Journal Logo

BASIC SCIENCES: Original Investigations

Muscular Performance after Concentric and Eccentric Exercise in Trained Men


Author Information
Medicine & Science in Sports & Exercise: October 2006 - Volume 38 - Issue 10 - p 1770-1781
doi: 10.1249/01.mss.0000229568.17284.ab
  • Free


It is thought that a crucial stimulus to increase muscular strength is the regular development of high force. Thus, resistance training has traditionally been carried out using a high external load in both the concentric and eccentric phases of a lift. Although it has been known for a long time that skeletal muscles can develop more force during eccentric actions than during concentric actions (21), Komi and Buskirk (20) were the first to address the potential difference of training with either maximum eccentric or concentric muscle actions in untrained healthy subjects. To our knowledge, all subsequent studies comparing the effects of eccentric and concentric resistance training have used previously untrained subjects (8,12,16-20,26). Untrained humans may be unable to activate their muscles optimally (11,25), and it is a common finding that during the early phases of resistance training, maximum strength increases rapidly and greatly exceeds the hypertrophy of the trained muscle (23). This deviation between increases in strength and muscle cross-sectional area after training is usually attributed to an increased motor neuron excitability and/or firing frequency of motor units, generally dubbed neural adaptations (1). In contrast, resistance-trained subjects are thought to have already developed greater muscular activation, and the increase in the performance after further resistance training is considered to be a result mainly of muscular adaptations (25).

In their classic elbow-flexion study, Komi and Buskirk (20) found that eccentric training improved eccentric strength more than concentric training did and that both groups improved similarly in concentric strength and isometric strength. However, upper-arm circumference increased only after eccentric training. Later studies have confirmed these findings in eccentric strength (12,16-18). However, in terms of gains in maximum concentric strength, the literature is inconclusive in that eccentric training has been reported to promote less (16,18), equal (8,17), and greater (12) increases than those found after concentric training. The changes found in isometric strength likewise differ between studies (8,17-19). On the other hand, most studies find greater hypertrophy after eccentric training than after concentric training as assessed by the cross-sectional area of single cells (17,18) or of whole muscles (12,16,26).

It is possible that the seemingly conflicting results on improvements in strength after eccentric and concentric training could be a result of the initial training status of the experimental subjects. Therefore, the purpose of this study was to examine both structural muscular adaptations and changes in strength and velocity performance after eccentric or concentric resistance training in previously resistance-trained subjects. Our first hypothesis was that neural adaptations would be less prominent in previously trained subjects. As a result, changes in muscular performance would be more closely related to muscular changes. Our second hypothesis was that eccentric training would lead to greater hypertrophy than concentric training and, consequently, would increase strength and velocity performance more than concentric training would. The results of this study may be important for professionals working with resistance-trained individuals.


Approach to the Problem

The duration of the experiment was 16 wk. Initial strength and velocity performances were tested during the first week, and pretraining muscle cross-sectional area and biopsies were taken during the second week. The subjects were subsequently randomized to either concentric training or eccentric training protocols as described in more detail below. All participants exercised two to three times a week over a 12-wk training period. In week 15, the tests of strength and velocity were repeated. Posttraining muscle cross-sectional images and muscle biopsy samples were taken in week 16. Maximum concentric and eccentric strength of the elbow flexors were tested to assess possible differences in the gain of maximum strength. Maximum angular velocity of the elbow joint at different loads relative to the maximum concentric strength was measured to examine for possible adaptations throughout the in vivo load-velocity relationship. Axial cross-sectional computer tomography (CT) images were taken to measure whole-muscle size changes of the elbow-flexor muscles. Muscle biopsies were taken from the biceps brachii muscle to analyze fiber-type proportions and single-cell cross-sectional areas.


Twenty-two healthy, resistance-trained men served as subjects. Fourteen of the subjects were recreationally resistance trained, and eight were athletes competing in track and field or powerlifting at the national elite level. The regional ethics committee for medical research (Health region 1, Oslo, Norway) approved the experiment. The subjects were informed about possible risks and negative effects as well as the procedures and purposes of the study both orally and in writing. Thereafter, each subject signed a written informed consent. None of the subjects reported using anabolic steroids. All of the subjects had performed resistance training for at least 2 yr, and their elbow flexors had been trained regularly at least one session each week for the last 6 months before the start of the study. The subjects were able to lift 26 ± 4 kg (~29% of their body mass) in one-arm elbow flexion (nondominant arm). Four subjects withdrew from the experiment for reasons not related to the study, and one subject dropped out during the training period because he experienced reoccurring pain in the trained arm. The anthropometric data on the remaining 17 subjects are presented in Table 1. Because of individual variation in strength, the subjects were ranked according to the maximum concentric strength (1RM, the one-repetition maximum). The subjects were then matched for strength, and within each pair, one subject was randomly assigned to one of two groups: concentric training (CON, N = 8) or eccentric training (ECC, N = 9) and the other subject to the other group. There were no statistically significant differences between the two groups for any of the dependent variables at baseline, with the exception of the proportion of type I fibers. Because the subjects were ranked and then randomized according to their concentric strength, we assume this difference between groups was a result of chance that may readily occur when a number of different comparisons are carried out.

Anthropometric data for the CON and ECC groups (mean ± SD).

