Background

Early work from Ikai and Fukunaga [1] found that those with smaller muscle cross-sectional areas had less arm strength than those with larger muscle cross-sectional areas. However, the question that remains is whether the exercise-induced change in muscle size contributes to the exercise-induced change in muscle strength. The current model for explaining changes in muscle strength is that neural mechanisms play a role initially followed by larger contributions from muscle hypertrophy.

Historical Context

In the previous century, there seemed to be much skepticism regarding the role that exercise-induced changes in muscle size played with exercise-induced changes in muscle strength. In 1939, Schneider suggested that “Casual observation is sufficient to prove that muscles do not make a similar gain in size [2].” In 1952, Rasch echoed his own skepticism of changes in muscle size and changes in muscle strength (among other things) in a paper titled “The Problem of Muscle Hypertrophy” [3]. In 1963 and 1976, it was written by Morehouse and Miller that “It has not been proved that hypertrophy is necessarily a desirable reaction. Some students are of the opinion that it may be simply a by-product of training, perhaps a noxious one [4, 5].” Following this point in time, the narrative began to be consistently told in a different manner. To illustrate, Brooks and Fahey suggested in 1985 that “Muscles are strengthened by increasing their size and by enhancing the recruitment and firing rates of their motor units. It appears that both of these processes are involved in the adaptive response to resistance exercise [6].”

What happened to change how the story was told? Evidence that muscle hypertrophy may be related to changes in maximum strength was provided early on by Ikai and Fukunaga in 1970 [7]. They had five individuals’ exercise one arm with isometric contractions while the other arm served as a non-exercise control. Following 100 days of exercise, both arms increased strength (trained arm: 115 kp vs. untrained arm: 37 kp) but only the arm that exercised observed a change in muscle size. These authors suggested that increases in muscle cross-sectional area and nerve discharge are the two primary factors leading to exercise-induced increases in strength.

Interestingly, the original work that most textbooks cite as evidence for the time course of changes in muscle strength is the study completed by Moritani and deVries [8]. This was an 8-week study in which five individuals did bicep curls in one arm and had the other arm serve as a non-exercise control. The outcomes of interest were changes in amplitude (measured by surface electromyography [EMG]) and arm circumference. The authors took a downward shift in the EMG amplitude relative to an absolute force measurement as evidence that growth had taken place. Of note, muscle growth was not actually measured. Following 2 weeks of training, there was an increase in strength but the authors reported no downward shift in EMG amplitude relative to absolute force. The authors concluded that this lack of shift indicated that the immediate increase in strength was neural in origin. Following 8 weeks of exercise, both arms saw an increase in strength (trained arm: 21 lb vs. untrained arm: 13 lb) but only the arm that exercised observed a downward shift in the EMG measurement. The conclusion of this study was “After the first 3–5 weeks, muscle hypertrophy becomes the dominant factor in strength gain.”

Though there are numerous studies showing that muscle hypertrophy occurs alongside changes in strength [7, 9, 10], this alone is insufficient to inform us on the importance of muscle growth for increasing strength. One point to consider when evaluating the classic studies used as support for this “neural first, followed by hypertrophy” narrative is what would have occurred had both arms exercised but one was designed to minimize growth? We have been working with a model where we are able to produce changes in strength without producing changes in muscle growth [11, 12]. Though this is not without its own set of limitations, this model may provide some insight into how important a change in muscle size is for a change in muscle strength.

Muscle Growth is not Necessary, Sufficient, or Contributory?

Necessary?

Muscle growth is not necessary for changes in muscle strength. For example, if one arm is trained and the other arm remains untrained, strength will increase in both arms [13]. This is called the “cross-over effect” and is thought to be largely due to neural adaptations because muscle growth does not occur to an appreciable degree in the untrained limb [14]. Individuals can also increase strength (in certain movements) by just imagining the execution of the movement and this is also hypothesized to be the result of neural adaptations [15]. Together, this provides support that muscle growth need not occur for an increase in strength.

Sufficient?

