Historical and recent observations suggest that unimanual motor practice increases the motor output not only in the practiced but also in the unpracticed homologous muscle (3,9,29). Neural mechanisms, which are unknown, must mediate such cross education because it occurs in the absence of muscle hypertrophy. One possibility is that unimanual practice leads to adaptive changes in the primary motor cortex ipsilateral to the practicing side (iM1) that is willfully not involved in the practice, and this increased excitability mediates the ensuing cross education. Indeed, EEG (38), fMRI (7), MEG (35), PET (19), and TMS studies (16,26,27,30) show that activation of iM1 and the excitability of the corticospinal path targeting the resting hand increase during unilateral voluntary isometric contractions. The increase in iM1 excitability increases with increasing contraction intensity of hand and wrist muscles (16,27,30), and the increase in neural drive from iM1 to the untrained muscles after unilateral training (23) all strongly point to the possibility that iM1 is a prime mediator of cross education.
Execution of a motor task with one hand also modifies inhibition that controls the balance of excitability between the task and the nontask hemispheres. In early experiments, interhemispheric inhibition (IHI) deepened with a slight unimanual muscle contraction (11), but recent experiments with careful adjustments of the conditioning and test pulse size found that IHI actually decreased from the active to the nonactive M1 (30). In addition, in cross-sectional experiments, unimanual serial reaction time task transferred implicit knowledge and unimanual pinching task transferred speed and accuracy to the nonpracticed hand so that the transferred motor behaviors were associated with the magnitude of decrease in IHI (2,32). It thus seems reasonable to hypothesize that chronic unimanual practice involving, as in numerous studies before (3,9,29), effortful isometric contractions would produce cross education and IHI would mediate this interlimb transfer of the ability to exert maximal force. Therefore, the purpose of the present study was to determine whether increases in excitability of the nontask M1 or reductions in IHI from the task to the nontask M1 contribute to cross education produced by chronic unimanual motor practice comprising serial effortful isometric contractions of a finger muscle.
Participants and Study Design
Twenty healthy adults volunteered for the study (12 men and 8 women; mean ± SE age = 30.9 ± 1.4 yr, height = 1.75 ± 0.02 m, mass = 72.4 ± 3.3 kg). Participants gave their informed consent for the experimental procedures that were approved by the local ethics committee. The study was performed in accordance with the Declaration of Helsinki. Participants were interviewed about their general health on the telephone and were included only if they were right-handed and, based on a medical examination, had no history of neurological disorders, including no head or hand injuries. After medical and neurological screening, volunteers visited the laboratory for a 1-h session and were familiarized with the equipment and testing and training procedures. At the time of the study, none of the subjects were practicing any activity that involved strengthening of the fingers. Volunteers were randomly assigned to an experimental (n = 12, 8 men and 4 women) and a control group (n = 8, 6 men and 2 women).
The stated goal of the program to the subjects was to increase muscle strength in the right first dorsal interosseus (FDI) without providing explicit details about a potential contralateral effect. Participants attended twenty 15-min-long training sessions on nonconsecutive days during an 8-wk period. At the start of each exercise session, subjects performed one trial of 5-s-long force production at what they perceived was 50%, 75%, and 90% maximal voluntary contractions (MVC) as a warm-up and then performed one 100% MVC to assess strength gains and used this value as a base to set the 80% target of training intensity. Subjects in the experimental group exercised the right FDI by performing five blocks of 10 isometric index finger abduction at an intensity of 80% of MVC, set as a target on an oscilloscope (model TDS220; Tektronix, Richardson, TX). The duration of each contraction and the intercontraction rest interval were both 5 s long. Participants ramped force production to the target in 0.5 s, held it for 4.0 s, and reduced force in 0.5 s. The interblock rest period was 60 s. To minimize mirror activity in the contralateral left FDI (14,33,40), we used 80% instead of 100% exercise intensity and subjects received the following standardized instruction before each block of contractions: "Get ready to contract on the right and completely relax the left side." In addition, the investigator reminded the subject to relax the arm and shoulder muscles and checked this by palpation. Auditory feedback of EMG activity from the contralateral left FDI also helped subjects to maintain relaxation in the left arm throughout each block of exercise. Preliminary experiments showed that the exercise program caused minimal fatigue because the MVC taken before and 1 min after the end of a session were not significantly different. Members of the control group did not exercise and participated in testing only. They came to the laboratory, placed their right hand in the dynamometer, and quietly sat or read a newspaper.
Sessions 1, 5, 10, 15, and 20 were longer, about 1.5 h, than the other sessions because maximal voluntary force MVC of the right FDI and left FDI and left abductor minimi digiti (ADM) was measured, and peripheral nerve stimulation and TMS experiments were conducted before and after the training bouts. ADM was used to determine the spatial specificity of cross education. In all experiments, subjects were comfortably seated in a reclining padded chair, equipped with an armrest on each side and an adjustable head support.