Test and Training Apparatus

A specially designed apparatus for training and testing of elbow flexion was used in the present experiment (24) (Fig. 1). In short, the subject stood upright with the upper arm abducted laterally to the horizontal position. The subject held a handgrip connected by steel wire to a lever arm at a length that was regulated individually to match the length of the lower arm of each subject. The lever arm was positioned parallel to the lower arm of the subject, and the rotary axis of the lever arm was parallel to the elbow joint. The lever arm was connected by steel wire via castors to a load sledge that moved vertically on two steel rails. An angular potentiometer (Spectrol mod 138 (spes), Spectrol Electronics Corporation, CA) was positioned in the rotation axis of the lever arm to measure the angular velocity. A length transducer (Houston Scientific, mod 1850-80, Houston, TX) was secured in the ceiling vertically over the load sledge to measure the position of the sledge. The transducers were connected to a microcomputer (Major, Ultra Tech As, Oslo, Norway) through an amplifier (SE bridge condition unit, mod SE 994, S.E. Laboratories Ltd, Feltham, UK) and a CIO-AD16fr AD-card (ComputerBoards Inc., Mansfield, UK). Data were sampled at 350 Hz. The angles and lengths were recorded in a custom-designed acquisition program made in Visual Basic and then analyzed in MatLab.

The test and training apparatus for the elbow flexion used in the present study.

Training Protocol

The subjects exercised both arms during the training sessions, but only the nondominant arm was tested and analyzed. The training program is given in Table 2 and consisted of 2-3 training sessions per week for 12 wk. The exercise sessions alternated between either maximum (repetition maximum, RM) or medium loads. The maximum load was defined as the greatest load that could be lifted a given number of repetitions and sets (4- to 8RM). The medium training load was set to a value of 85-90% of the maximum load. During any given 2 wk of training, each subject completed three sessions with the maximum load and two sessions with the medium load. As a warm-up before training, all subjects completed two to three sets of 6-12 repetitions of the designated action type while gradually increasing load. In the CON group, the subjects lifted the load from a starting angle of 160° to an end position of approximately 70° in the elbow joint. Similarly, the ECC group lowered the load from approximately 70 to 160°. An assistant returned the load to the starting position, thus letting the subjects train concentrically or eccentrically only. The subjects in ECC were instructed to use 3-4 s for each lowering of the load, and the subjects in CON were instructed to use maximum effort in every lift. To ensure progression in training, the relative training load was gradually increased every fourth week as the target number of repetitions in each set was decreased from eight to six repetitions in week 4 and then to four repetitions after week 8. In addition, the individual absolute training load was progressively increased as the subjects became stronger. If all repetitions and sets were completed during the maximum training bout, the load was increased (0.25-1 kg) at the next maximum session. If the subject failed to complete the session as described, the load was unaltered at the next maximum sessions until it was successfully completed. The number of sets in each maximum session rose from three during the first week to five during the last weeks. There was a 3- to 6-min pause between each set. Three subjects in the CON and ECC groups did not complete the final training session. Apart from the training in the experiment, the subjects were not allowed to perform exercises in which the elbow flexors were the prime movers (e.g., arm-curl exercises). However, the subjects were allowed to continue their regular training with exercise where the elbow flexors were only synergists to the movement (e.g., dorsal back-muscle exercises). These exercises had been a regular part of the subjects' training for months before this study. Because the elbow flexors are not the prime movers during such exercises, we assumed that their regular training would have minimal influence on the outcome of this study. Also, because the subjects were randomized, a possible effect should be similar in both groups.

The training program for the concentric exercise and eccentric exercise groups.

Performance Tests

The subjects were tested twice on each performance test on four separate days, both before and after the training period. The coefficient of variation of the strength tests were approximately 2%. The highest score from each test was used for later analysis.

Maximum concentric strength

Maximum concentric strength was measured as the concentric 1RM in the elbow flexion. The range of motion of the elbow joint started from a slightly flexed (160°) position to an end position of 75° in the elbow joint. The subjects completed three to four sets of three to five repetitions as a warm-up. Thereafter, single lifts were performed, with the load increased gradually until the maximum was reached. Each lift was separated by a pause of 3-5 min to ensure full recovery between the attempts. This same rest protocol was used for all performance tests in the study. The specific concentric strength was defined as the maximum concentric strength divided by the mean anatomical cross-sectional area of the elbow flexors.

Maximum eccentric strength

The eccentric 1RM was measured as the maximum load the subject could lower from 75 to 160° in the elbow joint with a constant velocity and with the movement lasting a minimum of 3.5 s. The load sledge was locked at the start position of 75° and released at a given signal. The subjects then resisted the downward movement of the load sledge through the entire range of motion. The warm-up consisted of two to three sets of three to five repetitions of traditional lifts with gradually increasing load. Specific eccentric warm-up started at loads corresponding to approximately 90% of the concentric 1RM; thereafter, each subject used three to five maximum lifts to establish the eccentric 1RM. The specific eccentric strength was defined as the maximum eccentric strength divided by the mean anatomical cross-sectional area of the elbow flexors.

Maximum angular velocity

The maximum angular velocity in the elbow joint was measured at loads of 30, 50, 70, and 90% (V30, V50, V70, and V90) of the pretraining concentric 1RM and at a common absolute load corresponding to the weight of the handgrip and lever arm (~2 kg, termed V2). The velocity was taken midway during the lift, at 115° of elbow joint flexion. The subjects completed a standard warm-up regime of two to three sets of three to five repetitions at 30-50% of the preconcentric 1RM and one set at the specific test load with maximum mobilization. The tests were taken in random order between the subjects; for any given subject, the same order was used after the training period.