Exercise-induced muscle growth may also not be sufficient for changes in strength. Studies providing support for this theory are beginning to appear in the literature since more researchers have begun investigating alternatives to high-load resistance exercise for increasing muscle size. To illustrate, studies implementing body weight exercise have found increases in muscle size but not necessarily changes in voluntary strength [16, 17]. This is corroborated by two additional studies with very low load resistance exercise where muscle growth appeared to change but strength in the task they were training did not [18, 19]. When considered together, there is strong evidence that muscle growth is not necessary for an increase in strength and some evidence that muscle growth is also not sufficient for an increase in strength.

Contributory?

Our hypothesis states that while both muscle size and strength can increase following exercise, these adaptations are separate and unrelated adaptions. The evidence that muscle growth contributes to strength appears to be largely based on two points: (1) the baseline relationship between muscle size and strength and (2) that the addition of contractile protein should lead to an increase in strength. The first point is reasonable but a baseline correlation does not necessarily reflect what happens in response to exercise. When examining correlations on the change scores, these analyses are completed on groups designed to increase both muscle size and strength and also appear to be primarily correlating the error/random biological variability with muscle size with the error/random biological variability in muscle strength [20]. The second point seems intuitive but there is a surprising lack of evidence showing that the increase in muscle size contributes to an increase in voluntary strength. Currently, the mechanistic importance of muscle growth is largely determined on whether a change in muscle size was detected relative to nothing [21, 22], a non-exercise control group [9, 23], or another group that may have also observed increases in muscle size [24,25,26]. If individuals increased both muscle size and strength, the conclusion is often that muscle growth played a role. If individuals increase strength but not muscle size, the conclusion is often that the mechanism is not growth but primarily neural.

We have tried to better address this question by using an intentional program design to create two separate groups. The first group is more traditional and is designed to increase both muscle size and strength, whereas the second group is designed to only increase strength through repeated practice with heavy singles. Using these designs, we have been successful in creating differences in muscle growth (of varying magnitudes) but the strength adaptation is remarkably similar [11, 12]. As previously noted, this design and these studies are not without their own limitations and these studies certainly need to be interpreted with those in mind. First, there was no comparison to a time-matched non-exercise control group. Though we interpreted our results with the error of the measurement in mind, the time-matched component is needed to fully capture the random error associated with an intervention [27, 28]. Second, we have primarily inferred muscle growth from the B-mode ultrasound and we do not know how findings might change had they been measured through other assessments. Nevertheless, the changes between B-mode ultrasound and magnetic resonance imaging appear to track similarly [29, 30]. Third, we have only studied this to a maximum duration of 8 weeks and have primarily studied this in untrained individuals. We are uncertain whether these findings would be different if they were studied over a duration of months/years or if they would differ in a larger sample of resistance-trained men and women. Nevertheless, this is the same duration and population used by the study commonly cited as support for this idea of “neural first, followed by hypertrophy” [8]. Last, although both conditions complete exercise at a relatively high percentage of their maximum strength level, the load has not been matched between conditions.

Other indirect evidence that changes in muscle size and strength may be separate and unrelated includes the low-load resistance training and detraining literature. A large portion of the low-load resistance training literature finds similar increases in muscle size as high-load resistance training but lower levels of strength in these exercises [31]. The gap in strength difference can often be decreased by having the low-load resistance training group periodically practice heavier lifting [25]. Given that muscle growth is similar but strength is different, this might suggest that muscle growth is not a mechanism of strength gain. This may also be supported by some of the detraining literature. In a paper by Bickel et al., a group of young adults were trained for 16 weeks and then detrained for an 6 additional months [32]. The authors observed that all of the muscle growth that occurred from training was lost with detraining but that the strength gained from training was largely maintained. This study was unique in that the one-repetition maximum was tested monthly to track strength. For a simple movement like the knee extension, this was possibly enough practice with the skill to maintain the strength gain over time. This study does not necessarily provide support against muscle hypertrophy as a mechanism but it is useful for considering alternative hypotheses for the mechanisms of strength gain (given a loss of exercise-induced growth did not appear to impact strength). Based on the aforementioned points, we hypothesize that changes in muscle size and strength are potentially separate and unrelated phenomena; meaning that increasing one does not cause an increase in the other.