A custom-designed hand dynamometer, described in details previously (15), was used for strength training and testing of the right FDI, and it was reconfigured for the testing of the left FDI and ADM. Briefly, the dynamometer's plexiglass base was affixed to the chair's armrest, supporting the hand and forearm. With the right wrist pronated, the palm rested on the plexiglass base. The index finger was isolated with the thumb extended and abducted and the third, fourth, and fifth fingers extended and, with a Velcro strap, abducted. The center of the proximal interphalangeal joint of the extended index finger was aligned with the center of the load cell (model 31; Honeywell-Sensotec, Columbus, OH). The wrist was stabilized with a Velcro strap, wrapped around the chair's armrest and the dynamometer's plexiglass base. The optimal position of the load cell was determined for each subject and adjusted individually. Such a setup ensured that the sole source of force production was index finger abduction. In separate measurements with the same dynamometer reconfigured, we measured MVC in the left FDI and ADM. Subjects were familiarized with the procedure by performing one trial of 5-s-long force production at 50%, 75%, and 90% MVC with each finger. Subjects were then instructed to perform two trials of 5-s-long MVC with 1 min of intertrial rest, and the force was recorded from these trials on the computer. Participants received biofeedback of their force production from an oscilloscope. The order of testing of the right FDI, left FDI, and left ADM was systematically rotated between subjects. MVC of the right FDI was measured in every session in the experimental group, and it was also measured in the left FDI and ADM in sessions 1, 5, 10, 15, and 20. MVC of the three fingers in the control group were measured in sessions 1, 5, 10, 15, and 20. Table 1 summarizes the schedule of measurements.
The EMG activity of the right FDI and left FDI and ADM was recorded using silver-silver chloride surface EMG electrodes (2-cm center-to-center interelectrode distance), placed over these muscles in a belly-tendon montage. EMG signals were amplified using a Nicolet Viking electromyography system (Madison, WI) and band-pass-filtered between 10 and 2000 Hz. Signals were digitized at a frequency of 5 kHz and fed into a laboratory computer for further offline analysis. The EMG activity was recorded during cortical and peripheral nerve stimulation. EMG activity was also recorded from the left FDI during the training contractions in sessions 1, 5, 10, 15, and 20 to determine the magnitude of mirror EMG activity as a percent of MVC.
Transcranial magnetic stimulation and peripheral nerve stimulation.
Before transcranial magnetic stimulation (TMS), a cap (Electro Caps International, Inc., Eaton, OH) was placed on the subject's head. For each subject, we marked individual anatomical landmarks on the cap that allowed us to place the cap on the head in the same position within and across sessions. The optimal coil position was also marked in a coordinate system drawn on the surface of the cap. Thus, the coil was placed on the surface of the cap in the same position within and across sessions. These procedures made it possible to stimulate the same area of the cortex within and across sessions. Motor evoked potentials (MEP) were elicited by TMS delivered from a Magstim 200 stimulator (Magstim Company, Dyfed, UK) through a figure-8 coil (external loop diameter, 8 cm; type number, SP15602) with a monophasic current waveform. The coil was moved in 0.5-cm steps over M1 to identify the optimal scalp position, i.e., hot spot, for activation of the target muscle for the right FDI overlying left M1 and left FDI and ADM overlying the right M1. The hot spot correlates well with stimulation of Brodmann area 4. The intersection of the coil was placed tangentially to the scalp with the handle pointing backward and laterally at a 45° angle away from the midline over the hot spot for the target muscle to activate the corticospinal system preferentially trans-synaptically via horizontal corticocortical connections. The direction of the current flow was posterior-to-anterior across M1. The coil was secured with a coil holder to ensure that the same area of the cortex was stimulated within a session and across sessions. Single pulse stimuli were delivered at an interstimulus interval varying between 5 and 6 s. During experiments, MEP were displayed on the monitor of the data collection computer, visually inspected, and stored on a computer for offline analysis. TMS measures included resting motor threshold (rMT), MEP amplitudes using recruitment curves (RC), short intracortical inhibition (SICI) and intracortical facilitation (ICF), IHI from left M1 to right M1, and MEP at 120% and 160% rMT conditioned with the contraction of the right FDI at 20 and 80% MVC.
rMT was determined with 1% increment in stimulator output as the minimal stimulus intensity required to produce MEP of at least >50 μV in peak-to-peak amplitude in at least 5 of 10 consecutive trials.
