Effect of Verbal Instruction on Muscle Activity During the Bench Press Exercise

Abstract

Snyder, BJ and Fry, WR. Effect of verbal instruction on muscle activity during the bench press exercise. J Strength Cond Res 26(9): 2394–2400, 2012—Recent research suggests that humans have some ability to selectively activate or relax some muscles during isometric or dynamic muscle actions without changing posture or position. This study sought to reveal whether trained athletes could isolate either the pectoral or triceps muscles, respectively, at different intensities when given verbal technique instruction. Eleven male Division III football players performed 3 sets of bench press at 50% 1-repetition max (1RM) and 80% 1RM while electromyographic (EMG) activity was recorded from the pectoralis major (PM), anterior deltoid (AD), and triceps brachii (TB). In the first set, the subjects performed the exercise without instruction. In the second set, the subjects were given verbal instructions to use only chest muscles. In the third set, the subjects were instructed to use only triceps muscles. Mean normalized root mean square EMG activity was calculated during 3 repetitions in each condition. Repeated-measures analysis of variance was used to detect differences from the preinstruction condition, with significance set to p ≤ 0.017 as indicated by a Bonferroni correction for multiple comparisons. During the 50% max lift with verbal instructions to focus on chest muscles, PM EMG activity increased by 22% over preinstruction activity (p = 0.005), whereas AD and TB activities were statistically unchanged. When the subjects were instructed to focus on only the triceps muscles, PM returned to baseline activity, whereas TB activity was increased by 26% (p = 0.005). When the lift was increased to 80% max, PM and AD activities were both increased with verbal instructions to use only chest muscles. The TB activity was unchanged during the 80% lifts, regardless of instructions. In conclusion, it is found that verbal technique instruction is effective in shifting muscle activity during a basic lift, but it may be less effective at higher intensities.

Introduction

Practitioners from a variety of fields have an interest in which muscles are involved in various activities, to what level they are activated, and how their involvement might be emphasized or perhaps even deemphasized. For example, physical therapists instruct patients on a certain exercise to activate a specific muscle or muscle group but may be unsure whether the proper muscles are activated or how their instructions affect the level of muscle activity. Occupational therapists seek to understand how to minimize fatigue during repetitive motor tasks and what effect verbal instructions may have. Strength and conditioning professionals give technique instruction to athletes and clients with the expectation that it will lead to specific muscle activation patterns but may not have a direct knowledge of the actual effects on muscle activity. Having this knowledge could prove valuable for practitioners hoping to enhance their training exercises and validate the efficacy of their instructions to patients and clients. Numerous studies have investigated the effect of alterations to motor tasks to increase or decrease the activity level of involved muscle, most of which involved significant changes to the exercise such as grip (^14–16), limb or body position (^6,8,25), torso stability (see Behm et al. ^[2] for a review), and range of motion (^5,20). Some investigators, however, have observed that rather than changing the exercise, a given exercise might instead be optimized for a selected muscle or muscle group, without any changes to the exercise itself. To this end, verbal instructions and various forms of biofeedback, often in conjunction, have been found to be effective (^{10,12,18,19,23}).

Verbal instruction alone has also been shown to have a measurable effect on muscle activity during a variety of activities without any changes to posture, form, or range of motion of the exercise. Trainers, coaches, and therapists often use verbal instruction during resistance training activities with their clients, and accurate instructions can help these clients to most efficiently activate and train targeted muscles groups while maintaining joint stability and avoiding injury. A number of studies have illustrated the power and specificity of verbal instructions. Sahaly et al. (²²) showed that for a variety of upper- and lower-body single-joint isometric contractions subjects developed higher electromyographic (EMG) activity and had a greater rate of force development when instructed to produce their most “explosive” contraction than when instructed to produce their maximal force as “hard and as fast as possible.” This indicated that even a subtle shift in the subject's focus during the activity could have a significant effect on muscle activity.

Other investigations have shown that verbal instructions can affect the distribution of activity in complementary muscle groups and postural muscles during single-joint isometric and dynamic activities. Specifically, research subjects have exhibited the ability to increase activity in a specified muscle or muscle group and voluntarily reduce activity in another agonist muscle (^12,18,19) without changing the performance of the exercise, although only one of these studies (¹²) showed reduction in muscle activity using only verbal instruction. Furthermore, the study by Palmerud et al. (^18,19) revealed that not all muscles can be voluntarily relaxed, even with the aid of EMG biofeedback.