If Not Growth, What?

The increase in strength following exercise is typically in proportion to the specificity of the skill and specificity of the load. The greatest change in strength occurs in the movement trained and those training heavier tend to be better at lifting heavier than those who have been training with a lighter load [31]. We hypothesize that these changes in strength may be due to changes within the nervous system and changes at the fiber level that occur independently of changes in muscle size. For example, a recent meta-analysis on fiber-level changes concluded that neural adaptations and mechanisms intrinsic to the fiber itself (independent of growth) may explain the change in strength following exercise [33]. Neural adaptations could be occurring in the primary motor cortex [34], spinal cord [35], and/or via alterations in the alpha motor neuron [36]. Changes intrinsic to the fiber may include a change in the composition of the myosin motors [37], changes in the pattern of calcium release [38], and/or changes in the major components of the excitation contraction coupling process [39].

Conclusions

It is our opinion that the current story for explaining changes in voluntary strength requires revision. We acknowledge that this has been, and continues to be, a difficult question to answer. We have attempted to test the causal relationship by creating studies designed to produce differential results on one dependent variable (muscle size) and test how this in turn impacts the other dependent variable (muscle strength). While we appreciate that there are some limitations to our research, the fact remains that there is still no available evidence supporting the claim that changes in muscle size lead to changes in voluntary strength. Our hope is that through appropriately designed studies, we and others will be able to better address this research question in the years to come.

Reply to Taber et al.

We wish to thank Taber et al. [40] for their thoughtful article and it is our hope that the dialogue within this point/counter point will be constructive for future research. The authors find the question of “Does hypertrophy contribute to strength gain?” less interesting than, “To what extent and under what circumstances does hypertrophy contribute to strength gain?” However, this question assumes that a change in muscle size contributes to a change in muscle strength. This is the same issue with the authors’ proposed hypothetical figure and the majority of literature on this topic. Aside from the theoretical reason for why a change in muscle size should contribute to a change in muscle strength, the authors present no experimental evidence demonstrating that a change in one contributes to a change in the other. Their position is that there is a contributory-causal relationship but are unable to provide the type of evidence required to make that claim, as their argument rests on correlations across time. It cannot be said that a result provides evidence for a claim if little has been done to rule out ways that the claim may be false [41]. The claim must be tested through experiments whereby muscle growth is manipulated across groups to determine the impact of that manipulation on changes in strength.

Consider for a moment if the current paradigm was that changes in strength were not causally related to changes in muscle size and it was up to Taber et al. [40] to convince the field otherwise. Would any of the evidence provided be enough to sway scientific opinion? Would the baseline correlation between muscle cross-sectional area/mass and strength be enough [42,43,44,45]? Would the correlations found on the change scores of muscle size and strength from studies designed to increase both be more convincing [22, 46]? Would correlations from longer term studies on the changes in skinfold thickness and muscle strength be enough [47, 48]? It is hard to imagine another scenario where correlations (between or within), in the absence of experimental evidence, are sufficient to make causal claims about the importance of one variable for another variable [40, 49]. If muscle growth is a meaningful contributor for strength adaptation, then this should be demonstrated through experiments. Until such data are produced, we find it unsatisfactory to claim that long-term increases in muscle strength are driven by hypertrophy.

In conclusion, there is a surprising lack of experimental evidence that muscle growth is a mechanism for changes in muscle strength. This is highlighted by our article and the article of Taber et al. [40]. It is our collective hope that this facilitates positive discussion in the field moving forward. Future work may want to consider mechanisms capable of explaining the observation that the changes in strength following resistance training appear to largely follow the principle of specificity. Potential candidates could be changes in the nervous system or changes at the fiber level that are specific to the contraction history of the muscle.