RC was measured in the left FDI and ADM with both hands at rest before and after sessions 1, 10, and 20 (Table 1). For the RC, stimulation started at 10% below rMT and increased in 10% steps of maximal stimulator output until the MEP amplitude did not experience additional increases or the maximal stimulator output was reached. The stimulation intensities were administered in a pseudorandom order, with the highest intensity always presented last. There were seven trials to determine MEP amplitude at each stimulation intensity. The amplitude of MEP was measured peak-to-peak, averaged offline, and expressed as a percentage of the maximal motor response (M max). To determine M max, the ulnar nerve was stimulated (1-ms rectangular pulse, model Viking IV; Nicolet Biomedical, Madison, WI) with supramaximal intensity using bipolar surface electrodes placed at the wrist with the subject's hand in the dynamometer. The stimulating electrode was placed the intersection of the two lines representing the flexor carpi ulnaris tendon and the ulnar styloid. The intensity of stimulation was increased from a subthreshold level until there was no further increase in the peak-to-peak amplitude of the M-wave with increasing stimulation intensity. At this final stimulation intensity, we recorded the twitch-evoked force in three trials. We normalized each participant's TMS data to the individual M max, making it possible to compare MEP amplitudes between different test sessions. Stimulation intensity was expressed as a percent of each participant's rMT. For the sake of group comparisons, all stimulus intensities were normalized to rMT of a stimulus-response curve determined before an intervention during the first training session. Such a normalization process makes it possible to reliably determine shifts in RC due to a specific intervention.
SICI, ICF, and IHI.
SICI and ICF, in one block, and IHI, in another block, were measured in the left FDI and ADM with both hands at rest before and after sessions 1, 5, 10, 15, and 20 (Table 1). We used the methods described by Kujirai et al. (21). The conditioning stimulus was set at ∼80% of rMT, an intensity that does not affect spinal excitability (8). The test stimulus was delivered 2 and 10 ms after the conditioning stimulus for SICI and ICF, respectively. Because there are two distinct phases of inhibition (12), we used the 2-ms interstimulus (21) interval to avoid a mixture of the two phases. Under these conditions, the conditioning stimulus produced ∼50% inhibition and ∼30% facilitation at 2- and 10-ms interstimulus intervals. The intensity used for the conditioning stimulus ranged between 29% and 56% of stimulator output, corresponding to 78.9% of the rMT, and was kept constant within and between sessions. The size of the test stimulus was set at ∼1.0 mV and was also kept constant. Seven test stimuli, seven SICI, and seven ICF were presented at random at rest and repeated two times to establish the stimulation parameters. There was 5 s of rest between trials. We also assessed ICF with a conditioning stimulus set at 90% of rMT as such a stimulus intensity is less likely to produce changes in excitability of motoneurons at the spinal level (27) and found similar results at 80% and 90% conditioning stimulus (n = 6, sessions 1, 10, and 20; data not shown).
In a separate block administered in an alternating order with the SICI and ICF block, we measured IHI from left to right M1 in the left FDI and ADM using two custom-built figure-8 coils (loop diameter = 8 cm; type numbers SP15602 and SP15001). The two coils were positioned at the optimal location for activating the left and right FDI, respectively, and in a separate run for the left ADM. The handles of the coils pointed 45° backward and laterally relative to the midsagittal line. A suprathreshold conditioning stimulus was set at an intensity that elicited inhibition of ∼75% to allow its up- or down-regulation. The intensity used for the conditioning stimulus ranged from 51% to 82% (69.9 ± 8.3 of stimulator output), corresponding to 126.4% of the rMT. The conditioning stimulus was delivered to the left M1 10 ms before the test stimulus given to the right M1. The size of the test stimulus was set at ∼1.0 mV within and between sessions. Twelve test stimuli and 12 IHI were presented at random with 5 s of rest between each trial. In these double-pulse experiments, the control MEP and the conditioned MEP were normalized to M max, averaged, and the conditioned MEP expressed relative to the control MEP, yielding one data point per subject for SICI, ICF, and IHI before and after sessions 1, 5, 10, 15, and 20 in the left FDI and ADM.
MEP conditioned with muscle contraction.
We evaluated the effect of exercise training of the right FDI on the size of MEP measured in the left FDI conditioned with a voluntary contraction of the right FDI before sessions 1, 5, 10, 15, and 20 (Table 1). In these experiments, the left FDI, from which the MEP were recorded, was always at rest. Block 1 consisted of TMS given at 120% of the rMT at rest and conditioned with 20% MVC. Block 2 consisted of TMS given at 160% of the rMT at rest and conditioned with 80% MVC. The order of the two blocks and the five trials at rest and five trials with contraction within each block, respectively, was rotated between subjects, but the order was kept the same in each subject across the five testing sessions. There was 5 s of rest between trials. TMS was delivered at 2.5 s of the 5-s-long conditioning contractions. Within each block, the control and conditioned MEP were normalized to M max and then, respectively, averaged, and the conditioned MEP were expressed relative to the MEP at rest, yielding one data point per block per subject before sessions 1, 5, 10, 15, and 20 in the left FDI.
Data in the text and figures are reported as mean ± SE. We tested for normal distribution with the Kolmogorov-Smirnov tests for each variable. Mauchly test of sphericity, conducted for each repeated-measures analysis, was followed by a group (experimental, control) by session (1, 5, 10, 15, and 20) ANOVA with repeated measures on session to determine changes in MVC of each of the three muscles and in mirror EMG activity. The amplitude of the maximal compound action potentials, M max, and the associated evoked twitch forces were measured at the start and end of sessions 1, 5, 10, 15, and 20 and were analyzed with a group (two) × session (five) × time (two) ANOVA in the right FDI and left FDI and ADM.
