Increased Inhibition in Non-Primary Motor Areas of String-Instrument Players: A Preliminary Study with Paired-Pulse Transcranial Magnetic Stimulation

INTRODUCTION

The non-primary motor areas (NPMAs) anterior to primary motor cortex (M1) have representations of the upper limb muscles in both non-human primates and humans [1–6], but the role of these representationsoutside M1 is unclear. It has been claimed that an assessment of the integrity of corticospinal tract fibers originating from dorsolateral premotor cortex (PMd) may have some predictive value in the assessment of stroke recovery [7]. Even though the degree of motor recovery in stroke patients is most strongly correlated to the integrity of corticospinal tract fibers originating from M1, a higher integrity of corticospinal tract fibers from PMd also predicts a better recovery (7). Furthermore, it has been reported that motoneurons in NPMAs are recruited when individuals are learning skills requiring fine motor control [8].

Navigated TMS is a technique that enables exploration of the motor representations in M1 and NPMA separately. Previously, we have stimulated M1 and NPMAs in healthy subjects and mapped cortical muscle representations [3, 9]. In addition to M1, muscle responses could be evoked by stimulations of the PMd and border region between the PMd and the supplementary motor area (SMA) without any simultaneous activation of M1 [3, 9].

Intracortical inhibition and facilitation in the motor representations can be studied by paired-pulse TMS with different inter-stimulus intervals (ISIs) [10, 11]. In paired-pulse TMS, a subthreshold conditioning stimulus (CS) precedes the suprathreshold test stimulus (TS). Inter-stimulus intervals shorter than 6 ms produce short-interval intracortical inhibition (SICI), which can be seen as a decrease in MEP amplitudes when compared to single-pulse stimulation alone. Intracortical facilitation (ICF) is most pronounced with ISI values longer than 7 ms and can be detected as an increase in the MEP amplitude [10–12]. In our previous study, SICI and ICF were observed in motor representations in M1 and in NPMA. We noted that SICI at 2 ms ISI measured in the NPMA representation of opponens pollicis (OP) muscle was slightly weaker than inM1 [9].

The aim of the current study was to compare SICI and ICF in the intrinsic hand muscle representations in M1 and in NPMA (corresponding mainly to PMd) in a cohort of 5 string-instrument players with a control group of participants not actively playing any musical instrument. In our previous study, the overall size of the cortical representation of the actively used hand muscle was smaller in string-instrument players than in controls and also in comparison to the representation area of less actively used hand in the contralateral hemisphere [13]. We also hoped to clarify the role of NPMA motor representations in the control of skilled finger movements in subjects with long traininghistories.

MATERIAL AND METHODS

RESULTS

DISCUSSION

We found that local inhibition (SICI) was increased in the non-primary motor area (NPMA) cortical representation of left ADM muscle in the string-instrument players which is a muscle actively used by right-handed string-instrument players. In M1, the effect was opposite, SICI was reduced in the string-instrument players. These results may indicate that cortical control of an actively used muscle is more effectively focused on the primary motor representation in M1 in fine motor skill specialists. The weaker inhibition in M1 in actively used hand muscle representation in string-instrument players may be interpreted as an increased readiness to recruit a local motoneuron population.

TMS studies have shown that during the initial phase of learning a new fine motor skill, SICI decreases, MEP amplitudes increase and motor representations enlarge, i.e. phenomena which represent increased activation of synapses and deactivation of silent synapses [38–41]. Continuing practice leads to the formation of new synapses [42, 43] and axonal sprouting [44]. In musicians, the synaptic plasticity is enhanced and cortical excitability is regulated more stringently than in non-musicians [45]. This more effective regulation of cortical excitability is seen as a steeper input-output curve of MEP amplitudes and a steeper SICI curve [45]. These findings can be interpreted as a higher capability to recruit fast volitional movements and a greater capability to prevent the unwanted spread of activation, both of which lead to more focused cortical activation.

