Journal Logo

Applied Sciences: Biodynamics

Greater cross education following training with muscle lengthening than shortening


Author Information
Medicine & Science in Sports & Exercise: January 1997 - Volume 29 - Issue 1 - p 107-112
  • Free


Cross education and bilateral deficit are manifestations of neural integration of interlimb coordination (14). Cross education, as first described by Scripture et al. (25), is a chronic effect that takes the form of performance enhancement in an inactive limb musculature following exercise training of a remote limb. Several prior reports found evidence for cross education of muscle strength, ranging from 0% to 56% (29) using isometric(26), isotonic shortening (32), isotonic (2) contractions of small(4) or larger (11) muscle groups. Using electromyostimulation, some researches succeeded(3) while others failed (4,7) in an attempt to separate central from peripheral factors mediating cross education. The significance of cutaneous effects was demonstrated by Matyas et al. (17): rapid brushing increased EMG activity by 25% in the contralateral quadriceps during isometric contractions in hemiplegics.

Despite being an intrinsic component of human motion, muscle lengthening has rarely been used to investigate the cross education phenomenon. Lagassé (16) and Smith (28) attempted to take advantage of the crossed extension-reflex(27) using a superimposed stretch of the quadriceps. The crossed extension-reflex inhibits the contralateral homologous muscles and simultaneously excites the contralateral antagonist muscles in the spinal cat. While Smith (28) reported a superiority of cross education with stretch versus isometric contraction, to date, Lagassé's(16) is the only study that reports a decrease in contralateral homologous muscle strength following ipsilateral training. Most recently muscle stretch in the form of pulsed eccentric loading was found to cause up to 135% cross education in the quadriceps muscle(1).

Our goal was to directly compare the magnitude of cross education following training with muscle lengthening or shortening. Specifically, we tested the hypothesis that the magnitude of cross education effect is greater with muscle lengthening than shortening. This hypothesis emanated from prior observations that in sedentary subjects there was a significantly greater difference between measured and expected forces during maximal effort, voluntary muscle lengthening compared with shortening based on forces evoked by a combination of voluntary effort, and electromyostimulation (31). A similar pattern of force deficit was apparent when sedentary and strength-trained subjects were compared (10). Thus, the purpose of the study was to compare the magnitude of cross education following training with muscle lengthening and shortening in the quadriceps muscle.


Subjects and design. Twenty-one sedentary male volunteers (21.3(±SD) 1.9 yr, 175.2 ± 2.6 cm, 75.4 ± 6.8 kg) who were randomly assigned to one of three groups: eccentric training (ECC, N= 7), concentric training (CON, N = 8), and control (KON,N = 6). A subject was included in the study if he did not participate in resistive training at least 1 yr prior to the study; was free of lower extremity orthopedic problems, and agreed to sign an informed consent form.

All subjects were tested twice initially (week 0), once at week 6 during training, and once after 12 wk of training (week 12). Two initial testing sessions, administered 3 d apart, were used to assess the reliability of the strength and EMG measures. The subjects in the control group performed all testing but did not train.

Strength testing and EMG. Unilateral maximal voluntary isometric and isokinetic eccentric and concentric strength of the left knee extensors and flexors were measured with a dynamometer (Kin-Com, 500H, Chattecx, Inc., Chattanooga, TN). Subjects sat with knees and the hip at a joint angle of 1.57 rads (90°) and arms folded in front of the chest. The anatomical zero was set at a knee angle of 3.14 rads. Extraneous movement of the upper body and the involved leg was limited by two crossover shoulder harnesses, a lap belt, a thigh strap, and an ankle cuff. The transverse axis of the knee joint was aligned with the transverse axis of the dynamometer's power shaft. The length of the lever arm was individually determined. Torque was measured by a strain gauge embedded in the ankle cuff. Gravity-corrected force was computed by the software based on lever arm length and torque data. Familiarization with the dynamometer included two trials of 50%, 75%, and 90% of perceived maximal isometric and dynamic contractions at each speed, separated by 1 min of rest.