Acute changes in rMT, SICI, ICF, and IHI, using the before and after data measured within sessions 1, 5, 10, 15, and 20, respectively, were analyzed with a group (two) × time (two) ANOVA with repeated measures on time. Acute changes MEP size (RC), using the before and after data measured within sessions 1, 10, and 20, respectively, were analyzed with a group (two) × time (two) × stimulation intensity (eight) ANOVA with repeated measures on time and intensity.
Chronic changes in rMT, SICI, ICF, and IHI measured before sessions 1, 5, 10, 15, and 20 in the two groups' left FDI and ADM muscle, respectively, were analyzed with a group (two) × session (five) ANOVA with repeated measures on session. Chronic changes in MEP size measured at eight intensities before sessions 1, 10, and 20 in the two groups' left FDI and ADM muscle, respectively, were analyzed with a group (two) × session (three) × intensity (eight) ANOVA with repeated measures on session and intensity. In case of significant interaction effects, we used Tukey post hoc to determine the means that were different at P < 0.05. Pearson correlation analysis was used to determine correlations as needed.
Voluntary Force and EMG
Figure 1A shows the group × session interaction in the right, trained, FDI (F 4,68 = 27.0, P = 0.001). The training program increased maximal voluntary force from baseline to session 20 by 14.5 ± 1.5 N or 49.9% ± 6.3% without changes in the control group. Figure 1B shows the group by session interaction in the left, untrained, FDI (F 4,68 = 18.9, P = 0.001). The time course of adaptation revealed 15.6% ± 1.7% cross education at session 10 (P < 0.05) that further increased 6.2% ± 2.0% by session 20 (P < 0.05), with a total amount of cross education of 5.7 ± 0.6 N or 21.8% ± 2.3% (both P < 0.05) from baseline to session 20. There were no changes in MVC of the control group's untrained FDI (−1.9 ± 3.5). The changes in maximal voluntary force of the ADM were not significant in the experimental (−0.8% ± 3.8%) and control groups (−0.5%± 3.5%) (group × time interaction, F 4,68 = 0.5, P = 0.747; Fig. 1C). There was a group × session interaction in the left, untrained, FDI's EMG activity normalized for M max (F 4,68 = 3.9, P = 0.006) with an increase of 27.6% ± 2.9% in the experimental group and 6.1% ± 3.0% in the control group (not significant). The changes in surface EMG activity were not significant in the experimental (3.1% ± 2.6%) and control group's (3.2% ± 3.4%) left ADM.
Responses to Peripheral Nerve Stimulation
There were no group by session by time or other two-way interactions for M max or twitch forces in the right FDI and left FDI and ADM. For example, the pre-to-post changes in M max of the right, trained, FDI were 2.0% ± 1.7% (P = 0.258) and the changes in the evoked forces were 3.6% ± 4.5% (P = 0.482).
Responses to TMS
There were no significant acute (within session) and chronic (across sessions 1, 5, 10, 15, and 20) changes in rMT. The largest pre-to-post change in rMT of all groups and muscles was a reduction from 50.3% ± 4.1% in session 1 to 47.5% ± 4.6% in session 20 in the right, trained, FDI (P = 0.111).
For the acute changes in MEP size measured at eight stimulation intensities before and after sessions 1 and 20, respectively, in the two groups' left FDI and ADM muscle, there were no significant main, two-way, or three-way interaction. For the chronic changes in MEP size measured at eight intensities before sessions 1, 10, and 20 in the two groups' left, untrained, FDI, there were a group × session (F 2,34 = 7.4, P = 0.002) and a group × session × stimulation intensity interaction effects (F 14,238 = 2.4, P = 0.004). Figure 2A shows that MEP size increased 6% (P = 0.010) from the 1st to the 20th session. In the control group's left FDI, MEP size did not change between the 1st (41.2% ± 4.7%) and the 20th session (43.1% ± 5.1%). None of main and interaction effects were significant in the ADM (Fig. 2B).
SICI and ICF.
There were no significant acute (within session) and chronic (across sessions) changes in SICI and ICF in the experimental and control groups' FDI and ADM. As a percent of the test MEP, the level of SICI across all sessions was 52.4% ± 2.9% and 53.9% ± 4.1% in the experimental and control group's left FDI. The largest change in SICI occurred in the experimental group's FDI in session 5 where SICI diminished from 48.0% ± 2.4% before exercise to 57.1% ± 3.9% of the test MEP after exercise without the group by time interaction reaching significance (F 1,17 = 1.4, P = 0.246). The overall level of ICF across all sessions was 137.0% ± 9.1% and 134.1% ± 6.2% of the test MEP in the experimental and control group's left FDI. The largest change in ICF occurred in session 5 where ICF increased from 133.0% ± 8% to 149.0% ± 11.0% of the test MEP in the experimental group's left FDI (group × time interaction, F 1,17 = 3.9, P = 0.064). In the SICI and ICF experiments, the size of the test pulse was constant within and across sessions. For example, the size of the test pulse in the experimental group's left FDI was 1.02 ± 0.06, 1.01 ± 0.05, 1.06 ± 0.06, 1.05 ± 0.07, and 1.10 ± 0.08 mV in sessions 1, 5, 10, 15, and 20, respectively (session effect, F 4,72 = 3.72, P = 0.071), with similar values in the control group (group × session interaction, F 4,72 = 0.9, P = 0.457). The size of the test pulse was similar (∼1.0 mV) in the experimental and control group's left ADM and was also stable across sessions.