In our previous report, more focused functional ADM muscle representations were also demonstrated in string-instrument players than in controls [13] even though earlier studies have reported also contrasting results [27, 46, 47]. Structural evaluations have demonstrated both decreased and increased grey matter density in highly active sensorimotor cortices in musicians [48, 49]. The conflicting results may reflect the different roles of the studied muscles in the trained skill as well as differences in the degree of co-activation with adjacent muscles. Spatially wider motor representations and the greater overlap of proximal and distal muscles are both related to skills demanding synchronous activation of different muscles [50–52] and more focused cortical activation may be associated with skills demanding accurate single-muscle movements. A more focused cortical activation in musicians has also been reported in fMRI studies [53, 54]. Even though increased activity in SMA and especially in caudal PMd, has been related to the learning phase of a fine motor skill, no evidence has been found for any higher activity of these areas in well-learned skill performance [53]. This may be because there is less need for preparation of movement sequences, error detection and correction; it is especially these types of activities which are processed in PMA and SMA [31, 55, 56]. In fMRI studies, the increased activity of any neurons is seen as an increase in the BOLD signal and it cannot separate activity of motoneurons from other cortical neurons. The SMA proper and posterior parts of PMA are connected strongly to M1 and the majority of neurons in these areas are responsible for the preparation and correction of movement sequences via M1 motoneurons [5]. Thus, more focused cortical activation in fMRI studies and the shift of activation towards M1 in learned skill performance cannot be unequivocally interpreted as more focused muscle representations i.e. activation of more focused motoneuron population but may be evidence of a reduction of PMA and SMA activation because of a diminished need to concentrate on preparation and timing of the movements. In the present study, the stronger SICI in the representation of the actively used hand muscle in NPMA of string-instrument players suggests that also the direct cortical control of voluntary movements has been shifted towards M1.

The sensory input (vibration) to the actively used hand muscle has been demonstrated to decrease SICI in multiple intrinsic hand muscles in musicians whereas in non-musicians this reduction in SICI was seen only in the vibrated muscle [57]. The representations in NPMAs have projections from somatosensory associative cortices (BA 5) but do not possess direct connections from S1 [58, 59]. The strong connections between S1 and M1 and the importance of sensorimotor organization in professional musicians could potentially explain the importance of motor control from M1 representations.

Based on stroke studies, the representation areas in NPMAs anterior to M1 may be relevant in the recovery from an M1 lesion, even though the best motor recovery is related to increased activation of M1. Previous studies have revealed a shift in the motor representations anterior to M1 in the affected hemisphere of stroke patients [60]. A positive correlation has been observed in stroke patients between the integrity of corticospinal tract fibers originating from M1 and PMd and grip strength pointing to an important role of PMd representations in stroke recovery [7]. In particular, those patients with poor recovery displayed increased activation in NPMAs when performing paretic hand movements [61–63]. These studies indicate that in severe M1 damage and intense motor impairment, the recruitment of motoneurons from NPMAs might be beneficial even if recovery will not be as complete as in patients in whom activation is being restored towards M1. Therefore, the role of NPMA representations is more likely to be compensatory for M1 in making rough limb movements while M1 representations are necessary in performing fine motor movements.

There were no significant differences in SICI and ICF between OP and ADM muscles in string-instrument players although ADM is more actively used. We could have expected a similar difference in inhibition between the muscles than between the inhibition in ADM representations in string-instrument players and controls. Although RMT should not be affected by long-term practice of fine motor skills [27], OP and ADM muscles may have a slightly different threshold which could have some effect on results. We determined RMT only for OP muscle and used CS and TS intensities proportional to OP RMT intensity.

Because of the small sample size, the results should be viewed as preliminary and interpretations discussed with caution. The results will need to be confirmed in future studies. The cortical excitability should also be studied between the hemispheres to confirm that the active use of left hand muscles in right-handed string-instrument players is related to altered cortical excitability. One confounding factor is the heterogeneity of subjects in the control group. One of the subjects was left-handed, all other subjects right-handed. Previous studies have shown heterogeneous results when cortical excitability has been compared between left- and right-handed subjects. Differences in SICI and ICF were found in the study of Civardi et al. showing stronger SICI and weaker ICF in both hemispheres in right- handed subjects when compared to their left-handed counterparts [64]. However, Bäumer et al. did not detect any differences in the values of SICI and ICF between left- and right-handed subjects [65]. Furthermore, no differences were observed between left- and right-handed subjects in SICI between dominant- and non-dominant hemispheres with classical paired-pulse protocol in the study of Ilic et al. [66]. Since there is no clear evidence about the dependency of handedness in SICI and ICF and the non-dominant hemisphere was stimulated in all subjects, it is unlikely that the single left-handed subject in a group of six controls would skew the results substantially, however the effect of handedness may not be totally ruled out. In addition, four of the subjects in the control group had a long practice history as figure skaters. Although a figure skater does not perform any complex motor sequences with the small hand muscles, the learning-related plasticity in other motor representations may exert some effect on SICI and ICF in the primary and secondary hand motor representations.