Anatomical Muscle Cross-Sectional Area

The anatomical cross-sectional area of the elbow flexors (brachialis, brachioradialis, and biceps brachii) was measured via CT (Somatom Emotion, Siemens, Munich, Germany). Four serial cross-sectional images were taken of the upper arm at one-, two-, three-, and four-eighths the length of the humerus (Lh) starting from the distal end at the base of capitulum humerii and moving proximally towards the base of caput humerii. Each CT slice was 10 mm thick. The muscle area was measured planimetrically and was conducted offline using ImageJ software (National Institute of Health, Bethesda, MD). It was not possible to determine the individual borders of the different muscles in the flexor group at all axial scans. Therefore, the flexor group as a whole was carefully followed using the trace function in the software. Each scan was measured twice by an investigator blinded to the identity of the subject and the time point of the scans, and the mean of both measurements was used in the analysis. The mean of the four individual sites was taken as the mean flexor area. One picture was randomly chosen among subjects and used for reliability testing of the procedures. Twenty measurements of the area were completed, and the coefficient of variation was 0.3%. Because of a procedural error, the post scans of one subject in the ECC group had to be excluded.

Biopsy Sampling

The subject lay in the supine position and was first given local anesthesia (Xylocaine 10 g·L−1) in the skin and thereafter in the fascia and muscle (Xylocaine with adrenaline 10 g·L−1) of the biceps brachii of the nondominant arm. Two 1-cm cuts were made in the skin and fascia midway in the long head of the biceps, and several muscle samples were taken using the percutaneous needle biopsy method (5). For one subject in the ECC group, only one sample was taken because the subject experienced the procedures as being unpleasant. The muscle biopsies were frozen in Isopentane cooled to its freezing point of −160°C by liquid nitrogen and stored in a cryofreezer at −80°C until later analyses. Posttraining biopsies were taken randomly, either 1 cm distally or proximally to the first biopsy site. Posttraining muscle samples were not obtained from two subjects in the ECC group for reasons not related to the study.


Before sectioning, the muscle samples were warmed to −20°C in a cryostat (HM 560 M, Microm International, Walldorf, Germany), then oriented and mounted on metal discs using an embedding medium (Tissue-Tek O.C.T. Compund, Sakura, Tokyo, Japan). Serial cross-sections of 10 μm were cut from two muscle samples from each subject and put on glass microscope slides. Myofibrillar mATPase histochemistry was performed (7) as modified by Andersen and Aagaard (4) to distinguish between five different fiber types. Muscle fibers were classified as type I, IIC, IIA, IIA/X, and IIX according to their stability to the acidic and alkaline preincubations. Glass slides of both pre- and posttraining muscle samples of each subject were stained simultaneously to minimize possible risks of different staining intensities between assays.

Image Analysis

The slides were placed on a microscope (BX50WI, Olympus, Tokyo, Japan) and magnified using a 10× water immersion objective and then photographed using a digital camera (Coolpix 995, Nikon) connected to the microscope. A composite photo montage of the images of each preparation (preincubation pH 4.37, 4.6, and 10.3) was then assembled in Photoshop 7.0 (Fig. 2). A mean ± SD of 742 ± 286 fibers before and 664 ± 254 fibers after the training period from two to three muscle samples per subject were classified by an investigator blinded to the identity of the subjects. Because there were relatively few fibers for one subject in ECC (N < 200), this subject's muscle samples were excluded from further analysis. Therefore, fiber analysis was completed in six subjects in ECC. However, when comparing the groups, there were no differences at any dependent variable after omitting these subjects, with the exception of the proportion of type I fibers. The fiber area was measured planimetrically using the image software ImageJ (National Institute of Health, Bethesda, MD). The images from preparation for pH 4.6 were used, and only fibers with clear cell boundaries and that appeared cross-sectioned were measured. The quality of the muscle cells was poorly preserved in one subject in CON, and his samples were therefore not included. Because of the small numbers of type IIC, IIA/X, and IIX fibers in most of the subjects, only the areas of type I and IIA were measured. A mean of 81 ± 18 type I and 88 ± 17 IIA fibers from two to three samples were measured before and after the training period from each subject. To determine the weighted mean fiber area and the relative cross-sectional area occupied by the type I and II fiber types, the IIC, IIA/X, and IIX fibers were first added to the type I or the type IIA fiber proportions by the following formulae:

Serial cross-sections of a muscle biopsy of the biceps brachii muscle taken from one subject in the eccentric exercise group before the study (A-C). Cross-sections were stained for myofibrillar ATPase activity after preincubation at pH 4.37 (A), 4.6 (B), and 10.3 (C). I, type I fibers; IIA, type IIA fibers; IIX, type IIX fibers. Bar, 200 μm.

True type refers to the fiber type assessed by the myofibrillar ATPase protocol as described above.

The weighted mean fiber area was then calculated on basis of the type I and II fiber proportions and the mean cross-sectional area of type I and type IIA as follows:

The relative cross-sectional area occupied by the type I and type II fibers was calculated using the formula of Castro et al. (10).