Figure 3A shows a single-subject example for IHI in the left FDI and Figure 3B shows the experimental group's data. In session 1, the magnitude of IHI in the experimental and control group's left FDI was statistically similar 74.3% ± 2.4% and 79.8% ± 3.2% of the test MEP. IHI in the two group's ADM was 76.9% ± 3.3% and 75.5% ± 3.9% of the test MEP. In the experimental but not in the control group's left FDI, IHI acutely diminished 8.8% ± 3.9% in session 1 (Tukey post hoc, P = 0.159), 11.6% ± 2.9% in session 5 (P = 0.016), 7.6% ± 4.5% in session 10 (P = 0.254), 10.4% ± 2.6% in session 15 (P = 0.048), and 6.3% ± 3.8% in session 20 (P = 0.141). The average acute within-session diminishment in IHI was 8.9%. In the control group, the average acute within-session change in IHI was 1.8%. Relative to session 1, in the experimental but not in the control group's left FDI, IHI chronically diminished (group × session interaction, F 4,72 = 8.2, P = 0.010) 5.4% ± 2.8% by session 5 (Tukey post hoc, P = 0.111), 23.4% ± 5.1% by session 10 (P = 0.002), 28.1% ± 3.8% by session 15 (P = 0.001), and 30.9% ± 3.8% by session 20 (P = 0.001). In the control group, the average chronic change in IHI between session 1 and session 20 was 3.5%. No significant acute and chronic changes occurred in IHI in the experimental group's ADM (Fig. 3C). In the IHI experiments, the size of the test pulse was constant within and across sessions. For example, the size of the test pulse in the experimental group's left FDI was 1.06 ± 0.13, 0.99 ± 0.18, 1.12 ± 0.18, 1.15 ± 0.27, and 1.11 ± 0.28 mV in sessions 1, 5, 10, 15, and 20, respectively (session effect, F 4,72 = 3,72, P = 0.268), with similar values in the control group (group × session interaction, F 4,72 = 1.9, P = 0.222). The size of the test pulse was similar (∼1.0 mV) in the experimental and control groups' left ADM and was also stable across sessions.
Cortical excitability during muscle contraction.
Figure 4 shows representative trials in one subject (A, B) and also the experimental and control groups' data for changes in cortical excitability measured in the left FDI, whereas the right M1 received TMS at an intensity of 120% (C) and 160% (D) of the rMT during rest and the right FDI was contracted at 20 and 80% MVC. At 120% TMS and 20% MVC, cortical excitability remained unchanged initially but, relative to session 1, increased 5.4% ± 1.5% by session 15 (P = 0.002) and 10.3% ± 2.3% by session 20 (P = 0.001) (group × session interaction, F 4,68 = 9.0, P = 0.008) (Figs. 4A, C). At 160% TMS and 80% MVC, cortical excitability, relative to session 1, increased 26.6% ± 3.8% by session 5 (P = 0.001), 40.7% ± 5.7% by session 10, 67.0% ± 7.2% (P = 0.001) by session 15, and 63.9% ± 8.4% by session 20 (P = 0.001) (group × session interaction, F 4,68 = 3.8, P = 0.012) (Figs. 4B, D).
The percent changes in cross education in the left, untrained, FDI were unrelated to the percent changes in strength gains in the right, trained, FDI in sessions 5 and 10, but by session 15 (r = 0.55, P = 0.021) and session 20 (r = 0.81, P = 0.001), cross education became strongly associated with strength gains. Subjects were instructed to contract the right FDI and keep the left FDI completely relaxed to avoid mirror EMG activity. The mirror EMG activity, expressed as a percent of the EMG during an MVC, in the left FDI measured in sessions 1 (0.8% ± 0.8%), 5 (2.0% ± 1.0%), 10 (1.2% ± 0.9%), 15 (1.2% ± 1.3%), and 20 (1.5% ± 0.9%), averaging 1.3% (±1.0%) did not change, and this mirror activity did not correlate with the magnitude of cross education (y = −0.15x + 28.6, r = −0.052, P = 0.873).