One further limitation in interpreting the results is the separation of muscle representations in NPMA and in M1. The electric field value remained below RMT in M1 when the optimal target in NPMA was stimulated with local RMT intensity. When the optimal target in NPMA was stimulated at 120% of local RMT intensity, the electric field value exceeded the MT value in M1 in some subjects. Since the optimal direction of the current differed between M1 and NPMA (from the lateral to the medial direction in NPMA and from the posterior to the anterior direction in M1), it seems improbable that there was any considerable activation of motoneurons in M1 when NPMA was stimulated. However, we cannot totally exclude the possibility of activation of motoneurons in M1 or in cortical area between the optimal targets in M1 and NPMA, especially when NPMA was stimulated with 120% RMT intensity. The only way to exclude the activation of M1 would have been to stimulate the optimal target in M1 with an identical electric field value as used in the induction of M1 when NPMA was stimulated with 120% RMT intensity and with the same direction of current that was used in NPMA. Therefore, it is possible that in our results, the MEPs, SICI and ICF have been affected by mild activation of M1 motoneurons when stimulating NPMA with 120% RMT intensity.

It is not possible to define the exact functional area in the NPMAs producing muscle responses because there are not any anatomical landmarks which would separate PMd from SMA. The PMd and SMA are functionally distinct motor areas anterior to M1 [5, 67]. Both areas are located in BA 6. Pre-SMA (the more frontal part of SMA) is located in BA 8. There are no clear anatomical landmarks with which these functionally distinct areas, PMd and SMA, can be separated. Based on the probabilistic maps of Geyer et al., the cytoarchitectural border between BA 6 and BA 8 is frontally in the level of anterior commissure [30]. SMA is located more medially (near to the interhemispheric fissure) than PMd. It is most probable that the MEPs evoked by the stimulations of sulcus between middle and superior frontal gyrus were produced from PMd. Since the current was directed towards thesuperior frontal gyrus, we cannot rule out the possibility of motoneuron activation in SMA. If TMS pulses activated motor representations in PMd and SMA, then the results may be affected by differences in the cortical excitability balances between these areas.

It appears unlikely that M1 motoneurons had been activated polysynaptically via connections from NPMA to M1 because the latencies were similar after stimulations of M1 and NPMA to ADM muscles in both groups and to the OP muscle in the string-instrument players. Even though the stimulus locations in NPMAs most probably are independent muscle representations, the possibility for separate clusters of primary representations cannot be excluded. This would mean that the primary muscle representations would extend anterior to the precentral gyrus. In monkeys, novel neural connections and functionally active synapses were observed from higher visual centers to the intraparietal cortex after tool-use training which demanded visual and somatosensory information processing [44]. Thus, one could speculate that the motoneurons in NPMAs are part of the primary representation and can be activated by TMS, even if recruited to use only after damage to the M1 representations. According to both the cytoarchitectural and functional classifications of the cortex, the primary motor cortex with giant pyramidal cells in cortical layer V is located in the posterior part of the precentral gyrus and areas anterior to this region with their own distinctive cytoarchitectural structures and characteristic functions are considered to be the premotor and supplementary motor areas (BA 6 and BA 8). Therefore it is perhaps more accurate to presume that the motoneurons/muscle representations originating in PMA and SMA are distinct from the primary representation in M1. The study of Schulz et al. also demonstrated that corticospinal tract fibers originating from different parts of NPMA exerted different effects on stroke recovery. The integrity of fibers originating from SMA did not exhibit any positive correlation with stroke recovery similar to those fibers originating from M1 and PMd, indicating that they have differing activities or functions [7].

The latencies were even faster to the OP muscle and OP MEP amplitudes were higher after stimulations of NPMA than to M1 in the control group. Since the MT was higher in NPMA, the higher amplitudes and faster MEP latencies most probably reflect the activation of a sparse but widely spread motoneuron population in NPMA. This finding is in line with our previous study [9].

Since the threshold to activate cortical interneurons may differ from the RMT, it is possible that the higher absolute stimulation intensities in NPMAs may have activated local inhibitory interneurons more effectively, thus producing stronger SICI in NPMA [68, 69]. This may explain, at least in part, the differences in SICI between M1 and NPMA within the groups at low CS intensities. Nevertheless this would not account for the differences in SICI between the groups.

In conclusion, the current study provides preliminary evidence for the need for more effective inhibition of motor representations of actively used hand muscles in NPMAs in string-instrument players. This may be beneficial in the prevention of any unnecessary spread of cortical activation when finely-controlled single muscle movements must be performed rapidly. Weaker SICI in string-instrument players in M1 in the actively used hand muscle representation may enable fast recruitment of M1 motoneurons when needed.

CONFLICTS OF INTEREST

Karhu J. is affiliated part-time as the Medical Advisor of Nexstim Plc. Vaalto S. and Julkunen P. have received unrelated consulting compensation from Nexstim Plc, the manufacturer of the navigated TMS devices. Könönen M., Määttä S. and Säisänen L. do not have financial interests or potential conflicts of interest.