|Home | About | Journals | Submit | Contact Us | Français|
The aim of this study was to further characterize surround inhibition (SI) in the primary motor cortex (M1) by comparing its magnitude and time course during a simple reaction time task (SRT) and a choice reaction time task (CRT).
In both the SRT and the CRT, subjects performed the same right index finger flexion in response to an acoustic signal. For CRT, the alternative choice was a similar movement using the left index finger, as distinguished by a different tone. In both tasks, single pulse transcranial magnetic stimulation (TMS) was applied at rest, 75ms (T1) and 25ms before EMG onset (T2), and during the first peak of EMG (T3) in the right first dorsal interosseous muscle (FDI). Motor evoked potentials (MEPs) were recorded from both FDIs, which act as synergists in the task, and the right surrounding, relaxed abductor pollicis brevis muscle (APB).
For right hand movement, SI started earlier and was more pronounced for CRT compared to SRT. For left hand movement in the CRT, SI was similar to that of right hand movement.
We conclude that SI occurs earlier and stronger with increasing task difficulty.
The timing as well as the bilateral effect of the inhibition suggests that motor areas involved in motor planning, proximate to the motor cortex, contribute to the genesis of surround inhibition.
Surround inhibition (SI) is a neural mechanism, first described in the visual system, to focus neuronal activity (Angelucci et al., 2002). In the retina, cells are excited by light that falls in the center of their receptive field, while light in the peripheral areas of the receptive field has an inhibitory effect on the same cell. SI has been described in the primary motor cortex (M1) (Hallett, 2004; Sohn and Hallett, 2004b) and deficient SI seems to be a component of the pathophysiology of focal hand dystonia (FHD) (Sohn and Hallett, 2004a; Stinear and Byblow, 2004; Beck et al., 2008).
Using single pulse transcranial magnetic stimulation (TMS), motor evoked potentials (MEPs) in the active muscle are facilitated before and during movement (Rossini et al., 1988; Tomberg and Caramia, 1991), whereas MEPs in antagonist muscles (Tomberg, 1995; Hoshiyama et al., 1996) and the contralateral homologous muscle (Leocani et al., 2000) are inhibited in a time interval from 60-100ms before EMG onset. While the time course of MEP modulation in the active muscle appears to be independent from the type of task that is performed (simple reaction time task, choice reaction time task or go/no-go-task) (Leocani et al., 2000), the amount of change in M1 excitability increases with increasing expectancy of a response signal in the muscle to be activated (van Elswijk et al., 2007). Synergistic muscles show the same type of modulation as prime movers (Sohn and Hallett, 2004b).
At this point, the timing as well as the exact mechanisms underlying the genesis of SI is unclear. There is evidence that the inhibition of unwanted movements can be modulated by volition (Liepert et al., 1998) and differs between tasks (Sohn et al., 2002). One possible mechanism is the so-called short intracortical inhibition (SICI), which reflects the inhibitory effect from the local GABAA-ergic inhibitory interneurons in the primary motor cortex (M1) and can be assessed using paired-pulse TMS (Kujirai et al., 1993; Di Lazzaro et al., 1998). Findings on the contribution of SICI to SI are controversial; while one study reported an up-regulation of SICI synchronous with SI, others did not (Beck et al. 2008; Sohn et al. 2004). This difference in result may be due to differences in tasks, target muscles as well as stimulation parameters. Alternatively, other motor areas, such as the contralateral primary motor cortex (M1c), parietal cortex, supplementary motor area (SMA), ventral (PMv) or dorsal premotor cortex (PMd), may contribute to the formation of SI. PMd in particular is known to be relevant for motor planning in tasks like CRT (Schluter et al., 1998), and this could be responsible for a difference in SI between SRT and CRT.
In this study, we assessed SI in the adjacent, non-synergistic abductor pollicis brevis muscle (APB) during an index finger flexion, which has been shown to inhibit MEPs in APB in a previous study (Beck et al., 2008). SI during the time course of a simple acoustic reaction time task (SRT) and a forced choice reaction time task (CRT) were compared. For the latter, subjects performed the same movement they executed during SRT using either the left or right hand. Two different acoustic signals were used to cue either right or left hand movement. The goal of this study was to compare the time course and amount of SI between SRT and CRT. In addition, SI was compared between right and left hand movement during CRT. We expected that SI would be more pronounced for CRT.
Eleven healthy volunteers (age 39-69 years, mean 54.5 ± 4.5 years; 6 males) participated in the study. All subjects were right-handed according to the Edinburgh handedness inventory (Oldfield, 1971). Participants had never been exposed to neuroleptic drugs and had no history of other neuropsychiatric disorders, neurosurgery, or metal or electronic implants. All participants gave their informed consent prior to the experiments, which were approved by the Institutional Review Board (IRB) of the National Institute of Neurological Disorders and Stroke (NINDS).