The changes in MEP size measured with RC at rest did not correlate with changes in cross education at session 10 (r = 0.33, P = 0.172), and this relationship remained unchanged by session 20 (r = 0.20, P = 0.293). The changes in the amount of facilitation of the MEP when the right FDI contracted weakly (20% MVC) and the right M1 was stimulated with low intensity TMS (120% rMT) did not correlate with the changes in cross education at session 5 (r = 0.19, P = 0.549), session 10 (r = 0.43, P = 0.167), and session 15 (r = 0.49, P = 0.105), but the correlation was significant at session 20 (r = 0.68, P = 0.014). Figure 5 shows that the changes in the amount of facilitation of the MEP when the right FDI contracted strongly (80% MVC) and the right M1 was stimulated with high-intensity TMS (160% rMT) correlated significantly with the changes in cross education already at session 5 (r = 0.58, P = 0.047), at session 10 (r = 0.66, P = 0.020), at session 15 (r = 0.65, P = 0.022), and at session 20 (r = 0.88, P = 0.001).
Figure 5 also shows that the percent changes in cross education were initially unrelated to the percent changes in IHI in session 5 (r = 0.21, P = 0.511), but moderate relationships emerged by session 10 (r = 0.61, P = 0.036) and session 15 (r = 0.66, P = 0.020) that further strengthened by session 20 (r = 0.72, P = 0.008). At no time point was there a significant correlation between the changes in IHI and changes in MEP size measured with RC at rest (P > 0.05).
Repeated effortful contractions of a finger muscle increased maximal voluntary force of the homologous muscle in the unexercised contralateral hand. A reduction in IHI at rest from the trained to the untrained hemisphere and an up-regulation of iM1 excitability, especially during muscle contraction, mediated this cross-education effect. These results provide the first evidence for the evolution and consolidation of interhemispheric plasticity in intact humans.
For more than 100 yr, most studies on cross education have reported an increase in motor function in the contralateral homologous muscle after unilateral practice involving strong voluntary contractions with no skill (3,9,29) or involving complex motor skills (24). The present results extend these findings by showing that serial voluntary isometric contractions at submaximal intensity involving minimal skills can also produce functionally and statistically significant interlimb transfer of maximal voluntary force. Two-thirds of cross education appeared by the 10th session, with its gain slowing to about 7% between sessions 10 and 20, suggesting that cross adaptations peaked by session 20, after 1000 contractions.
The value of 21.8% cross education observed here agrees with the magnitude (27%, adjusted for the changes in the control group) reported previously for a similarly unfamiliar movement, ulnar deviation (9) and, as expected, exceed the 8% value reported in meta-analyses for familiar movements (3,29). During effortful unilateral muscle contractions, the contralateral homologous muscle can exhibit as high as ∼20% mirror EMG activity and involuntary force (40). Keeping the mirror activity low under these conditions was important because exercise with loads as light as 10% of maximum can increase maximal force (22). Therefore, higher levels of mirror activity could have confounded the observed cross education. However, the average mirror EMG activity in the present study was only 1.3% of maximum EMG activity and did not correlate cross education. Therefore, it is unlikely that mirror EMG activity contributed to the 21.8% cross education. The cross education observed in the present study had a high level of spatial specificity because it was confined to the homologous left FDI without changes in the left ADM although both muscles are innervated by the deep branch of the ulnar nerve. The evolution of cross education was linked to a strong dependence on persistent and large gains (49.9%) in motor output of the exercised, right, FDI because the strength gains in each FDI became more strongly correlated as the study progressed from session 5 (r = 0.01), 10 (r = 0.17), 15 (r = 0.55), and to session 20 (r = 0.81). Because there were no changes in the twitch-evoked forces produced by electrical stimulation in the untrained, left, FDI, cross education was the result of neural mechanisms and not muscle hypertrophy, as suggested previously (3,9,29). The progressively stronger association between the increases in EMG activity recorded during test MVC of the untrained, left FDI and the increases in cross education suggests that an increase in gross neural drive to the untrained, left, FDI contributed to cross education, strengthening the argument for a neural, possibly supraspinal mechanism mediating cross education (23).
Corticospinal excitability of IM1 contributes to interlimb force transfer.
TMS studies provide evidence that activation of iM1 and the excitability of the corticospinal path targeting the resting hand increase during contralateral unilateral voluntary isometric contractions (16,26,27,30). Especially relevant to the present study are the observations from cross-sectional studies that in hand and wrist muscles iM1 excitability increases with increasing contraction intensity in a task-specific manner (16,27,30) and that twitch interpolation after chronic unilateral training revealed increases in neural drive from iM1 to the untrained muscles (23). Together, these data point to the existence of a model that forceful unilateral motor practice would repeatedly and bilaterally increase excitability of both M1 that in turn would consolidate the increase in excitability in iM1 and mediate interlimb force transfer.