More recent work has shown that untrained subjects can voluntarily increase the activity of the latissimus dorsi during a low-intensity lat pull-down, a common multijoint dynamic resistance training exercise (²³). However, receiving only verbal instruction, the novice weightlifters in this study did not exhibit the ability to voluntarily reduce activity of targeted muscles below the level seen during normal performance of the exercise, and it is not known whether the same results would be found in trained or experienced individuals or in subjects using higher-intensity loads.

The purpose of this project was to observe the electrical activity of the agonist and antagonist muscles of resistance-trained individuals during a bench press exercise at 50% 1RM and 80% 1RM, before and after verbal instructions to subjects to alter the involvement of specified agonist muscles. It was hypothesized that, in response to verbal instructions, the subjects would exhibit the ability to increase the activity of a specified agonist muscle while also decreasing the activity of a complementary agonist at a 50% 1RM load but that this ability would be limited or absent at an 80% 1RM load.

Methods

Experimental Approach to the Problem

Because of the importance to practitioners of understanding the effects of their verbal instructions to clients and patients, this study seeks to answer 2 questions: (a) “What is the effect of verbal instructions on muscle activity during the bench press exercise in experienced lifters?” (b) “Does increasing the load to near-maximal levels alter subjects' ability to respond to specific instructions to increase or decrease the activity in specific muscles?” Anecdotal evidence suggests that subjects can indeed emphasize a muscle when asked to do so using palpation or mental imagery, but few researchers have investigated this phenomenon experimentally, and usually not in trained subjects. Likewise, it was important to discover whether increasing the load to match typical training loads would abolish the efficacy of verbal instructions, if it existed. Therefore, using a repeated-measures analysis in which the subjects served as their own controls, EMG activity of the agonist and antagonist muscles for the bench press activity was recorded during the bench press under 3 conditions: (a) nonspecific instructions, (b) instructions to emphasize chest muscles, and (c) instructions to emphasize arm muscles. The subjects were not told the nature of the specific instructions because that knowledge may have confounded the results of the nonspecific instructions, so each subject performed the sets in the same order. A small number of repetitions and rest periods between each set were used to minimize the possible confounding effects of fatigue, and EMG power spectrum analysis was used to detect the presence of fatigue.

Subjects

Eleven male Division III football players with at least 6 months of continuous experience with the bench press activity and no recent orthopedic or neurologic issues completed the study. The subjects completed spring drills, including agility and speed drills, and a full body 4 d·wk⁻¹ lifting protocol, but did not perform heavy bench press at the time of the study. All the subjects were carefully informed of the protocol to be followed and advised of their right to withdraw from the study at any time without penalty. The research protocol was approved by the Shenandoah University Human Subjects Review Committee.

Procedures

The general procedure is outlined in Figure 1. The subjects reported to the laboratory on 2 separate occasions. On the first occasion, 1RM bench press max was determined using standard procedures (¹) using a standard bench press (Cybex International, Medway, MA, USA). On the second occasion, the subjects were prepared for electromyographic data collection from 3 agonist muscles—pectoralis major (PM), triceps brachii (TB), and anterior deltoid (AD) and 2 antagonist muscles—biceps brachii (BB) and posterior deltoid (PD). Next, maximal isometric activity of both agonist and antagonist muscles was recorded during a maximal effort bench press at a supramaximal resistance and again during a maximal isometric pull with the subjects lying with their chests down on a freestanding bench. For the maximal isometric bench press, the subjects reclined on a bench underneath a bench press bar suspended on a squat rack (Cybex International) and immobilized by supramaximal resistance. The bar height was altered if necessary so that the subjects' elbows were flexed to approximately 90° when they grasped the bar with their self-selected grip. At this time, grip width was measured and recorded. The subjects were instructed to gradually increase their pushing effort up to maximal during a verbal 3-second count by the investigator and then to sustain a maximal effort during the next 3 seconds. The subjects then lay down on their chests on a separate bench underneath which was an Olympic bar with a supramaximal weight. The bar was aligned with the subject's sternal angle and the maximal EMG activity was recorded during a maximal pull using the same procedure just outlined. The subjects were verbally encouraged during both isometric efforts.

Figure 1: Outline of methodology.