Subjects were seated in a chair with both arms resting on board in front of them to which the force transducers were attached. In some subjects, the wrist was supported by a towel to help the subject keep the hand muscles as relaxed as possible. Disposable surface silver-silver chloride EMG electrodes were placed on the right APB and on both FDI in a belly-tendon montage. Impedance was reduced below 5 kΩ. The EMG signal was amplified using a conventional EMG machine (Nicolet Viking, Skovlunde, Denmark) and bandpass filtered (20-2000 Hz). The signal was digitized at a frequency of 4 kHz and fed into a computer for off-line analysis. The individual MEP amplitudes were measured in four phases (rest, 75ms and 25ms before and during the first peak of the EMG in FDI; see motor task). Background EMG was calculated in all three muscles by assessing the root mean square over 20ms prior to MEP onset in the same four phases.
The task for SRT has previously been used successfully to induce SI in APB (Beck et al., 2008; see Fig. 1). In brief, subjects performed an index finger flexion, which they practiced at the beginning of the experiment to attain a consistent motor performance. With their right hand lying flat on a table beside them, subjects pushed down on a small force transducer (Strain Measurement Devices, Inc, Meriden, CT, model S215 load cell) with the tip of their index finger in response to an acoustic signal. This led to a flexion in the metacarpo-phalangeal joint (MP joint) of the index finger. FDI participates as synergist rather than as prime mover in this motion, but it has been shown that the modulation of cortical excitability is similar in prime movers and synergists (Sohn and Hallett, 2004a). Subjects produced 10% of their maximum force (Fmax) of isolated index finger flexion as fast as possible after the onset of the tone and maintained the contraction for about 500ms. The force level was individually adjusted and displayed as a line on an oscilloscope in front of them. The output of the force transducer was also displayed on the oscilloscope as feedback. The acoustic signal was present for 200ms.
For CRT, subjects performed the same movement as described above either using their right index finger (same tone as in the SRT task; 1kHz) or left index finger (1.5kHz). The likeliness for each signal to be presented was 50%.
In SRT and CRT, four different phases of the movement were assessed: rest (100ms before the onset of any tone), 75ms before the onset of the EMG in FDI (T1), 25ms before the onset of the EMG in FDI (T2) and the first peak of EMG in FDI (T3). Similar timings were used in both tasks, as it has been shown that the time course of modulation of cortical excitability in the active, synergistic muscle is independent from task difficulty and reaction time (Leocani et al. 2000), while the amount of modulation reflects task difficulty (Leocani et al., 2000).
In all subjects, SRT and CRT tests were performed in two different sessions in a randomized order. All participants performed SRT and CRT reliably with an error rate < 5%. The incorrect trials were excluded from the analysis.
For TMS, a high-power Magstim 200 machine (Magstim Co., Whitland, Dyfed, UK) was connected to a custom made figure-of-8 coil with an inner loop diameter of 35mm. At the beginning of each experiment, the “motor hot spot” for eliciting MEPs in right APB was determined over left M1. This position was marked on the scalp to ensure proper coil placement throughout the experiment. Coil orientation was tangential to the scalp with the handle pointing backwards and laterally at a 45-degree angle away from the midline. Resting motor threshold (MT) was determined for APB to the nearest 1% of maximal stimulator output. MT was defined as the minimal stimulus intensity required to evoke MEPs of at least 50μV in 5 out of 10 consecutive trials. The stimulus intensity applied during both tasks was 140%MT. MEP size was determined by averaging peak-to-peak amplitudes. Trials with a background EMG of more than 0.02mV (assessed as root mean square) over 20ms before the onset of the MEP were rejected.
The amount of SI in APB is expressed as ratio between MEP size during the phasic phase in percent of the MEP size at rest:
SI = (MEPphasic/MEPrest)*100 [%]
Two different comparisons were conducted: The comparison between SRT and CRT during right hand movement and the comparison between right and left side movement for CRT.
Since the data were not Gaussian for all conditions, a non-parametric Wilcoxon test was used to compare the unconditioned MEP size to the conditioned MEP size for the 3 phases for the 2 tasks. This was done to identify significant SI, which was expressed as ratio between conditioned and unconditioned MEP, in percent (SI = (MEPcond/MEPtest)*100[%]). Then Conover's distribution-free method, a non-parametric ANOVA based on ranks was used to compare SI in APB (Conover and Iman, 1982). Two within-subject factors were compared: PHASE (3 levels: T1, T2 and T3) and TASK (2 levels: SRT and CRT). If effects were found to be significant at the 0.05 level, a contrast analysis was performed.