The present data provide evidence for this model on multiple levels. First, chronic effortful motor practice made iM1 more excitable during contralateral muscle contraction. In previous studies and at baseline in the present study, conditioning of weak TMS with a weak muscle contraction did not facilitate MEP in iM1 (27). Although motor practice was done at 80% MVC, adaptations also occurred at weak contractions: MEP size still increased 10.3% (P = 0.001) when weak TMS at 120% rMT was combined with a weak (20%) muscle contraction (Figs. 4A, C), suggesting that iM1 became more sensitive to weak input. Second, the capacity of iM1 to respond to TMS also increased because when strong TMS at 160% rMT was conditioned with strong, 80% MVC, MEP size increased 63.9% by session 20 (Figs. 4B, D). In previous acute experiments, iM1 facilitation was normally saturated when TMS at 160% rMT was conditioned with strong voluntary contraction (27). but here we found a profound increase in the ipsilateral corticospinal path's ability to respond to strong TMS conditioned with a strong muscle contraction. Finally, the 6% increase in MEP size at rest also supports the idea of iM1 becoming more sensitive to TMS and undergoing plastic changes with chronic unilateral motor practice, confirming a previous report that found bilateral increase in corticospinal excitability after unimanual ballistic motor practice (4). Although this increase in MEP size at rest did not correlate with cross education and probably plays a minimal role in between-limb force transfer (see also Perez et al. ), the significant associations that evolved by session 20 between changes in MEP size during contraction and cross education suggest, for the first time, a functional role for these adaptations in the corticospinal paths projecting to the nonexercised, left FDI. These results are compatible with the model that unilateral motor practice can upregulate the excitability of iM1, especially during muscle contraction, and improve motor behavior.
Several studies support the view that GABAa-mediated SICI contributes to M1 plasticity (39). Indeed, MVC or skill practice by one hand cause reductions in SICI in the other hand (2,27,32), and this diminishment in SICI strengthens with increasing unimanual force generation (30). Therefore, it was reasonable to expect that, with the evolution and consolidation of cross education, SICI would decrease, and this would increase corticospinal excitability in iM1. Contrary to this expectation, we observed no changes in SICI and ICF. In previous studies, motor skill training acutely, within a single session, reduced SICI in small hand muscles (6) and a leg muscle (31). Although SICI is lower in musicians versus nonmusicians (34), it is unclear if such adaptations are not the result of a selection bias because long-term sensorimotor training produced no changes in SICI (1). One possibility is that we instructed subjects to suppress mirror activity during exercise, and this volitional inhibition of mirror activity in the left FDI during contractions of the right FDI negated any training-induced decrease of SICI and increase in ICF. This is because volitional inhibition deepens SICI, lessens ICF, and suppresses corticospinal excitability (37). Thus, had subjects not suppressed mirror activity in the left FDI, SICI could have diminished, ICF could have increased, and cortical excitability measured with the RC could have also increased more than the observed 6%.
The locus and the cellular mechanisms that contribute to persistent changes in corticospinal excitability in iM1 after unilateral motor practice are not known. The mechanisms could involve changes in membrane properties of corticospinal neurons and in the efficiency of excitatory synaptic inputs onto corticospinal neurons produced by the forceful muscle contractions. Whether the site of these adaptations is in iM1 circuits is unclear because some human and animal experiments suggest that the nature of the motor task has to be complex to produce plasticity in the task and nontask hemisphere (18). However, several studies, using magnetic and electrical stimulation (see references in Hortobágyi et al. ) and imaging (9) show increases in excitability of supraspinal structures after chronic high-intensity motor training using a simple motor task. rTMS studies strongly suggest that motor practice-induced increases in corticospinal excitability were caused by cortical rather than spinal adaptations (28). Although based on imaging and TMS studies, the possibility cannot be excluded that remote cortical areas subserving M1 such as supplementary motor areas (32), caudal cingulate, cerebellum (36), and parietal cortices (20) contributed to the increase in iM1 excitability; the most likely scenario involves the modulation of transcallosal paths. Finally, it is still possible that spinal mechanisms played a role in cross education because there were no changes in SICI and ICF, and spinal mechanisms can also increase MEP in the RC caused by a more synchronized corticospinal conduction, producing a more efficient summation of descending volleys at the spinal motoneurons.
To the best of our knowledge, the present data provide the first longitudinal evidence for unimanual motor activity producing a regulatory effect on IHI in intact humans. Repeated effortful contractions of the right FDI increased maximal voluntary force in the left, untrained, FDI by 21.8% and decreased IHI from the trained to the untrained hemisphere by 30.9%, and these changes in IHI and the transferred force became progressively more strongly and significantly correlated across 20 sessions. The data suggest that interhemispheric plasticity most likely contributed to cross education.