Maximal isometric force was measured at a knee angle of 2.36 rads. Two, maximal effort, 5-s trials were performed with 1 min of rest between trials. Maximal concentric and eccentric strength of the knee extensors and flexors was measured at 1.05, 2.09, and 3.14 rad·s-1, but only the data at 1.05 rad·s-1 are reported. Subjects performed two trials with a 1-s pause at either end of the range of motion to avoid the facilitating effects of the prior action. The order of isometric versus dynamic actions and eccentric versus concentric actions was counterbalanced across subjects and the order of speeds randomized. The concentric and eccentric force-angle curves were digitized at 2.36 rads. Because variability from trial-to-trial was not significant, the average of two trials was used as the criterion measure. Reliability, as estimated with an intraclass correlation coefficient(R) over 2 d and two trials, was R = 0.89 for concentric, R = 0.94 for eccentric, and R = 0.91 for isometric strength.

According to standard procedures, the bellies of the vastus lateralis and biceps femoris and the lateral aspect of the knee joint were shaved and alcohol-washed. Two ECG electrodes, spaced 2.5 cm apart, were placed on each muscle belly and one ground electrode on the lateral aspect of the uninvolved knee. The electrode sites were marked weekly with an indelible marker for reliable repositioning. The signals were preamplified (HDX-82, Oxford Medical, Oxford, England) with a nominal gain of 1000, input impedance of 0.5 MΩ, bandwidth of 0.4-400 Hz, and common mode rejection ratio ≥ 100 dB. The signal passed through an isolation amplifier for the subject's safety and a buffer amplifier followed by a 60-Hz notch filter. The signal was then bandpass filtered between 58 and 490 Hz at 3 dB and passed through a variable gain amplifier (21-41 dB). A linear envelope was generated using a 10-ms window averaging and the peak value of this linear envelope was digitized. The peak EMG associated with the peak concentric, eccentric, and isometric force were used in the statistical analysis. Reliability of the EMG data was R = 0.80 for concentric, R = 0.83 for eccentric, and R = 0.82 for isometric muscle actions.

Training. The training regimen consisted of 12 wk of isokinetic eccentric (ECC) or concentric (CON) quadriceps strengthening of one leg at 1.05 rad·s-1. The subjects performed 1,890 maximum effort repetitions distributed over 36 training sessions fluctuating four to six sets of 8-12 repetitions from week to week according to the periodization principle(8). Every subject exercised the left leg. To specifically exercise the quadriceps muscle group, a technician returned the subject's leg and lever arm unit to the starting position.

Subjects were constantly remined during the training sessions to relax the contralateral leg. We performed several pilot experiments to identify a position in which the contralateral leg would be relaxed. Such positions included the placement of the contralateral foot on a chair on the side or in front of the seat of the dynamometer, pulling the heel up to the seat under the buttocks, and the normal seated position-used throughout the experiment by all subjects. In six subjects of both exercise groups, the EMG activity of the vastus lateralis and biceps femoris was measured every third week in one session.

Statistical analysis. The BMDP PC-90 statistical package was used to perform the Group (CON, ECC, KON) by Condition (Concentric, Eccentric, Isometric test contraction) by Time (Pre-, Mid-, Post-training) repeated measures analysis of covariance followed by a Tukey's post-hoc contrast. The pre-training values were used as covariates to account for the initial between-group differences. The post-training values are reported with and without an adjustment according to the covariance analysis. One such analysis was done for quadriceps strength, hamstrings strength, quadriceps peak vastus lateralis EMG, and hamstring peak EMG data. The level of significance was set at P < 0.05.


Strength and EMG.Table 1 and Figure 1 show the changes in contralateral quadriceps muscle strength. There was a significant Group (Concentric, Eccentric, Control) by Condition (Concentric, Eccentric, Isometric test contraction) by Time (Pre-, Mid-, Post-training) three-way interaction (F = 7.8, P = 0.003). In the concentric group, concentric strength increased 154 N or 30% and isometric strength increased 140 N or 22% (both P < 0.05). In the eccentric group, eccentric strength increased 469 N or 77% and isometric strength increased 273 N or 39% (both P < 0.05). The eccentric group improved eccentric strength three times more than the concentric group improved concentric strength (469 N/154 N, P < 0.05). The eccentric group improved isometric strength about two times more than did the concentric group (273 N/140, P < 0.05). The eccentric group improved significantly from pre- to mid-training in eccentric and isometric strength (P < 0.05). The control group showed no significant changes (P < 0.05).