During a 2-minute rest, the subjects were informed that they would perform 6 sets of 3 repetitions of the exercise alternating between 50% 1RM and 80% 1RM. To prevent corruption of the data, the subjects were not told that they would receive new instructions for each set. After positioning on the bench, grip width was matched to that used during the maximal isometric push to ensure consistent grip width throughout the entire procedure. The subjects then performed a set of 3 repetitions of bench press at 50% 1RM and again at 80% 1RM with 30 seconds of rest between sets to change the weights. During this set, the subjects were instructed only to perform each repetition at a pace of 2 seconds down and 2 seconds up and to maintain proper form and full range of motion at all times. At no time were the subjects given any verbal encouragement during the performance of the exercise. After a 3-minute rest period, the subjects performed another 2 sets of 3 repetitions at 50% 1RM and 80% 1RM (separated by 30 seconds), this time with verbal instructions designed to isolate the horizontal adductors as follows: “During this set, try to use only your chest muscles, and not your arm muscles. To do this, attempt to push your hands together, while still maintaining your grip on the bar.” The second half of the phrasing was chosen based on pilot data indicating that the subjects needed an additional mental imagery to successfully alter muscle activity. Clarification of these instructions was given only on specific subject request. After another 3-minute rest period, the subjects were given the opposite instructions—to attempt to use only arm muscles and not chest muscles to complete the lift. This time the verbal cue was “During this set, try to use only your arm muscles, and not your chest muscles. To do this, attempt to push your hands apart, while still maintaining your grip on the bar.” At all times, the speed of movement and range of motion were closely monitored by the investigators, and subjects were immediately corrected if any variance from instructions was noted. Repetitions not performed through the full range of motion or at the prescribed speed were noted and eliminated from the analysis, because movement speed and range of motion have been shown to affect EMG amplitude (^3,13).

Electromyographic Data Collection and Analysis

Electromyographic data were recorded using an 8-channel Delsys Myomonitor IV (Boston, MA, USA). Electrode placement followed recommendations of the Surface Electromyography for Non-Invasive Assessment of Muscles (SENIAM) project (⁹) Electrodes were placed while the subject was in a standing position. For the TB, the electrode was placed 2 cm medial to and halfway along the line between the posterior aspect of the acromion process and the olecranon process. For the PM, the electrode was placed on the halfway point of a line between the sternal angle and the anterior aspect of the acromion process. The electrode was oriented parallel to the ground, along the expected parallel path of the muscles fibers at that point. For the AD, the electrode was placed 1 cm distal and anterior from the anterior aspect of the acromion process and oriented to be parallel along the line between that location and the attachment point of the deltoid on the deltoid tuberosity. For the biceps brachii (BB), the electrode was placed one-third of the way proximal to the antecubital fossa, on a line to the acromion process, parallel to that line. For the PD, the electrode was placed 2 cm behind the angle of the acromion and oriented to be parallel along the line between that location and the attachment point of the deltoid on the deltoid tuberosity. The single reference electrode was placed over the spinal process of the seventh cervical vertebra. Before testing, the skin surface over the center of the muscle belly was shaved, rubbed with light abrasive paper, and cleaned with an alcohol pad to optimize the strength of the EMG signal. Polycarbonate-enclosed single-differential 99.9% Ag parallel-bar electrodes with a 10-mm contact interspace were affixed to the skin using double-sided tape. No gel was used, because it has been shown to increase movement artifact with this system (²¹). Once preparation was complete, signal integrity was checked by having the subjects contract against resistance with a movement involving each tested muscle.

The EMG signals were filtered using a bandwith of 20–450 Hz with the gain set to 1,000×. Samples were collected at 1,024 Hz and stored for later analysis. During analysis, a window containing all 3 repetitions was selected, and the average root mean square (RMS) was calculated by the EMGWorks Analysis software (Delsys, Boston, MA, USA) using a 0.30-second moving window with a 0.0625-second overlap. Root mean square values for each muscle were then normalized by dividing activity from the dynamic bench press by the processed signals collected during the maximal isometric efforts. A recent review of EMG normalization methods reports that the use of EMG activity during maximal isometric contractions at an arbitrary joint angle produces interindividual reliability that is at least as high as with other methods (⁴). Intraclass correlation coefficients for surface EMG measurements during maximal isometric, submaximal concentric and submaximal eccentric contractions have been shown to have intratester reliabilities of 0.93, 0.87, and 0.96, respectively (⁷). Power spectrum analysis was carried out on a 1-second period of isometric contraction before dynamic contractions began, for both the first exercise set and the final exercise set. The EMGWorks software's built-in power spectral density analysis carried out a Fast Fourier Transform of the selected data, which was then analyzed to find the mean frequency of the signal distribution.

Statistical Analyses

A one-way repeated measures analysis of variance was used to detect differences from the preinstruction condition, with instructions as the independent variable and EMG activity as the dependent variable (within-subjects comparison). Significance level was set to p ≤ 0.017 as indicated by a Bonferroni correction for multiple comparisons. For comparison of mean EMG frequency in the first exercise set vs. the final exercise set, a paired Student's t test was used, with the level of significance set to p ≤ 0.05.