Short-interval IHI is mediated by glutaminergic excitatory transcallosal motor fibers that arise from the M1 that receives the conditioning TMS pulse and project onto local GABAergic inhibitory interneurons located in the opposite M1 that receives the test TMS pulse (11). Under nonpathological conditions, IHI from the active to the "nonactive" hemisphere increases with increasing intensity of unimanual muscle contraction, presumably to suppress unwanted mirror (co)activation of the voluntarily "nonactive M1" and to minimize mirror EMG activity in the nontask homologous muscle (17). In the present study, subjects were successful in volitionally suppressing mirror EMG activity in the left FDI as they strongly contracted the right FDI. Although volitional suppression of voluntary command decreases M1 excitability and deepens SICI in the task (37) and nontask hemisphere (25), the specific effects of such suppression are unknown on IHI. In the present study, the net effect of repeated forceful muscle contractions on one side and suppression of mirror activity on the other side was that the muscle contractions had a conditioning effect on IHI, which, during 20 sessions, became progressively and significantly more diminished from the active to the nonactive M1. Several lines of evidence support this scenario. First, although in the original experiments IHI deepened with a slight unimanual muscle contraction (11) when the size of the conditioning and test MEP are properly adjusted across force levels, IHI actually decreases (30). Second, a suppression of mirror EMG activity has little influence on the duration of ipsilateral silent period, a measure of IHI (13). Third, although the unimanual serial reaction time task that transferred implicit knowledge and the unimanual pinching task that produced transfer of speed and accuracy are eminently different from each other and still are more complex than the simple task of serial isometric contractions used in the present study, each produced reductions in IHI from the practiced to the nonpracticed M1 (2,32). It thus seems that IHI is amenable to diminishment independent of the nature of the task. Fourth, although in cross-sectional studies IHI influenced SICI during unimanual wrist flexions (30), we found no changes in SICI and the changes in IHI and SICI did not correlate at any time point, further supporting the contention that IHI acted as a key moderator of the increase in voluntary force of the untrained, left, FDI. Together, the behavioral outcome in the present study was the transfer of voluntary force from the trained to the nonpracticed finger so that the magnitude of transfer (i.e., cross education) became strongly associated with the magnitude of attenuation in IHI. These changes were topographically confined to the homologous, left, FDI because there were no changes in IHI and force transfer in the nonhomologous, control ADM muscle. Therefore, it seems that IHI contributed to interlimb transfer of the ability to produce maximal voluntary force by the untrained homologous muscle after a unimanual exercise intervention involving a simple motor skill.
When healthy adults execute a unilateral motor task, there can be as much as 20% of maximum mirror EMG activity (13,40) and large increases in excitability of iM1 (16,27). There has been a lingering suspicion in previous longitudinal exercise studies that the rarely quantified but undoubtedly present mirror EMG activity in the "inactive" contralateral homologous muscle (14) and the increased excitability of the nontask M1 contributed if not caused the observed cross-education effects (3,5,9,29). However, we controlled for mirror EMG activity by repeatedly instructing the subjects in a standardized manner to relax the homologous muscle pair contralateral to unimanual muscle contractions (16). Therefore, mirror EMG activity (∼1.3%) very likely did not cause cross education. Yet, the present and previous results (2,32) shed paradoxical light on IHI concerning its role in mirror EMG activity and mediating cross education. IHI's function is to eliminate unwanted mirror EMG activity and mirror movements in the resting limb when muscles of the other limb execute a unilateral motor task (17). The prediction is that unimanual motor practice, by repeatedly activating IHI networks, would tend to inhibit the homologous and even nonhomologous muscles (25) in the opposite limb and strengthen interlimb independence. In contrast, the concept of cross education predicts that unilateral exercise would increase performance in the untrained limb and making, perhaps, the two limbs not more but less independent. Indeed, we observed here and others also reported previously (2,32) that unimanual motor activity decreased IHI and reductions in IHI correlated with transfer of motor function. Such a diminishment would tend to actually increase instead of decrease mirror EMG activity and, more broadly, interference between limbs. Because cortical excitability of iM1 at rest increased little (6%) and mirror EMG activity was virtually absent, the diminishment of IHI seems to be a key contributor to the observed cross education. Although cross-sectional studies link mirror activity to IHI (5,17), under the current chronic experimental conditions practice of a simple unilateral voluntary force production task reduced IHI with virtually no covariation or progressive increases in mirror activity (∼1.3%), suggesting that the two variables are independent. Again, the current experimental approach does not allow us to exclude the possibility of changes in spinal excitability contributing to cross education and controlling mirror activity.
The present results have clinical relevance to neurological and orthopedic conditions. Unilateral exercise of the free limb reduced voluntary strength loss produced by experimental immobilization (10). It is possible that unilateral interventions can modulate excitatory balance between the two hemispheres and yield diagnostic or therapeutic benefits in certain clinical conditions such as multiple sclerosis where IHI can be affected because of demyelination of callosal fibers, cortical myoclonus where deficient IHI can facilitate the spread of myoclonic activity between hemispheres, and subacute stroke where reduced IHI can lead to hyperexcitability of M1 in the unaffected hemisphere.
Supported in part by National Institutes of Health NS049783, National Institutes of Health, National Institute of Neurological Disorders and Stroke intramural program, and an East Carolina University Research Development Award.
The authors thank Dr. Paul DeVita for his helpful comments and Mr. Patrick Rider and Mr. Jonathan Gomez for preparing the figures.
The results of this study do not constitute endorsement by the American College of Sports Medicine.