There were no significant changes in contralateral hamstrings muscle strength (P < 0.05). The largest increase, 16%, occurred in the eccentric group for the pre- to post-training comparison in the eccentric testing mode (data not shown).

Although there was no significant three-way interaction for the EMG activity of the contralateral vastus lateralis muscle (F = 1.8,P = 0.49, cell means not shown), there were trends for EMG activity during isometric and eccentric test contractions in the eccentric group to be substantially greater after training compared to the EMG activity during isometric and concentric test contractions in the concentric group. This greater EMG is reflected by the significant Group by Time two-way interaction(F = 5.6, P = 0.003) and Figure 2 shows the percent changes. Surface EMG activity increased significantly in the eccentric and concentric groups initially and in the eccentric group mid- to post-training P = 0.05). Surface EMG activity of the vastus lateralis increased 2.2 times (pre- to mid-training), 2.8 (mid- to post-training) and 2.6 more (pre- to post-training) (P < 0.05) in the eccentric than concentric group (data pooled across concentric, eccentric, and isometric testing modes). Neither training group showed significant changes in the EMG activity of the biceps femoris muscle during eccentric or concentric test contractions. No significant changes in EMG activity occurred in the control group (P > 0.05).

Contralateral muscle activity. The EMG activity of the contralateral vastus lateralis during training averaged 8% (± SD = 9%, range = 2%-12%) of the maximal isometric knee extension EMG activity.Figure 3 shows the magnitude of contralateral EMG activity in one subject performing eight eccentric contractions during training. The contralateral biceps femoris EMG activity during ipsilateral training averaged 21% (±19.6%, range 6%-29%) of maximal isometric biceps femoris EMG activity.

Ipsilateral strength gains. This portion of the experiment was reported earlier (9). In brief, eccentric training increased eccentric strength 46% (P < 0.05) and concentric training increased concentric strength 13%. Eccentric training increased concentric strength and concentric training increased eccentric strength by about the same magnitude (5 and 10%, P > 0.05). Eccentric training increased EMG activity during eccentric testing 86% and concentric training increased EMG activity during concentric testing 12%. Eccentric training increased the EMG activity measured during concentric tests 8% and concentric training increased the EMG activity measured during eccentric 11%(P > 0.05). Note that these changes were expressed as concentric-to-isometric and eccentric-to-isometric ratios(9).


The aim of the present study was to directly compare the magnitude of cross education consequent to training with muscle lengthening and shortening. This is an important issue because prior studies focused on isometric or isokinetic shortening contractions(2,4,11,26,29,32) although muscle lengthening occurs as frequently during human movements as does muscle shortening. The main finding was that ipsilateral exercise training with muscle lengthening increased contralateral strength of the homologous muscle group significantly more than did ipsilateral exercise with muscle shortening.

Our hypothesis was that cross education is greater following training with muscle lengthening than shortening. In sedentary subjects there is a significantly greater difference between measured and expected forces during maximal effort, voluntary muscle lengthening compared with shortening based on forces evoked by a combination of voluntary effort and electromyostimulation(31). The cause of this greater force deficit is probably due to some neural inhibition that protects against excessive muscle tension during muscle lengthening (30,31). It is thus conceivable that training with muscle lengthening would result in a greater neural adaptation both on the trained (9) and untrained side. The current results seem to confirm this hypothesis.

The magnitude of cross education was a dramatic 77% following eccentric training and it was 30%, as reported in the literature, following concentric training when tested with the same contraction mode used during training(Table 1). Such large cross education effects are in line with data from a prior study that reported up to 135% cross education following training with pulsed eccentric loading (1). The present results suggest a certain degree of specificity. Cross education was the greatest in the same contraction mode subjects used for training. There was a somewhat less of a specificity associated with eccentric training because eccentric compared with concentric training increased isometric strength significantly more. Because subjects did not train with isometric contractions, a significant increase in isometric strength must be due to factors other than learning. It is interesting to note that cross education was only about 16% when tested with a contraction mode opposite to the one used during training. This finding would suggest eccentric and concentric contractions are associated with unique motor unit activation and that the process of inducing cross education may be different for training with concentric, eccentric, and isometric contractions.

