there is no information about individual behavior of the subjects. It is possible that none of the subjects had a bilateral deficit before training, unlike our investigation where some subjects showed a bilateral deficit or bilateral facilitation at baseline.

Considering individual variability in the BI, UG had uniform changes in the BI_1RM ​​after training, showing a reduction in variability, unlike the heterogeneous response observed in BG. It is possible that less between-subjects variability in unilateral adaptations might have resulted in the lateral specificity found after unilateral training. Moreover, Häkkinen et al. (⁵) and Janzen et al. (¹¹) did not show any effect of unilateral training on BI, probably because the percentage of change of unilateral and bilateral strength were similar for UG but not BG, while in present study the BG had percentage values more similar between conditions that resulted in only 3% of change in the BI.

The bilateral deficit is probably a result of neural limitations (¹⁶), so it would be expected that lower torque values in bilateral tests were accompanied by lower levels of activation. The bilateral index with muscle activation values was performed only on data obtained during the maximal isometric test. At pre, all groups showed ​bilateral deficit for muscle activation in accordance with the results found in BI_PT. At post, both training groups still demonstrated a bilateral deficit, while the CG demonstrated none. Other cross-sectional studies have not find proportional lower values between torque and muscle activation (^8,9,20).

It should be noted as potential limitation of this study that it is not possible to extrapolate our results to other muscle groups and different exercises because only the knee extension movement was investigated. Furthermore, the current study was conducted on young women, but no control for the menstrual cycle variations was done. Thus, the generalization of the current findings should be done with caution to other populations and other strength exercises.

Practical Applications

The use of unilateral or bilateral exercises does not seem to be decisive in improving neuromuscular adaptations to strength training in untrained young women. However, if the training aim is to optimize the increase in force produced for each lower limb separately, the unilateral training may be recommended. The increase of BD post unilateral training means that force in each limb was maximized and more force is produced for each one separately than simultaneously. This may imply in improved performance in daily activities and in athletic performance during unilateral movements, such as kicking and running. Although unilateral training enabled the use of heavier training loads, there were no greater superior gains in muscle mass compared with bilateral training for 12 weeks. It is worth noting, however, that longer training periods may have resulted in greater neuromuscular adaptations in the UG, since differences in training loads were only evident in the last weeks of training.