Statistical Analysis

We used eight CON and nine ECC subjects. If the true underlying effect of training differed by 1 SD between the groups in a parameter, the statistical tests used had a power of 0.62. If the effect differed by 1.5 SD, the power was 0.90, and if the effect differed by at least 2 SD, the power was 0.984 or more, using a level of 0.05 for statistical significance. The effect of slightly smaller sample sizes for analyses of the muscle biopsies had no major effects on the power (not shown). Thus, true differences in the effect of training less than 1 SD may have been masked by random variations in this study, whereas true differences of 1.5 SD or more were most likely to result in statistical significance. All results are given as mean ± standard deviation. Paired Student's t-test was used to test for changes in the dependent variables within groups. Two-sample t-tests were used to test for possible differences in the changes between the groups. We have also analyzed the data using ANCOVA. Because the outcomes of these latter analyses were almost identical to those of the t-test, we do not report the results of the ANCOVA. Pearson correlation coefficient was used to evaluate possible relationships between selected variables. P values less than 0.05 were considered statistically significant, and P values between 0.10 and 0.05 were given as trends.


Training progression

All participants completed 29 training sessions during the 12-wk experimental period, and all but three participants in each of the CON and ECC groups completed the final session. The loads increased steadily in both groups as a function of increased strength as well as decreased number of repetitions in each set (Fig. 3A), but the increase in training load was larger for ECC than for CON (59 vs 41%, P < 0.001 for ECC vs CON). The mean training load in ECC was 21% larger than that used by CON at the first training session, and at the end of the training period this difference had increased to 37%.

A. Progression of the training load during the training period from sessions 1 to 29 in the concentric exercise group (CON; filled circles, solid line) and the eccentric exercise group (ECC; open circles, dashed line). The horizontal arrows show the training sessions that employed the described number of repetitions (rep). B. The maximum concentric strength and eccentric strength before and after the training in CON (filled bars) and ECC (open bars). The data are given as mean ± SD. *Significantly different from pre values, P < 0.001; †significantly greater increase than for CON, P < 0.001.

Maximum strength

The concentric 1RM increased to a similar extent for the two groups during the training period (P < 0.001 for post- vs pretraining tests; Fig. 3B). The improvements were 4.7 ± 2.2 kg (18%) and 3.9 ± 1.3 kg (14%) after training for CON and ECC, respectively. The eccentric 1RM improved more for ECC (8.6 ± 3.3 kg, 26%) than for CON (3.1 ± 1.3 kg, 9%; P < 0.001 for ECC vs CON). Consequently, the ratio between the eccentric 1RM and the concentric 1RM developed differently (P < 0.001). For CON, the ratio decreased from 1.30 ± 0.10 to 1.20 ± 0.12 during the training period (P = 0.007), whereas for ECC, the ratio increased from 1.21 ± 0.08 to 1.33 ± 0.12 (P = 0.008).

Maximum angular velocity

Both groups increased the maximum angular velocities at all loads during the training period (P < 0.05; Fig. 4), and there was no difference between the two groups. The increase in the absolute angular velocity was similar in all tests (V2-V90) and ranged from 18 (V2) to 53°·s−1 (V90). Because the absolute angular velocity decreased as the load increased, the relative effect of training was greatest at the highest loads. When the test loads of 90, 70, 50, and 30% of the preconcentric 1RM were normalized to the postconcentric 1RM, the loads were reduced to mean values of 78 ± 4, 60 ± 3, 43 ± 3, and 26 ± 2%, respectively. After the normalization of the test loads, there was little difference in the velocity of shortening from that before the training period (Fig. 5).

Maximum angular velocity at loads 90% (V90), 70% (V70), 50% (V50), and 30% (V30) of the maximum concentric strength at pretest and the common 2-kg load (V2). All velocities are taken at 115° in the elbow joint before (filled circles, solid lines) and after (open circles, dashed lines) the study in the concentric exercise group (CON; panel A) and the eccentric exercise group (ECC; panel B). * Significantly different from pre values, P < 0.05. The data are mean ± SD.
Relative load-velocity relationship in the concentric exercise (CON; panel A) and eccentric exercise (ECC; panel B) groups before (filled circles, solid lines) and after (open circles, dashed lines) the study. The loads V90, V70, V50, and V30 are normalized to both pre- and postconcentric strength. The test loads were reduced at posttesting to 78 ± 4, 60 ± 3, 43 ± 3, and 26 ± 2% the post 1RM, respectively. The data are mean ± SD.

Fiber-type proportions

The proportion of the type I fibers differed between the two groups before the training period (P = 0.005; Table 3). Before the study, the subjects had relatively few IIX (2%) and IIA/X fibers (8%); five subjects had no IIX fibers, and six subjects had less than 2% type IIA/X fibers. The proportions of type I fibers did not change in either group during the training period, although there was a trend toward a reduced proportion in the ECC group (P = 0.08). Also, there was no difference between the two groups in the proportions of the type I fibers after the training period. There were likewise few alterations in the subgroups of the type II fibers. Only the 2.8% reduction of IIX fibers in CON was statistically significant, whereas there was a trend for reduction in the proportion of the IIAX fibers. There were no other significant differences or changes neither within nor between the two groups. In addition, when the two groups were pooled together, we found no statistically significant changes in the fiber-type proportions.

Fiber-type proportions (%) in the CON and ECC groups before and after the training period.

