Free weights are generally preferred over machines by body builders and strength-trained athletes because they are thought to provide a more unstable exercise, requiring a greater recruitment of trunk musculature (5). Machines, however, are easier to use by beginners and require spotters less often than free weight exercise (5). Few studies have compared free weight and machine exercises of similar movement pattern to determine which is optimal for muscle activation.
Free weight squat is one of the most popular exercises used by strength-trained athletes for training leg musculature. The Smith machine allows one to perform a similar type of movement, where the barbell is stabilized in 2 parallel tracks, allowing a more stable exercise. Anderson and Behm (1) compared electromyographic (EMG) activity during squat exercises using free weights and the Smith machine. They found that there was a trend for EMG activity of the trunk musculature to be greater during the free weight squat; however, EMG activity of the quadriceps (i.e., vastus lateralis) was highest during the Smith machine squats. One limitation of this study was that the same absolute weight was used for both exercises. The relative intensity during the most stable condition (i.e., Smith machine squats) would have been low because muscle force production during this movement is superior to the more unstable free weight squat (3). The purpose of the current study was to reassess EMG activity of prime movers and stabilizers during the squat exercise using free weights and the Smith machine where the load lifted was the maximal one could lift over 8 repetitions on each exercise (i.e., loads were set relative to each exercise; therefore, different absolute loads were used). We believe this has greater applicability to actual resistance-training situations when one is selecting training loads to achieve a desired number of repetitions. It was hypothesized that (a) in a stable environment (Smith machine) prime movers (i.e., muscles of the upper leg) activity would be higher and (b) in an unstable environment (free weight) stabilizer (i.e., muscles of the trunk and lower leg) activity would be higher.
Experimental Approach to the Problem
This study was designed to compare electromyographic activity of the legs (i.e., tibialis anterior, gastrocnemius, vastus medialis, vastus lateralis, biceps femoris) and trunk stabilizers (i.e., lumbar erector spinae and rectus abdominus) during free weight and Smith machine squat exercises to determine which exercise was optimal for activating musculature. Resistance was selected for each exercise to achieve 8 maximal repetitions (8RM) to match what a typical strength-trained athlete would select during a training set.
Six healthy participants (3 males and 3 females, 22 ± 1.2 years, 171 ± 12 cm, 71.5 ± 12.7 kg), all with previous strength training experience (2-5 years), were involved in this study. All participants were active in sports (i.e., basketball, track, or squash) and had trained on both free weights and machines. All participants were currently involved in resistance training at least 3 days per week. The study took place from January to March. Subjects were informed of the experimental risks and signed an informed consent document prior to the investigation. The investigation was approved by an Institutional Review Board for use of human subjects.
Each participant performed 8RM of Smith machine squat and free weight squat with the order randomized while EMG activity of 7 muscles was assessed. Testing sessions were at least 3 days apart.
Recording sites for EMG were prepared by shaving the area and wiping with alcohol pads to decrease electrical impedance. Electrodes were placed on the dominant side of the participant's body; this was the right side for all 6 participants. Electrodes were placed two thirds distally between the greater trochanter and the lateral condyle on the vastus medialis and vastus lateralis, mid-belly of the biceps femoris parallel to the quadriceps landmark, distal part of the lateral head of the gastrocnemius, mid-belly of the tibialis anterior, medial rectus abdominus (approximately 6 cm superior and 4 cm lateral of the umbilicus), and lumbar erector spinae (approximately 6 cm lateral to the L1-L2 spinous processes). For the purpose of this study the prime movers were considered the vastus medialis, vastus lateralis, and biceps femoris. The lower-limb stabilizers included the tibialis anterior and lateral gastrocnemius. The trunk stabilizers included the rectus abdominus and the lumbar erector spinae. Ground sites included the tibial tuberosity, iliac crest, and lateral malleolus of the fibula.
The EMG main amplifier unit included single differential electrodes with a bandwidth of 10 Hz to 1,000 Hz. The overall amplification or gain per channel was between 5,000 and 10,000 dB and was set according to individual subjects to maximize the digital range without saturating the signal. The electrodes were Ag/AgCl surface electrodes (2.4 cm × 2.4 cm) spaced 2 cm apart and aligned parallel to muscle fiber orientation. The EMG was recorded as raw EMG (V) and stored in the computer for analysis. The sampling rate was set at 500 Hz for 60 seconds. After data collection, the raw EMG data were used to calculate a mean absolute value (MAV) for each repetition in the 8RM. The EMG signal was not normalized because the experiment was a repeated-measures design comparing within individuals. Testing sessions were at least 3 days apart for the 2 conditions. A permanent marker was used to identify electrode placement on the first testing session so that identical electrode placement could be used during the second testing session. Identical amplification was used for each collection period within individual subjects. Intraclass correlation coefficients for EMG MAV for test-retest reliability (i.e., at least 3 days apart for the 6 individuals) were 0.93 for tibialis anterior, 0.95 for gastrocnemius, 0.80 for vastus lateralis, 0.84 for vastus medialis, 0.82 for biceps femoris, 0.90 for medial rectus abdominus, and 0.67 for erector spinae.