Since MEP sizes were not Gaussian distributed for all conditions, Conover's distribution-free method, a non-parametric ANOVA based on ranks was used to compare SI in APB (Conover and Iman, 1982). Two within-subject factors were compared: PHASE (3 levels: T1, T2 and T3) and SIDE (2 levels: right and left). If effects were found to be significant at the 0.05 level, we performed a contrast analysis. To test for significant SI, SI was compared applying Friedman test followed by a non-parametric Wilcoxon test using PHASE (3 levels: T1, T2 and T3) and TASK (2 levels: SRT and CRT) as within-subject factors.
Background EMG in APB and FDI were not Gaussian distributed for all conditions. Therefore, we used again Conover's distribution-free method (Conover and Iman, 1982). Two within-subject factors were compared: PHASE (4 levels: rest, T1, T2 and T3) and TASK (2 levels: SRT and CRT) or SIDE (2 levels: right and left), respectively.
MT and Fmax were Gaussian distributed, we therefore used paired t-tests to compare these parameters between tasks (2 levels: SRT and CRT). All results are presented as means and standard error of means. P-values less than 0.05 are considered significant. For analysis, SPSS 11.5.0 was used.
Mean reaction times were 169 ± 4ms (mean and standard error; ranging from 150ms to 200ms) for SRT, 327 ± 15ms (ranging from 250ms to 425ms) for CRT for right side movement and 329 ± 17ms for CRT and left side movement. For CRT, reaction times were not different between right or left side movement (F = 0.05, p = 0.834). The average performance error rate was < 5% in all subjects. Fmax was not different between tasks (p = 0.54, SRT: 4.2 ± 0.4N, CRT: 3.9 ± 0.4N), neither was MT (SRT: 48 ± 2%; CRT: 50 ± 2%; p = 0.1).
For SI in APB, there was a significant main effect for PHASE (F = 8.2, p = 0.003), TASK (F = 5.6, p = 0.028) and for the PHASE-by-TASK interaction (F = 5.3, p = 0.015; see Fig. 2). Contrasts revealed differences between T1 and T2 (F = 15, p = 0.001) and T1 and T3 (F = 4.8, p = 0.041), but not between T2 and T3 (F = 0.1, p = 0.98). The Wilcoxon test revealed first that SRT had less SI at T1 compared to CRT (p = 0.01) and a trend for less inhibition at T2 for SRT (p = 0.08), while there was no difference between tasks for T3 (p = 0.62). Second, when testing against the rest condition, there was significant SI in APB for CRT in all three phases tested (T1: 76.7 ± 5.6%, p = 0.021; T2: 74.9 ± 5.2%, p = 0.008; T3: 75.8 ± 6.6%, p = 0.004). For SRT, there was significant inhibition for T2 (87.4 ± 2.1%, p = 0.004) and T3 (84 ± 5.2%, p = 0.016), but not for T1 (100.7 ± 3.8%, p = 0.66; see Fig.2). MEP size at rest was not different between SRT (2.8 ± 0.2mV) and CRT (2.9 ± 0.3mV, p = 0.09). The analysis of APB EMG 20ms before the MEP did not show any significant effects (all p > 0.1; see Table 1).
In FDI, there was a significant main effect for PHASE (F = 20.2, p < 0.001), but not for TASK (F = 0.16, p = 0.69, see Fig. 3). The PHASE-by-TASK interaction was also not significant (F = 0.24, p = 0.79) indicating that the modulation was similar in both tasks. MEP size at rest was not different between tasks (p = 0.12; SRT 4.3 ± 0.5mV, CRT 3.4 ± 0.5mV). The Wilcoxon test revealed that there was no significant facilitation in FDI for T1 (SRT: 102.8 ± 3%, p = 0.25; CRT: 95.7 ± 8.2%, p = 0.53, see Fig. 3). For T2, there was significant facilitation for SRT (114 ± 4%, p = 0.026), but only a trend for facilitation for CRT (117.5 ± 7.7%, p = 0.09). At T3, facilitation was observed for both tasks (SRT: 125.3 ± 5.4%, p = 0.04; CRT: 125.1 ± 9.1, p = 0.021).
The analysis of FDI EMG 20ms before the MEP did not show any significant effects (all p > 0.1; see Table 1).