Muscle cross-sectional area

As seen in Figure 6, the anatomical cross-sectional areas were greater in the distal regions (one- and two-eighths Lh) than in the proximal regions of the elbow-flexor group (three- and four-eighths Lh). The mean anatomical elbow-flexor cross-sectional area (mean of one- to four-eighths Lh) of 26.8 ± 4.9 cm2 of the CON group did not change during the training period (+0.7 ± 1.1 cm2, +3%; P = 0.1). For the ECC group, the mean area of 25.4 ± 3.4 cm2 rose by 2.8 ± 1.4 cm2 (11%; P < 0.001) during the training period. The effect thus differed between the two groups (P = 0.004; Fig. 6). Increases in absolute area were greater where the flexor group had its largest portions (one- and two-eighths Lh), but the relative hypertrophy was comparable between the four individual regions (one- to four-eighths Lh) in ECC (10-12%). The corresponding nonsignificant increases of 2-4% in CON were likewise similar between the different regions. The specific concentric strength of the elbow-flexor group (concentric strength/mean cross-sectional area) increased in CON from 0.96 ± 0.06 to 1.10 ± 0.10 kg·cm−2 (P = 0.006). There was no change in specific concentric strength in ECC (from 1.05 ± 0.12 to 1.08 ± 0.09 kg·cm−2). The specific eccentric strength (eccentric strength/mean cross-sectional area) increased both in CON (from 1.24 ± 0.12 to 1.32 ± 0.09 kg·cm−2, P = 0.02) and in ECC (from 1.27 ± 0.12 to 1.46 ± 0.18 kg·cm−2; P = 0.02).

Serial anatomical cross-sectional areas (ACSA) of the elbow flexors in the concentric exercise (CON; panel A) and eccentric exercise (ECC; panel B) groups before (filled circles, solid lines) and after (open circles, dashed lines) the training period. Serial ACSA are given at intervals of one-eighth the length of humerus from the distal (left) to the proximal (right) end. * Significantly different from pre values, P < 0.005. The data are mean ± SD.

Fiber cross-sectional areas

Before the study, the cross-sectional area of the IIA fibers was 70% larger than that of the type I fibers (P < 0.001; Table 4). The fiber area of type I and IIA fibers and the weighted mean area increased only in ECC during the training period (P < 0.01; Table 4; Figs. 7 and 8). For this group, the area of type IIA fibers increased 2.8 times more than the type I area did (P = 0.002), and consequently the type IIA/I area ratio rose from 1.78 ± 0.25 to 1.99 ± 0.29 during the training period (P = 0.007). The relative cross-sectional area occupied by type II fibers before the training period was less for ECC than for CON (P = 0.01). As a result of training, the relative area occupied by the type II fibers increased and the area of the type I fibers decreased by 9% for ECC only (P = 0.05; Table 4). When the data of the two groups were pooled (N = 13), there was a positive corelation between the mean flexor area and the weighted mean fiber area (r = 0.87; P < 0.001). Thus, subjects with the largest flexor muscles also had the largest muscle fibers. In addition, the individual changes in muscle area correlated with the changes in the fiber area (r = 0.59; P = 0.03).

Single-fiber cross-sectional area, weighted mean cross-sectional area, and relative cross-sectional fiber area in CON and ECC before and after the training period.
Frequency distribution of type I fiber area in the concentric exercise (CON; panel A) and eccentric exercise (ECC; panel B) groups before (solid lines) and after (dashed lines) training. The areas have been grouped in bins of 500 μm2.
Frequency distribution of type IIA fiber area in the concentric exercise (CON; panel A) and eccentric exercise (ECC; panel B) groups before (solid lines) and after (dashed lines) training. The areas have been grouped in bins of 500 μm2.

Relationship between contractile area and muscular performance

Before the study, the mean flexor area was correlated with both concentric and eccentric strength (r = 0.81 and 0.79; P < 0.001) and with the maximum angular velocity at both the 2-kg and the 30% load (r = 0.83, P < 0.001; r = 0.71, P = 0.002). Similar correlations were also found between the weighted mean fiber area and concentric strength (r = 0.77; P = 0.002), eccentric strength (r = 0.74; P = 0.004), and angular velocity at the 2-kg load (r = 0.78; P = 0.002). Thus, the strongest subjects had the largest muscle fibers and flexor muscles. Also, the individual changes in weighted mean fiber area in ECC (N = 6) correlated with changes in the maximum angular velocity at the V70 load (r = 0.93; P = 0.007), whereas at loads V30, V50, and V90, the correlations were only trends (r = 0.78, 0.75, and 0.79; P < 0.1).


One of the major findings in the present study was that eccentric training in contrast to concentric training induced hypertrophy as measured by both the mean anatomical flexor cross-sectional area and the cross-sectional area of single fibers in our well-trained subjects. Eccentric training led also to a larger gain in eccentric strength than concentric training did. However, both the concentric 1RM and the angular velocity at different loads were as much enhanced after the concentric as after the eccentric training. Finally, the training regimes did not lead to any major changes in fiber-type proportions.

Strength and hypertrophy.

After the 12-wk training period, the anatomical cross-sectional area of the elbow-flexor group was only enhanced in the eccentric training group. This finding differs from previous experiments on untrained subjects that have reported increased muscle area after both concentric and eccentric training (8,12,16,19,26). However, several of these studies have shown larger hypertrophy after eccentric than after concentric training (12,16,26). For the subjects in ECC, the gains in the absolute cross-sectional area were greater in the largest distal portion of the elbow-flexor group, although in relative terms, the cross-sectional area increased uniformly (10-12%) in all regions of the muscle group. This is consistent with the results from the study of Farthing and Chilibeck (12) that also used elbow flexion as a model. However, there have been reports of nonuniform hypertrophy after eccentric training. For example, Seger et al. (26) noted that the greatest relative increase in muscle area was located in the distal part of the trained quadriceps muscle. These different findings might be related to the different muscles studied and also to the methods used for calculation of muscle area.