Participants were required to attend 2 pre-experimental exercise testing sessions during which a weight that could be lifted for 8 repetitions during a squat exercise using free weights (i.e., a barbell and weights) and the Smith machine were determined. This process consisted of performing 2 to 3 warm-up sets and then 2 to 3 working sets. The working sets consisted of choosing a weight that the participants thought they could do for 8 repetitions and was adjusted throughout these sets to meet the desired repetitions. Rest intervals were 4 to 5 minutes between sets. The Smith machine was composed of a rack that fully supports a regular Olympic barbell, therefore completely stabilizing the barbell. The barbell can be moved up and down the rack so that the user can perform a variety of exercises including squats. Approximately 1 week later participants were randomly assigned to perform 8 repetitions of squat exercise with free weights or the Smith machine while their EMG activity was recorded. Participants returned within a minimum of 3 days to have their muscle activity measured while performing 8 repetitions on the opposite exercise. We chose to make comparisons between the 2 training modes with the same relative load (i.e., 8RM) rather than the same absolute load because this provides a greater simulation of real-life training practices (i.e., one usually aims for a given number of repetitions, rather than selecting the same absolute load, when training for 2 different exercises). This resulted in a heavier weight used (by 14-23 kg) by each individual on the Smith machine compared to the free weight squat. Each testing session consisted of 2 to 3 warm-up sets with light weight and 1 working set. One repetition was performed prior to the working set with the appropriate weight for 8 repetitions. This allowed the amplification of the EMG to be adjusted to prevent saturated signals. Participants were reminded of the basic squatting technique but were encouraged to use their natural technique they were accustomed to from their previous experiences. Participants were instructed to go to approximately 90 degrees of knee flexion and were given feedback when this was achieved.
A repeated measures analysis of variance (with mode of testing: free weights vs. Smith machine as the factor) for each of the 7 muscle groups was used (Statistica Version 6.0 Chicago, Illinois, USA). Statistical significance for all analyses was set at p ≤ 0.05. Data are reported as means ± standard deviation.
The free weight squat elicited a 34% higher EMG MAV from the gastrocnemius, a 26% higher EMG MAV from the biceps femoris, and a 49% higher EMG MAV from the vastus medialis compared to the Smith machine squat (p < 0.05; Figure 1). The free weight squat also elicited a 25% higher EMG MAV from the vastus lateralis, with each subject having a higher EMG MAV during the free weight squat compared to the Smith machine squat; however, the differences between exercises did not reach a level of statistical significance (p = 0.057). Averaged over all muscle groups, the free weight squat elicited a 43% higher EMG MAV compared to the Smith machine squat (p < 0.05).
This study was designed to have applicability to typical training sessions, where participants did free weight and Smith machine squats with heavy weights for a desired number of repetitions (i.e., 8RM). Contrary to our hypotheses, muscles of the legs (specifically the vastus medialis and biceps femoris) displayed greater EMG activity during the free weight squat compared to the Smith machine squat, whereas there were no differences between exercises for EMG activity of trunk stabilizers. In support of our hypothesis, 1 of the stabilizing muscles of the lower leg (i.e., gastrocnemius) displayed greater EMG activity during the free weight squat.
The pattern of muscle activation during our free weight squat and Smith machine squat was similar to Anderson and Behm (1), who evaluated the same exercises. The knee extensors (i.e., vastus lateralis) and erector spinae displayed a large amount of EMG activity, whereas the abdominal stabilizers, biceps femoris, and plantar flexors displayed relatively less EMG activity. Several important differences were seen between studies regarding activation of specific muscle groups across the 2 different exercises; these differences are most likely a result of the different type of loading used across the 2 studies. These differences are outlined in detail later.