The comparison of SI between right and left side movement showed no significant main effect for PHASE (F = 1.6, p = 0.24), SIDE (F = 2, p = 0.17) or for the PHASE-by-SIDE interaction (F = 0.2, p = 0.83; see Fig. 4). Friedman test showed no significant differences between the three phases for SRT (p = 0.63) or CRT (p = 0.23). To test for significant SI, the Wilcoxon test was used and showed that there was significant SI for all three phases during right hand movement (T1: p = 0.029; T2: p = 0.004; T3: p = 0.005, see Fig. 4). For left side movement, only T1 and T2 showed significant SI (T1: 85.1 ± 3.6%, p = 0.006; T2: 82.6 ± 4.3%, p = 0.003, T3: 89.8 ± 5.9, p = 0.23). Comparing the two sides directly, there was no difference for T1 (p = 0.23), T2 (p = 0.18) or T3 (p = 0.08). The analysis of APB EMG 20ms before MEP onset did not show any significant effects (all p > 0.1; see Table 1).
The current results show that in reaction time tasks, SI can be observed up to 75ms before EMG onset. SI in APB started earlier and was more pronounced during CRT compared to SRT. Because the active movement of FDI was similar, as underlined by the synchronous modulation of MEP facilitation in FDI, the current findings suggest that the difference found in SI is due to task difficulty. During CRT, comparable SI was induced by movement of the contralateral (right) hand and the ipsilateral (left) hand, which may reflect a trans-callosal contribution to SI or indicate inhibitory influences from other, secondary motor areas, such as prefrontal cortex, premotor cortex or SMA.
The reaction times in this study were similar to earlier reports for both types of reaction time task without differences between sides (Leocani et al., 2000). In general, investigation of M1 excitability during the time course of a movement of the contralateral hand shows that changes in excitability precede the EMG onset starting from around 120-100ms in healthy subjects (Rossini et al., 1988; Tomberg, 1995; Hoshiyama et al., 1996). This finding is consistent with reports from intracortical recording in animals (Evarts, 1966; Fetz and Finocchio, 1972). In the current study, there was no correlation between individual reaction time and the amount of SI during SRT or CRT. Although it is not possible to calculate the individual onset of SI in the current study, this finding may suggest that during skilled motor behavior the modulation of cortico-spinal excitability precedes movement selection (Leocani et al. 2000). A non-selective inhibition may be followed by a selective excitation of the cortical representation of the muscles that are to be activated (Sohn and Hallett, 2004b), while surrounding muscles are inhibited (Stinear and Byblow, 2004; Beck et al. 2008).
The time course of facilitation of agonist muscles in humans seems to be consistent between different types of reaction time tasks, such as simple and forced choice reaction time tasks and go/no-go tasks (Leocani et al., 2000), whereas the amount of facilitation increases with increasing expectancy (van Elswijk et al., 2007).
Earlier studies on SI used self-initiated movements, during which the TMS pulse was applied as soon as a certain EMG threshold was reached (Sohn and Hallett, 2004b). As a consequence of this type of task, the beginning of SI in the time course before EMG onset could not be assessed. SI was observable right from the beginning of the recordings, which is approximately 20ms after the onset of EMG, because the nerve conduction time needs to be taken into account (Sohn and Hallett, 2004b). In contrast, the reaction time tasks used in the current setup allowed the assessment of earlier phases. While there was no significant SI at T1 (75ms before EMG onset) for SRT, CRT already induced SI at that phase. For the later phases of the movement, both tasks induced similar SI. These findings are in line with earlier reports about SI in the phase before EMG onset (Molloy et al., 2002; Beck et al., 2008) and during the first peak of EMG (Stinear and Byblow, 2004; Beck et al., 2008).
Looking at actual reaction times in this type of task, TMS has been found to disrupt motor performance, when applied 30-50 ms before EMG onset to M1 (Day et al., 1989; Pascual-Leone et al., 1992; Leocani et al., 2000). Single pulse TMS also interferes with motor performance, when given at time intervals of 80-120ms before EMG onset to areas involved in motor planning, such as SMA (Perez and Cohen, 2008) and dorsal premotor cortex (PMd) (Schluter et al., 1998). Therefore, several inhibitory circuits may contribute to SI.
At the first phase tested (T1), SI was more pronounced during CRT compared to SRT. Although subjects were trained in the tasks until they reached consistent motor performance with an error rate below 5%, the stronger inhibition may be due to an increase in attention or a difference in the current cognitive motor state between the two tasks (Conte et al. 2007; Liepert et el. 1998). In contrast to simple motor execution, attention to action leads to greater activation of PMd and the anterior cingulate cortex (Jueptner et al. 1997). There is evidence that attention influences the amount of MEP facilitation induced by high-frequency rTMS in healthy subjects through premotor-motor connections (Conte et al. 2007). For example for the no-go condition in a go/no-go paradigm, it is known that projections to muscles involved in both possible movements show facilitation initially. After the cue, the unwanted movement is actively suppressed (Leocani et al. 2000; Sohn et al. 2003). This inhibition may contribute to the inhibition we observed in the surrounding muscle.