The time course of cross education was different between eccentric and concentric training. At week 6, eccentric training caused 20% and 40% cross education in isometric and eccentric strength, respectively, about the same amount concentric training yielded in 12 wk. Thus, cross education occurs substantially faster with eccentric than concentric training. This finding may be important in an in-patient rehabilitation setting where time is limited due to treatment cost and patient availability. These data, however, should not be interpreted to include maximal effort eccentric actions in a rehabilitation program; instead we suggest to incorporate eccentric actions in paradigms that would conventionally use pure concentric actions.

Cross education is a form of neural adaptation. Figure 2 shows that there was a greater increase in EMG activity following eccentric than concentric training. Because muscle fiber size of the contralateral vastus lateralis did not change (9), we interpret the increase in surface EMG activity as an increase in muscle activation due to an increase in central drive (24). However, interpretation of increased surface EMG activity with resistive exercise training can be complicated. If in fact muscle activation increases with (eccentric) training due to an increased central drive(24,30) surface EMG activity is expected to increase. Yet one could also expect a decrease or no change in surface EMG activity under such conditions, because as muscles become more efficient with conditioning the same amount of work is done by less muscle(20), resulting in reduced surface EMG. Cannon and Cafarelli (4) reported contralateral strength gains without changes in surface EMG activity, suggesting that central motor adaptation may not have manifested in an increased neural drive or EMG activity. Whether or not we are capable to measure neural adaptations with surface EMG, the nature of cross education adaptation must be neural because no or little changes occur in muscle size(9,11,12).

The concept of common central drive distributed to homologous extremity muscles has recently been revisited (5). Using multiunit cross-correlation analyses, these authors presented evidence for a common central input being shared by diaphragm, rectus abdominis, and masseter muscle pairs that are normally coactivated, but saw no evidence for a common drive shared by left and right homologous biceps and deltoid muscles that often act independently. The functional significance of these findings become apparent for individuals with unilateral cerebrovascular accidents. The clinical usefulness of cross education remains to be tested because in such patients axial muscle functions (e.g., respiration, mastication) are generally preserved, while the extremity muscles of the involved side become severely debilitated (6).

Although neural adaptation to resistive exercise has been proposed to occur in the form of learning due to postular stabilization(23), such a mechanism is not likely to fully explain the magnitude of neuromuscular adaptations we observed in the contralateral muscles. The finding that the amount of contralateral coactivation was, as shown in Figure 3, mild and actually greater in the hamstrings than quadriceps muscles but not different between concentric and eccentric training-precludes the explanation that contralateral strength gains were due to postural contractions. The minimum stimulus known to result in strength gains is about 60% of maximum (8), substantially higher than the ≈15% forces back-calculated from the force-EMG relationship for the quadriceps.

The precise afferent and efferent mechanisms underlying cross education are unknown and could not be identified in the present study. Lagassé(16) and Smith (28) suggested that the crossed extension-reflex (27) may be the mediating mechanism. However, the crossed extension-reflex is under powerful supraspinal influence and its action in humans could be heavily modified compared to its operation in the spinal cat (22). Cross education is certainly a robust phenomenon because hemiparetic patients revealed neuromuscular activity in the contralateral homologous muscles in response tounresisted ipsilateral exericse-considered as evidence for afferent-system mediated effects (18). In contrast, unilateral tasks are often associated with bilateral activation of the motor cortex (15). Another sensory mechanism could be related to ipsilateral cutaneous stimulation. For example, ipsilateral cutaneous stimulation resulted in an increased motoneuron excitability of the contralateral homologous muscle groups (19,21). In addition, an increase in maximal voluntary ipsilateral isometric force occurred when the effort was coupled with contralateral electromyostimulation(13). A modulation of the descending command and/or the stimulation-evoked pain might mediate such an effect. The neuroanatomical basis of an efferent mechanism is probably associated with the 2% undecussated portion of the anterolateral corticospinal tract (22). This is a minimal but perhaps a sufficient neuroanatomical representation of central motor drive intended ipsilaterally that is present to also modulate muscle activation contralaterally. As stated earlier, the force deficit between measured and expected forces is greater during eccentric compared concentric actions (31), so improved muscle activation ipsilaterally due to eccentric training also presents a potential for improved muscle activation contralaterally. Because muscle activation is expected to increase with eccentric compared to concentric training(10), contralateral effects were also greater.