Our finding of a higher biceps femoris and gastrocnemius activity during the free weight squat may be attributed to the increased role that the knee flexors play in stabilizing and supporting the ankle, knee, and hip joints in a more unstable environment. Behm et al. (2) found significantly higher EMG activity of the biceps femoris during an unstable leg extension on a Swiss ball. This may be attributed to the antagonistic role that the biceps femoris plays in relation to the vastus medialis and vastus lateralis. As a result of a muscle contraction, the antagonist may be trying to control the placement of the limb (4). To increase joint stability and stiffness, the antagonist muscle activity increases (6). Behm et al. (2) also suggest that improved balance and motor control may be attributed to increased antagonist activity. In contrast to our results, Anderson and Behm (1) found no significant difference in muscle activity of the biceps femoris between a free weight squat and Smith machine squat. These results are likely a result of differences in design compared to the current study. Most notably, they had participants perform squats with standardized submaximal loads, whereas the current study used a load specific for each exercise for which participants could complete 8RM. We feel this is more applicable to an actual training situation, where one usually selects loads based on a target number of repetitions.
Our finding of a higher gastrocnemius EMG during the free weight squat compared to the Smith machine squat is again different from the results of Anderson and Behm (1), who found similar EMG activity of the plantar flexors (i.e., soleus) across the 2 exercises. Their data, however, showed a trend for EMG activity during free weight squat to be higher. Because the distal attachment of the gastrocnemius crosses the ankle joint and the proximal attachment crosses the knee joint, this muscle probably plays an important stabilizing role during movements such as the squat, which involve both of these joints and would be activated to a greater extent during a more unstable movement (7).
In contrast to Anderson and Behm (1), we did not find the Smith machine squat to be superior for activation of the knee extensors (i.e., vastus lateralis). Our free weight squat elicited higher activity of the vastus medialis and a trend for higher activity of the vastus lateralis (p = 0.057) compared to the Smith machine squat. This result was contrary to our hypothesis, which was based on the fact that one can lift heavier loads during the Smith machine exercise because of the greater stability (3). The higher vastus medialis and vastus lateralis recruitment during the free weight squat may be attributed to the potential increase in the stabilization roles these muscles play during this exercise.
One might postulate that there would be higher activation of trunk musculature during the more unstable free weight exercise compared with the more stable Smith machine exercise (1,5); however, our findings indicate the differences only existed in activation of leg musculature. There were trends for trunk musculature to have higher activation during the free weight squat; however, the differences when compared to Smith machine squat were not significant (Figure 1). Our study is limited by a lack of power because we only had 6 subjects participate in the study. After the study we determined the number of individuals required to reach statistical significance at alpha of 0.05 and power of 80%. For our nonsignificant results, this ranged from 8 required subjects for the vastus lateralis measurement to 29 subjects for the rectus abdominus measurement. Power for comparisons of muscle groups between the 2 exercises was 48% for tibialis anterior, 59% for gastrocnemius, 51% for vastus lateralis, 84% for vastus medialis, 83% for biceps femoris, 50% for rectus abdominus, and 25% for erector spinae. Another limitation of the study is the use of both genders. The small number of subjects did not permit a statistical comparison for differences across genders; however, the differences for activation between the free weight squat and the Smith machine squat were similar between the genders (i.e., an average difference in activation of 47% for males and 37% for females).
We found a 43% higher muscle activation during the free weight squat compared to the Smith machine squat. Activation of the knee extensors and flexors and ankle plantar flexors were higher during free weight squat, whereas activation of the trunk stabilizers was similar across the 2 exercises. This indicates that the free weight squat may be superior to the Smith machine squat for training the major muscle groups of the legs and possibly would result in greater strength development and hypertrophy of these muscle groups with long-term training.
We would like to acknowledge Doug Jacobson and Heather Whelan for their technical help and the participants who volunteered their time for this study.
1. Anderson, K and Behm, DG. Trunk muscle activity increases with unstable squat movements. Can J Appl Physiol
30: 33-45, 2005.
2. Behm, DG, Anderson, K, and Curnew, RS. Muscle force and activation under stable and unstable conditions. J Strength Cond Res
16: 416-422, 2002.
3. Cotterman, ML, Darby, LA, and Skelly, WA. Comparison of muscle force production using the Smith machine and free weights for bench press and squat exercises. J Strength Cond Res
4. De Luca, CJ and Mambrito, B. Voluntary control of motor units in human antagonist muscles: Coactivation and reciprocal activation. J Neurophysiol
58: 525-542, 1987.
5. Haff, G. Roundtable discussion: Machines versus free weights. Strength Cond J
22: 18-30, 2000.
6. Hogan, N. Adaptive control of mechanical impedance by coactivation of antagonist muscles. IEEE Trans Automat Contr
AC29: 681-690, 1984.
7. Sherbondy, PS, Queale, WS, McFarland, EG, Mizuno, Y, and Cosgarea, AJ. Soleus and gastrocnemius
muscle loading decreases anterior tibial translation in anterior cruciate ligament intact and deficient knees. J Knee Surg
16: 152-58, 2003.