The gain in relative muscle cross-sectional area was about two to three times larger than that reported by Higbie et al. (16), Jones and Rutherford (19), and Seger et al. (26) in studies of comparable training durations. However, a direct comparison is difficult because those studies were based on the quadriceps muscle.

In agreement with the observations on whole muscles, only the ECC group showed hypertrophy of single fibers. In untrained subjects, increased fiber area has been reported after both concentric and eccentric training, although larger increases were seen after eccentric training (17,18). In the present study, the hypertrophy was most evident in the type IIA fibers, which, in absolute terms, increased three times as much as the type I fibers. This finding seems consistent with the observation that even before the training period, the type IIA fibers were 70% larger than the type I fibers. Greater hypertrophy in type II fibers is a typical finding after traditional resistance training (15,22), and it does not seem to be an effect of eccentric training in particular. Taken together, the results from the muscle and single-cell cross-sectional areas suggest that development of high force is an important factor for gains in muscular hypertrophy in trained subjects. The present results can be seen as an adaptation continuum from that of untrained subjects because most studies on these subjects also find greater hypertrophy after eccentric than concentric training. Results from previous studies indicate that for resistance-trained subjects, it becomes gradually more difficult to enhance muscle area with prolonged training (2,13). It is therefore possible that the load used in the training becomes increasingly more important to promote further muscular hypertrophy in resistance-trained subjects. In the present study, the training load for the ECC group was about 20% higher than that for the CON group during the first week of training, and this difference increased to around 35% during the final week of training. Our findings thus support the idea that development of high muscular force during training is essential to promote muscle hypertrophy, at least in well-trained subjects.

Similar to some (8,17,20) but not all (12,16,18) earlier studies on untrained subjects, both the CON and ECC groups displayed a similar increase in concentric strength. This finding was unexpected because we hypothesized that a larger hypertrophy in ECC would consequently lead to a greater increase in concentric strength in the resistance-trained subjects. The comparable gains in concentric 1RM thus seem to be a result of a different adaptive mechanism because hypertrophy was detected in ECC only. Both concentric strength and hypertrophy increased uniformly in ECC, and there was consequently no change in the specific concentric strength of the elbow flexors. Thus, it seems that the gain in muscle area could account for almost all of the increase in the concentric 1RM. In contrast, there was a large increase in the specific concentric strength in CON, which indicates either a change in muscle quality per se or a change in muscle activation. One possible explanation for the increased force per cross-sectional area in CON is that the intrinsic strength of the single cells increased as a function of the concentric training. However, changes in the specific force of single cells has not been reported after resistance training in adult humans (27,30). It is therefore more likely that the increases in strength in CON were a result of increased neural activation as reported previously in untrained subjects (16,17).

Because we hypothesized that the resistance-trained subjects would already have optimized their neural drive in previous training, the present finding in CON was unexpected. It has been suggested that decreased activation of antagonistic muscles could contribute to the enhancement of strength in previously untrained subjects (9), but findings are divergent between studies (14,17). In a study by Amiridis et al. (3), coactivation by antagonistic muscles was shown to be significantly lower in highly skilled athletes than for sedentary subjects, which suggests that reduced coactivation is a less probable cause for the gain in strength in the present study. Even though the elbow flexors of the subjects were resistance trained, they were unaccustomed to the specific exercise, and it could be argued that they were not well trained for this particular exercise. It is known that increases in strength during a training period in formerly untrained subjects are specifically connected to the exercise carried out, with less transfer effect to other exercises employing the same muscles (28). Thus, it appears that resistance training does not optimize neural drive in a generalized manner. Moreover, different exercises may require some specific motor-unit activation even in simple tasks such as one-joint movements. If so, starting training with an unfamiliar exercise might allow for improved neural adaptations, even for resistance-trained individuals.

The maximum eccentric strength increased significantly more in ECC than in CON, a finding that seems to be very consistent throughout previous studies on untrained subjects (12,16-18,20). The relative gain in eccentric 1RM in ECC was far greater than the gain in anatomical cross-sectional area (26 vs 11%); as a result, the specific eccentric strength rose after training. Thus, it appears that the hypertrophy can explain only part of the increase in eccentric strength. Some studies have reported that the intrinsic strength of type II fibers is greater than that of type I fibers (6,30). It is therefore possible that the specific eccentric strength of the whole muscles increased as a function of the preferential hypertrophy of the type II fibers; this could explain part of the discrepancy between gains in eccentric strength and muscle cross-sectional area. However, if there was an increase in a specific strength as a result of preferential hypertrophy of the type II fibers, we would expect to find a similar increase in specific concentric strength of the elbow flexors, which was not the case. It is thus more likely that a large part of the gain in eccentric strength in ECC was caused by improved neural activation, as has been reported in untrained subjects (16,17). This finding indicates that maximum eccentric strength is not optimized during traditional resistance training and that maximum eccentric strength might require a different activation than during traditional concentric-eccentric lifts, even for resistance-trained subjects. Hence, because the subjects were unaccustomed to maximum eccentric lifting before the study, it is possible that they could be termed less trained for maximum eccentric contractions. The lack of both cell growth and muscle hypertrophy in the CON group indicates that the gain in eccentric 1RM is attributable to neural adaptations, as we have suggested previously for the improved concentric 1RM. Hence, unfamiliar resistance training seems to allow for improvements in neural activation to increase the force output, also for trained individuals.