In summary, the present data suggest a substantially greater cross education coupled with EMG activity following ipsilateral eccentric compared to concentric training of the quadriceps in man. It is concluded that the greater cross education following training with muscle lengthening is most likely being mediated by both afferent and efferent mechanisms that allow previously sedentary subjects to achieve a greater activation of the untrained limb musculature.

Figure 1-Percent pre- to post-training changes in contralateral quadriceps muscle strength as assessed with concentric (
Figure 1-Percent pre- to post-training changes in contralateral quadriceps muscle strength as assessed with concentric (:
filled column), eccentric (open column), and isometric (hatched column) test contractions, following ipsilateral concentric or eccentric training. Vertical bars denote ± SD* Significant change and significantly more change than the concentric group. Note: The control group's data in absolute units are shown inTable 1.
Figure 2-Changes in contralateral vastus lateralis EMG activity following ipsilateral quadriceps eccentric (
Figure 2-Changes in contralateral vastus lateralis EMG activity following ipsilateral quadriceps eccentric (:
filled column) or concentric (open column) strength training at 1.05 rad·s-1. * significant change and significantly more change than the other group (Pb <0.05). Bars are ± SD. Note: Data pooled for concentric, eccentric, and isometric testing modes.
Figure 3-Recordings of force (A, highest force in this tracing is 970 N), knee joint angle (B, 0°-90° knee angle), and EMG activity of the vastus lateralis (C) and biceps femoris (D) in one subject performing eight eccentric contractions at 1.05 rad·s-1 during the second week of training. Remaining tracings correspond to contralateral vastus lateralis (E) and biceps femoris (F). Calibration for all EMG channels is 1 mV for the total window height. Note a lack of contralateral quadriceps activity and the large biceps femoris activity during ipsilateral eccentric contractions.
Figure 3-Recordings of force (A, highest force in this tracing is 970 N), knee joint angle (B, 0°-90° knee angle), and EMG activity of the vastus lateralis (C) and biceps femoris (D) in one subject performing eight eccentric contractions at 1.05 rad·s-1 during the second week of training. Remaining tracings correspond to contralateral vastus lateralis (E) and biceps femoris (F). Calibration for all EMG channels is 1 mV for the total window height. Note a lack of contralateral quadriceps activity and the large biceps femoris activity during ipsilateral eccentric contractions.