Results

The % max normalized RMS figures and percent change from the no instructions condition for PM, AD, and TB can be seen in Table 1. During the 50% max lift with verbal instructions to focus on chest muscles, PM EMG activity increased by 22% (effect size = 0.43) over preinstruction activity (p < 0.017), whereas AD and TB activities were statistically unchanged (Figure 2). Both PM and AD activities were also increased at 80% 1RM, by 13.3% (effect size = 0.19) and 17.3% (effect size = 0.26, p < 0.017), respectively, compared with preinstruction activity (Figure 3). When the subjects were instructed to focus on only the triceps muscles at 50% 1RM, PM returned to preinstruction levels, whereas the TB activity was increased by 25.7% (effect size = 0.29, p < 0.017). The TB activity was unchanged during the 80% lifts, regardless of instructions. Under no conditions did the activity in either the PD or the BB change significantly. Because of the need to perform the different exercise sets in the same order, fatigue was a possible confounding factor. However, the subjects did not report fatigue or exhibit excessive effort during exercises at 50 or 80%, and power spectrum analysis did not reveal any statistical difference in the mean frequency of EMG signals between the first and last exercise sets p ≤ 0.05. Thus, no effects of fatigue were detected.

Table 1: Average normalized root mean square EMG activity at 50 and 80% 1RM.*†

Figure 2: Root mean square electromyographic activity as a percent of maximum isometric activity during bench press at 50% max with no instructions, chest-only instructions, and arms-only instructions. Error bars indicate SEE. ^*Indicates a significant difference from No Instructions (p ≤ 0.017).

Figure 3: Root mean square electromyographic activity as a percentage of maximum isometric activity during the bench press at 80% max with no instructions, chest-only instructions and arms-only instructions. Error bars indicate SEE. ^*Indicates significant difference from No Instructions (p ≤ 0.017).

Discussion

The results of this study show that trained subjects can alter the participation of muscles in both moderate and higher-intensity multijoint resistance training exercises in response to verbal instructions, because both TB and PM activities were increased selectively in response to 2 different sets of instructions at 50% 1RM and 80% 1RM. This indicates that verbal instructions from trainers, therapists, and coaches are likely to have a measurable effect on muscle involvement, although it is unclear how generalizable this effect might be to all training exercises. Previous research from our laboratory (²³) indicated that untrained subjects performing a lat pull-down at 30% max isometric load could respond to verbal instructions to increase back muscle involvement by increasing latissimus dorsi activity while maintaining proper form and similar speed of movement. The subjects in that study increased latissimus dorsi activity by 17.6%, whereas in the current study, verbal instruction resulted in a 22.3% increase from baseline at 50% 1RM for PM and a 25.6% increase for TB. However, antagonist activity was not measured by Snyder and Leech (²³), and it was possible that the subjects activated antagonist muscles to offset additional force produced by agonist muscles. This study addressed this possibility, but no changes were seen in antagonist muscle activity with verbal instructions. The question of the effect of higher testing loads was also addressed by this study, and it was found that at 50% 1RM, the subjects were capable of altering muscle participation of both the horizontal adductors and the elbow extensors, but at 80% 1RM, only the horizontal adductors were affected.

We did not see functional ‘isolation’ of these muscles, because activity in the alternate muscle (i.e., triceps activity during chest muscle only instructions) was not decreased as seen in some previous research (^12,18), although these studies also revealed inconsistent control of muscles with instructions. For example, Karst and Willett (¹²) showed that subjects performing an abdominal crunch could voluntarily increase external oblique activity and decrease rectus abdominus activity in response to verbal instructions, but they could not do the opposite when instructed. The authors speculated that this result indicated that the rectus is the dominant muscle in this exercise and therefore cannot be activated to a higher level. Palmerud et al. (¹⁹) explored activity of shoulder muscles during an isometric abduction exercise, to discover how to minimize fatigue and reduce injury during repetitive motor tasks. Following verbal cues, and using EMG biofeedback, the subjects were able to selectively reduce the activity of the upper trapezius to 67% of the value at the onset of a 1 minute submaximal contraction. During this time, rhomboids major and minor and the transverse trapezius muscles increased their activity by up to 232% of original muscle activity. An earlier study by the same author (¹⁸), however, showed that no other shoulder muscles could be relaxed, even with the aid of EMG biofeedback.

In contrast to these studies, the subjects in this study were instructed to focus on increasing the activity of certain muscles but were never instructed specifically to focus on relaxing the alternate muscle. Likewise, the subjects did not use EMG biofeedback to augment their efforts to isolate the requested muscles. These differences may explain the failure of ‘muscle isolation’ in this study.