Fiber-type proportions and angular velocity.

The two groups differed in their proportion of type I fibers before the study started. We assume that this difference is a random effect that may occur by chance because the use of the needle biopsy method for determining fiber-type proportion for a whole muscle has a relatively large inherent methodological uncertainty. Moreover, an apparent statistically significant difference may occur by chance when many independent comparisons are carried out, as was the case in this study. After training, there were no significant alterations in the proportions of type I fibers in either group, which seems to be a consistent finding after resistance training in humans for both previously untrained and well-trained subjects (2,15,22,28). However, we observed a trend toward reduction in the proportions of the type I fibers in ECC in that a decrease was seen in five out of the six subjects analyzed. We saw only small changes in the proportions of the subgroups of type II fibers. It is a common finding that regular resistance training in previously untrained subjects results in transformation of IIX(B) fibers to type IIA fibers (4,15,17,30). Our subjects had only few IIX (2%) and IIA/X fibers (8%) before the study started, which was probably an effect of their high training state before the study. Consequently, there were few fibers that could be transformed to type IIA. However, as a function of the preferential hypertrophy of the type IIA fibers as well the nonsignificant reduction in the proportion of type I fibers, the relative cross-sectional area occupied by the type II fibers increased from 64 to 73% in the ECC group.

ECC and CON increased the maximum angular velocity equally at each of the loads relative to the pretest concentric 1RM. These findings are in agreement with previous studies that have reported similar gains in concentric torques at low to medium velocities after eccentric and concentric training on untrained subjects (8,26). The relative gain in maximum angular velocity was greater at high loads than at low loads for both groups, a finding that is supported by previous studies after traditional and isometric resistance training (24,29). However, when the standard test loads of 90, 70, 50, and 30% of the preconcentric strength were normalized to the postconcentric strength, there were no changes in velocities. The gain in velocity at the standard loads seems, therefore, to be an effect of increased concentric strength and not increased maximum speed of shortening of the muscles, which is in accordance with the force-velocity relationship. Even though we found an increase in the relative area occupied by the type II fibers in ECC, it appears that this increase did not contribute significantly to the gains in the maximum angular velocity in this study. Because improvements in concentric strength in ECC are probably explained mainly by hypertrophy, the gains in angular velocity in ECC are likely a result of overall muscle cross-sectional area. Because we did not observe hypertrophy in the CON group, we suggest that the gain in maximum angular velocity was largely mediated by neural adaptations in this group.

Implications for functional training.

These results indicate that eccentric training is important to promote hypertrophy in already resistance-trained men, at least in the elbow flexors. Because eccentric strength increased significantly more after eccentric than after concentric training, eccentric training could also be important in sports subjected to large eccentric forces. However, because there was no difference between groups in the gain in performance in concentric strength and maximum angular velocity, it is possible that both maximum eccentric and concentric actions should be part of the training. Hence, further studies should investigate the effects of combined maximum eccentric and concentric training in trained athletes. Some caution should be made when interpreting the results of this study to practical training in athletes. First, we have compared the effects of eccentric versus concentric training, two training models that are seldom used in practical resistance training. It may therefore be difficult to compare these results directly with traditional resistance training. Second, even though only one subject dropped out of the study because of reoccurring pain in his arm, most subjects in ECC perceived the training as hard to complete. Before planning long-term training using eccentric models, caution is advised because the risk for injury may be higher than that of traditional resistance training. Third, we used the elbow flexors as a model in this study, and it is possible that the results are more relevant for non-weight-bearing muscles than for weight-bearing muscles such as the knee extensors.

In conclusion, only eccentric training increased anatomical muscle and fiber cross-sectional areas in previously resistance-trained men. The results of this study suggest that it is important to create high muscular force in training to promote further hypertrophy in already well-trained human muscles. Maximum eccentric and concentric training increased both concentric strength and maximum angular velocity to a similar degree despite hypertrophy in ECC only. On the other hand, eccentric strength increased significantly more after the eccentric training. These results suggest that in the ECC group, gains in concentric 1RM and maximum angular velocity were largely mediated by hypertrophy, whereas eccentric strength increased as a function of both hypertrophy and neural adaptations. In the CON group, the gains in strength and velocity performance were probably mainly a result of neural adaptations.

This study was partially funded by grants from The Norwegian Olympic Sports Centre. The authors would like to thank Vidar Jakobsen, Michael Guttormsen, Jo C. Bruusgaard, and Steinar Messel for technical support, and Arne Høiseth at Sentrum Røntgeninstitutt, Oslo for his excellent assistance in taking the computer tomography images. We further thank Professor Johan B. Steen and Robert C. Reid for giving valuable comments to the manuscript and suggestions for improvement, and Professor Knut Liestøl for his advice on the statistics.