1. Anson, M. R., A. A. Halpern, and P. M. Clarkson. Pulsed eccentric loading effects on cross-education. Med. Sci. Sports Exerc. 25(Suppl.):S164, 1993.
2. Brown, A. B., N. McCartney, and G. S. Digby. Positive adaptations to weight-lifting training in the elderly. J. Appl. Physiol. 69:1725-1733, 1990.
3. Cabric, M. and H. J. Appell. Effect of electrical stimulation of high and low frequency on maximum isometric force and some morphological characteristics in men. Int. J. Sports Med. 8:256-260, 1987.
4. Cannon, R. J. and E. Cafarelli. Neuromuscular adaptations to training. J. Appl. Physiol. 63:2396-2402, 1987.
5. Carr, L. J., L. M. Harrison, and J. A. Stephens. Evidence for bilateral innervation of certain homologous motoneurone pools in man.J. Physiol. 475:217-227, 1994.
6. Colebatch, J. G. and S. C. Gandevia. The distribution of muscular weakness in upper motor neuron lesions affecting the arm.Brain 112:749-763, 1989.
7. Davies, C. T. M., P. Dooley, M. J. N. McDonagh, and M. J. White. Adaptation of mechanical properties of muscle to high force training in man. J. Physiol. 365:277-284, 1985.
8. Fleck, S. J. and W. J. Kraemer. Designing Resistance Training Programs. Champaign, IL: Human Kinetics, 1987. pp. 20-46.
9. 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.
10. Hortobágyi, T. and F. I. Katch. Eccentric and concentric torquevelocity relationships during arm flexion and extension: Influence of strength level. Eur. J. Appl. Physiol. 60:395-401, 1990.
11. Housh, D. J., T. J. Housh, G. O. Johnson, and W. K. Chu. Hypertrophic response to unilateral concentric isokinetic resistance training. J. Appl. Physiol. 73:65-70, 1992.
12. Houston, M. E., E. A. Froese, S. P. Valeriote, H. J. Green, and D. A. Ranney. Muscle performance, morphology and metabolic capacity during strength training and detraining: a one leg model. Eur. J. Appl. Physiol. 51:25-35, 1983.
13. Howard, J. D. and R. M. Enoka. Enhancement of maximum force by contralateral-limb stimulation. J. Biomech. 20:980, 1987.
14. Howard, J. D. and R. M. Enoka. Maximum bilateral contractions are modified by neurally mediated interlimb effects. J. Appl. Physiol. 70:306-316, 1991.
15. Kristeva, R., D. Cheyne, and L. Deeke. Neuromagnetic fields accompanying unilateral and bilateral voluntary movements: topography and analysis of cortical sources. EEG Clin. Neurophysiol. 81:284-298, 1991.
16. Lagasse, P. P. Muscle strength: Ipsilateral and contralateral effects of superimposed stretch. Arch. Phys. Med. Rehabil. 55:305-310, 1974.
17. Matyas, T. A., M. P. Galea, and S. D. Spicer. Facilitation of maximum voluntary contraction in hemiplegia by concomitant cutaneous stimulation. Am. J. Phys. Med. 65:125-134, 1986.
18. Mills, V. M. and L. Quintana. Electromyography results of exercise overflow in hemiplegic patients. J. Am. Phys. Ther. Assoc. 65:1041-1045, 1985.
19. Pierrot-Deseilligny, E., B. Bussel, G. Sideri, H. P. Cathala, and P. Castaigne. Effect of voluntary contraction on H reflex changes induced by cutaneous stimulation in normal man. EEG Clin. Neurophysiol. 34:185-192, 1973.
20. Ploutz, L. L., P. A. Tesch, R. L. Biro, and G. A. Dudley. Effect of resistance training on muscle use during exercise. J. Appl. Physiol. 76:1675-1681, 1994.
21. Robinson, K. L., J. S. McIlwain, and K. C. Hayes. Effects of H-reflex conditioning upon the contralateral alpha motoneuron pool.EEG Clin. Neurophysiol. 46:65-71, 1979.
22. Rothwell, J. C. Control of Human Voluntary Movement. London: Croom-Helm, 1987. pp. 105-125.
23. Rutherford, O. M., C. A. Greig, A. J. Sargeant, and D. A. Jones. Strength training and power output: transference effects in the human quadriceps muscle. J. Sports Sci. 4:101-107, 1986.
24. Sale, D. G. Neural adaptation to resistance training.Med. Sci. Sports Exerc. 20:S135-S145, 1988.
25. Scripture, E. W., T. L. Smith, and E. M. Brown. On education of muscular control and power. Studies Yale Psychol. Lab. 2:114-119, 1894.
26. Shaver, L. G. Cross transfer effects of conditioning and deconditioning on muscular strength. Ergonomics 18:9-16, 1975.
27. Sherrington, C. S. Flexion-reflex of the limb, crossed extension-reflex, and reflex stepping and standing. J. Physiol. 40:28-121, 1910.
28. Smith, L. E. Facilitatory effects of myotatic stretch training upon leg strength and contralateral transfer. Am. J. Phys. Med. 49:132-141, 1970.
29. Stromberg, B. V. Contralateral therapy in upper extremity rehabilitation. Am. J. Phys. Med. 65:135-143, 1988.
30. Tesch, P. A., G. A. Dudley, M. R. Duvoisin, and B. M. Hather. Force, and Emg signal patterns during repeated bouts of concentric or eccentric muscle actions. Acta Physiol. Scand. 138:263-271, 1990.
31. Westing, A. H., J. Y. Seger, and A. Thorstensson. Effects of electrical stimulation on eccentric and concentric torque-velocity relationships during knee extension in man. Acta Physiol. Scand. 140:17-22, 1990.
32. Wickiewicz, T. L., R. R. Roy, P. L. Powell, J. J. Perrine, and V. R. Edgerton. Muscle architecture and force-velocity relationships in humans. J. Appl. Physiol. 57:435-443, 1984.


©1997The American College of Sports Medicine