One possible explanation for greater muscle activity with specific instructions comes from the ‘constrained action hypothesis.’ This theory, posited by Wulf et al. (²⁸) and subsequently supported by later studies (^17,24,26), states that an internal attentional focus hampers the body's automatic control of movements, making them less efficient. Studies investigating this theory have generally found that, during a variety of activities, including golf (²⁷), balance tasks (²⁸), maximal elbow flexion (¹⁷), and vertical jump (²⁶), focusing on the results of muscle activity (i.e., the object being moved) rather than the muscle or limb movements themselves results in generally improved success of a motor task. Some of these studies have included measurements of EMG activity in an attempt to find a mechanism for such changes in performance. In these studies, improved performance with external focus was coupled with decreased EMG. For example, Marchant et al. (¹⁷) showed a small increase in net maximal isokinetic torque during a single-arm bicep curl but decreased muscle activity of the BB using an external focus. Vance et al. (²⁴) compared internal and external focus during a submaximal biceps curl, finding that even when movement speed was controlled, there was a small decrease in the EMG activity of the BB during the flexion phase of a biceps curl. Interestingly, the researchers also measured antagonist EMG activity, finding it unchanged, indicating that neither speed of movement increased coactivation can explain the changes in EMG activity with an external attentional focus. This led the investigators to conclude that internal focus leads to a decreased efficiency of motor unit recruitment. Since the current study essentially asked the subjects to maintain an internal attentional focus, the finding of increased EMG activity with verbal instructions is in agreement with the finding of these studies, although our study involved a multijoint exercise with multiple muscles and included a ‘control’ set without any instructions.

In this study, movement speed was controlled, and no change in antagonist activity was found, so there is a question as to whether any additional force was created by the additional muscle activity with verbal instructions. Certainly, there are many situations in which muscle force and EMG activity are uncoupled, but neither fatigue, altered range of motion, nor altered speed of movement were present during any of the lifts, which would seem to indirectly confirm that decreased motor unit recruitment efficiency played at least some role in the EMG changes with verbal instructions. However, because the subjects explicitly attempted to increase activity of specified muscles, the more obvious explanation may be that verbal instructions can result in enhanced, if somewhat incomplete, voluntary control of muscles without significant changes in force production that would alter the movement. Regardless of a mechanistic explanation of the phenomenon, if increased activity of a specific muscle is the desired result, then an internal focus may be more appropriate than an external focus, whereas if motor task performance is of greater interest, an external attentional focus would be more desirable.

Although this study sheds further light on the importance and efficacy of verbal instructions during resistance training activities, several questions remain unanswered. First, the failure of the subjects to reduce activity of the TB after the ‘chest only’ instructions (or alternately, the failure to reduce PM or AD activity during ‘arms only’ instructions) leaves open the question of how increased activity in a desired muscle can occur without altering some aspect of the movement. Future studies should make some attempt to measure motor unit recruitment efficiency to attempt to answer this question. Previous research indicates that muscle isolation (increased activity in one muscle with concurrent relaxation of other agonist muscles) might only occur if instructions focus on relaxing rather than activating certain muscles, a protocol that has previously been successful (^18,19) with light loads during extended isometric actions in conjunction with EMG biofeedback. If this is possible with normal training loads during dynamic activity, the potential exists to improve performance by reducing fatigue of specific muscles. The EMG biofeedback has been shown to be an effective tool to enhancing control of specific muscles, even to the point of allowing subjects to activate separate compartments within muscles (^10,11). Future research will investigate the effects of both verbal instruction and EMG biofeedback on more complex resistance training activities and, more importantly, certain types of sport skills during athletic competition to discover if fatigue can be delayed by increasing the economy of these movements.

Practical Applications

After verbal instruction, trained subjects exhibit an ability to alter the participation of various muscles during a moderate intensity resistance training exercise, but at higher intensities, that ability may be muscle specific. Although EMG technology is more available than ever, it may not yet be reasonable for various practitioners to use this on a daily basis, so verbal instruction is still the frontline tool for enhancing movement technique. The results of this study confirm that personal trainers can target desired muscles in their client using well-crafted instructions, strength coaches can reasonably expect that their technique instructions help athletes to use the correct combination of muscles in their lifts, and therapists can help patients to target specific muscles in the process of recovery from injury. In higher-intensity lifts and activities, however, athletes, clients, and patients may have a difficulty responding to such instructions, suggesting that practitioners, while continuing to offer instruction, may sometimes have to alter their exercise prescription when technique begins to deteriorate with heavier loads. In addition, practitioners must construct their instructions carefully, because subtle changes in the verbal instructions can have unintended effects on the activity of involved muscles.

Acknowledgments

No external grant support was received for this project.