1. Aagaard, P. Training-induced changes in neural function. Exerc. Sport Sci. Rev. 31:61-67, 2003.
2. Alway, S. E., W. H. Grumbt, J. Stray-Gundersen, and W. J. Gonyea. Effects of resistance training on elbow flexors of highly competitive bodybuilders. J. Appl. Physiol. 72:1512-1521, 1992.
3. Amiridis, I. G., A. Martin, B. Morlon, et al. Co-activation and tension-regulating phenomena during isokinetic knee extension in sedentary and highly skilled humans. Eur. J. Appl. Physiol. 73:149-156, 1996.
4. Andersen, J. L., and P. Aagaard. Myosin heavy chain IIX overshoot in human skeletal muscle. Muscle Nerve 23:1095-1104, 2000.
5. Bergstrøm, J. Muscle electrolytes in man. Scand. J. Clin. Lab. Invest. 14:1-110, 1962.
6. Bottinelli, R., and C. Reggiani. Human skeletal muscle fibres: molecular and functional diversity. Prog. Biophys. Mol. Biol. 73:195-262, 2000.
7. Brooke, M. H., and K. K. Kaiser. Muscle fiber types: how many and what kind? Arch. Neurol. 23:369-379, 1970.
8. Carey Smith, R., and O. M. Rutherford. The role of metabolites in strength training. I. A comparison of eccentric and concentric contractions. Eur. J. Appl. Physiol. 71:332-336, 1995.
9. Carolan, B., and E. Cafarelli. Adaptations in coactivation after isometric resistance training. J. Appl. Physiol. 73:911-917, 1992.
10. Castro, M. J., D. F. Apple Jr., R. S. Staron, G. E. R. Campos, and G. A. Dudley. Influence of complete spinal cord injury on skeletal muscle within 6 mo of injury. J. Appl. Physiol. 86:350-358, 1999.
11. Enoka, R. M. Neural adaptations with chronic physical activity. J. Biomechanics 30:447-455, 1997.
12. Farthing, J. P., and P. D. Chilibeck. The effects of eccentric and concentric training at different velocities on muscle hypertrophy. Eur. J. Appl. Physiol. 89:578-586, 2003.
13. Häkkinen, K., P. V. Komi, M. Alén, and H. Kauhanen. EMG, muscle fibre and force production characteristics during a 1 year training period in elite weight-lifters. Eur. J. Appl. Physiol. 56:419-427, 1987.
14. Häkkinen, K., M. Alen, W. J. Kraemer, et al. Neuromuscular adaptations during concurrent strength and endurance training versus strength training. Eur. J. Appl. Physiol. 89:42-52, 2003.
15. Hather, B. M., P. A. Tesch, P. Buchanan, and G. A. Dudley. Influence of eccentric actions on skeletal muscle adaptations to resistance training. Acta Physiol. Scand. 143:177-185, 1991.
16. Higbie, E. J., K. J Cureton, G. L. Warren III, and B. M Prior. Effects of concentric and eccentric training on muscle strength, cross-sectional area, and neural activation. J. Appl. Physiol. 81:2173-2181, 1996.
17. Hortobágyi, T., J. P. Hill, J. A. Houmard, D. D. Fraser, N. J. Lambert, and R. G. Israel. Adaptive responses to muscle lengthening and shortening in humans. J. Appl. Physiol. 80:765-772, 1996.
18. Hortobágyi, T., L. Dempsey, L. Fraser, et al. Changes in muscle strength, muscle fibre size and myofibrillar gene expression after immobilization and retraining in humans. J. Physiol. 524:293-304, 2000.
19. Jones, D. A., and O. M. Rutherford. Human muscle strength training: the effects of three different regimens and the nature of the resultant changes. J. Physiol. 391:1-11, 1987.
20. Komi, P. V., and E. R. Buskirk. Effect of eccentric and concentric muscle conditioning on tension and electrical activity of human muscle. Ergonomics 15:417-434, 1972.
21. Levin, A., and J. Wyman. The viscous elastic properties of muscle. Proc. Roy. Soc. B. 101:218-243, 1927.
22. McCall, G. E., W. C. Byrnes, A. Dickinson, P. M. Pattany, and S. J. Fleck. Muscle fiber hypertrophy, hyperplasia, and capillary density in college men after resistance training. J. Appl. Physiol. 81:2004-2012, 1996.
23. Moritani, T., and H. A. DeVries. Neural factors versus hypertrophy in the time course of muscle strength gain. Am. J. Phys. Med. Rehabil. 58:115-130, 1979.
24. Moss, B. M., P. E. Refsnes, A. Abildgaard, K. Nicolaysen, and J. Jensen. Effects of maximal effort strength training with different loads on dynamic strength, cross-sectional area, load-power and load-velocity relationships. Eur. J. Appl. Physiol. 75:193-199, 1997.
25. Sale, D. G. Neural adaptations to resistance training. Med. Sci. Sports Exerc. 20:135-145, 1988.
26. Seger, J. Y., B. Arvidsson, and A. Thorstensson. Specific effects of eccentric and concentric training on muscle strength and morphology in humans. Eur. J. Appl. Physiol. 79:49-57, 1998.
27. Shoepe, T. D., J. E. Stelzer, D. P. Garner, and J. J. Widrick. Functional adaptability of muscle fibers to long-term resistance exercise. Med. Sci. Sports Exerc. 35:944-951, 2003.
28. Thorstensson, A., B. Hultén, W. von Döbeln, and J. Karlson. Effects of strength training on enzyme activities and fibre characteristics in human skeletal muscle. Acta Physiol. Scand. 96:392-398, 1976.
29. Toji, H., and M. Kaneko. Effect of multiple-load training on the force-velocity relationship. J. Strength Cond. Res. 18:792-795, 2004.
30. Widrick, J. J., J. E. Stelzer, T. C. Shoepe, and D. P. Garner. Functional properties of human muscle fibers after short-term resistance exercise training. Am. J. Physiol. 283:R408-R416, 2002.


©2006The American College of Sports Medicine