At this point, it is still unclear, how SI is generated in healthy volunteers. While one study showed an up-regulation of SICI reflecting the inhibitory influence from local, GABAA-mediated, inhibitory interneurons in M1 (Stinear and Byblow, 2004), consecutive studies could not confirm this result (Sohn and Hallett 2004; Beck et al. 2008). There is also inhibitory interaction between the two primary motor areas (inter-hemispheric inhibition (IHI); Ferbert et al. 1992). IHI is mediated through excitatory transcallosal fibers that project onto a subset of the local, GABAA-ergic, inhibitory interneurons in M1 (Ferbert et al. 1992; Chen 2004). IHI most likely uses different interneurons from those used by SICI, but the two mechanisms interact (Daskalakis et al. 2002). IHI seems to be involved in the coordination of the two M1 areas in finger movements (Chen et al. 2004), but there is no increase in IHI between homologuous surrounding muscles during any phase of this motor task in healthy controls (Beck et al. 2009). Therefore, IHI does not seem to contribute to the generation of SI, although it is reduced in FHD patients leading to the clinical phenomenon of mirror dystonia (Beck et al. 2009a).
The relative increase in SI for CRT compared to SRT found in the recent study may implicate higher order motor areas involved in movement selection, such as PMd, PMv or SMA, or the parietal cortex. PMd is thought to play an important role in the selection of movement execution in response to learned associations (Rushworth et al. 2003). Functional imaging studies show a side-dominance of the left PMd, which is activated during right- and left-handed movements, while the right PMd is more active during movement of the left hand depending on the task (Schluter et al. 2001), Similarly, the application of TMS over the left PMd has been shown to disrupt motor performance bilaterally, while stimulation of the right PMd only impaired the performance of the left hand (Schluter et al. 1998). A recent study assessed the contribution of left PMd to SI in healthy volunteers and FHD patients using an established paired pulse TMS paradigm over left PMd and left M1 (premotor-motor inhibition (PMI); Civardi et al. 2001). No increase in PMI was found for the surrounding muscle, neither before and during the first phase of EMG in the active muscle (Beck et al. 2009a), indicating that this interaction may not contribute significantly to the up-regulation of SI.
In humans, there is evidence from neuroimaging studies that the dorsal medial frontal cortex (dMFC) including pre-SMA, SMA and anterior cingulate are involved in free selection of motor behaviour (Brass and Craymon 2002) and the solution of conflicts in action selection (Taylor et al. 2007). These circuits may be involved in the generation of SI, but have not been assessed using TMS so far.
The bilaterality of SI observed in the current study further supports the hypothesis that areas involved in motor planning, such as PMd or parietal cortex, may contribute to SI (Rushworth et al. 2003). Both areas are involved in the preparation of movements and to switch between intended movements, as is required in the current experiments (Rushworth et al. 2003). However, as noted earlier, we have already demonstated that the PMd is not responsible (Beck et al. 2009b).
More pronounced SI was observed in adjacent and distal muscles than in homologous and proximal muscles (Sohn et al., 2003) underlining that surround inhibition is shaped in space. The current study extends these findings to the phase before EMG onset and demonstrates that there is no significant difference in the amount of SI between the two sides. It is not clear how early SI actually starts, since inhibition was already maximal at the first phase tested in the current experiment. The inhibition may be influenced by circuits connecting prefrontal and subcortical areas with the primary motor areas (Aron et al. 2004) and the local inhibitory network in M1 (Stinear et al. 2004; Beck et al. 2008). Further studies will be needed to determine the onset.
In summary, the current results show for the first time that SI can be observed as early as 75ms before EMG onset. With increasing task difficulty, SI increased and started earlier. Although it remains unclear, how SI is generated in healthy subjects, the early onset of SI during CRT and the observation that SI was induced bilaterally to a similar extent may indicate that some “higher order” circuit, such as those involving premotor areas or fronto-thalamo-cortical circuits may contribute to the genesis of SI. Further studies will be needed to assess the specific input from these motor areas onto M1 in the time course of SI in order to better describe the contributing network.
This work was supported by the Deutsche Forschungsgemeinschaft (DFG, BE-3792/1) and by the Intramural Research Program of the NINDS, NIH.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.