In the present experiment, we set out to investigate the combined effects of high-frequency rTMS and peripheral iHFS. Immediately after the 5-Hz rTMS, subjects showed a significant increase in cortical excitability as measured by a decrease in intracortical suppression, which continued to build up for at least 25 min after stimulation. When peripheral iHFS was applied, however, this continued increase was prevented. In contrast, rTMS produced an improvement in tactile acuity, which remained stable for at least 25 min after the end of stimulation, and was not affected by the additional application of iHFS.
Paired-pulse suppression of the median nerve somatosensory evoked potential
During the last few years, stimulation with pairs of stimuli in close succession (paired-pulse stimulation) has become a common tool to investigate short-term plasticity. This is a useful technique to investigate changes in, and the balance between, cortical excitation and intracortical inhibition. Paired-pulse suppression describes the phenomenon that, at short ISIs, neuronal responses to the second stimulus are significantly reduced. Paired-pulse suppression is quantified in terms of the ratio of the amplitude of the second response divided by the first. That means that large ratios are associated with reduced paired-pulse suppression, and small amplitude ratios are associated with stronger paired-pulse suppression. The fact that the second response of two stimuli given in short succession is strongly suppressed has often been denoted as a special form of short-term plasticity, which describes changes of neural behaviour resulting from prior activity (Zucker, 1989; Zucker & Regehr, 2002). Paired magnetic stimulation of the human motor cortex is frequently used to characterize different forms of intracortical inhibition and facilitation (Kujirai et al., 1993; Chen, 2004; Di Lazzaro et al., 2005). In these studies, GABAergic interneurons have been suggested as mediators of paired-pulse inhibition. However, the cellular mechanism underlying paired-pulse suppression of SEPs is not yet fully understood. According to in vitro studies, GABAergic inhibition appears to also play an important role in paired-pulse suppression (Porter & Nieves, 2004; Torres-Escalante et al., 2004). Höffken et al. (2010) reported that, with an ISI of 30 ms, there is no paired-pulse suppression of potentials originating in the cranial medulla, suggesting that, at this ISI, paired-pulse suppression must occur at least at the level of the thalamus or intracortically.
Effects of repetitive transcranial magnetic stimulation and intermittent high-frequency tactile stimulation on cortical excitability
The increase in cortical excitability after the 5-Hz rTMS stimulation was similar for both groups. This finding is consistent with previously published results, where this effect was seen after a similar rTMS application (Ragert et al., 2004). Furthermore, there was a significant further increase in excitability demonstrated in the last measurement for the group that did not receive iHFS. This suggests that there is a time window in which the effect of rTMS on cortical excitability continues to build up, even after stimulation has ceased, before it begins to return to baseline. Similar findings have been reported elsewhere, e.g. Peinemann et al. (2004). In their study, 1800 pulses of rTMS applied to the primary motor cortex, also at a rate of 5 Hz, produced an increase in MEP amplitude that continued to build up after the stimulation ceased, as demonstrated by a second measurement taken 15 min after the end of the stimulation session. Conceivably, this observation might reflect a common finding in rTMS studies, in which repeated post-stimulation assessments have been performed. The data from Peinemann et al. (2004) suggest that the amount of stimulation used might play a crucial role in determining the time course. It is possible that, depending on the stimulation, different populations of neurons are involved, which react with different time courses due to saturation effects. It should be noted that, in in vitro synaptic plasticity experiments, which use much higher frequencies (e.g. 100 Hz), typically maximal effects are observed immediately after the stimulation.
In our study, application of iHFS clearly cancelled this further increase in cortical excitability. Both groups exhibited an almost identical increase in excitability immediately after rTMS (Δbaseline – rTMS), but the last measurement (Δbaseline – last) demonstrated a marked difference between them (Fig. 4B).
Other studies have shown such interactions between tTMS stimulation and subsequent interventions. Delvendahl et al. (2010) showed that pre-treatment with very low-frequency rTMS at 0.1 Hz inhibits the effects of subsequent PAS, whether in its excitatory or inhibitory form. A further study has described a similar effect of 5-Hz rTMS on the subsequent application of either continuous or intermittent theta burst stimulation (Iezzi et al., 2011). In these two studies, the effects of priming are attributed in one case to “antigating” (Delvendahl et al., 2010) and in the other to another non-homeostatic form of interaction (Iezzi et al., 2011). Our experiment resembles these studies in that 5-Hz rTMS effectively abolished the effect of subsequent iHFS on cortical excitability. However, our study differs in that our “priming” intervention produced a strong effect in excitability, the temporal course of which was altered by subsequent iHFS, in a way that might indicate a homeostatic interaction.
Influence of the baseline state of excitability
In the group without iHFS, the change in paired-pulse suppression seen at the end of the experiment (Δ last – baseline) was strongly dependent on the baseline state of excitability, as demonstrated by a highly significant inverse correlation (Fig. 6D) between the final change in the PPR and the naive state values. Taking this into account, it is possible that normal fluctuations in the population in terms of their state of cortical excitability could explain the observed variability in responses to interventions such as rTMS. The importance of the baseline state of excitability of the brain in shaping the effect of an intervention such as rTMS is becoming increasingly recognized (Silvanto & Pascual-Leone, 2008; Silvanto et al., 2008). Indeed, the main goal of homeostatic plasticity studies is to control this directly by means of a ‘priming’ stimulus, as opposed to letting it vary normally, so as to optimize any effect of an intervention protocol (Fricke et al., 2011).
The correlation between the change in the PPR and baseline state was not evident in the measurement taken immediately after rTMS, although the average increase in the PPR even at that point was statistically significant. This is notable as it indicates that the influence of the baseline state of excitability on the response to rTMS is not present immediately after the stimulation has ended, but rather requires a time lapse to build up. This may indicate that the changes observed in the final measurement represent something closer to a ‘final’ size of response, before the effect begins to fade. However, this cannot be ascertained without a more prolonged period of post-stimulation testing.
In the group that also received iHFS, this correlation between the baseline condition and the final measurement was not present, indicating that iHFS had a disruptive effect on the normal time course of the response to rTMS. It is important to note that, in the group that received rTMS alone (Group 2), the PPR increased significantly after 25 min compared with the values obtained immediately after rTMS. This makes it unlikely that the lack of further increase in the PPR after iHFS in Group 1 was simply due to a ceiling effect, as after rTMS the PPR value was almost identical for both groups. Furthermore, in the group that received iHFS alone (Group 3), the baseline value of the PPR approximated the value found in the other two groups after rTMS. This did not prevent iHFS from producing a significant increase in the PPR, suggesting that the lack of effect of iHFS in Group 1 depended on the previous history of activity rather than on the value of the PPR at the time of stimulation.
Effects on tactile acuity
In contrast to the results obtained for cortical excitability, rTMS and iHFS showed no significant interaction in their effect on tactile acuity. Both groups experienced a significant improvement in two-point discrimination immediately following rTMS, which remained unchanged in the last assessment, with or without iHFS.
A previous report, in which a similar rTMS protocol was used, also showed that the induced change in tactile acuity was strongest immediately after stimulation, and slowly reverted to baseline values over the following hours (Tegenthoff et al., 2005). This represents a marked difference from the effect of rTMS on cortical excitability, which, as was shown above, is considerably stronger 25 min after the end of stimulation than immediately after. In addition, the effect on the PPR was highly sensitive to iHFS, whereas iHFS had almost no influence on the rTMS-induced change in tactile acuity.
It seems likely that this difference is due in part to the fact that two-point discrimination is a complex perceptual task that engages many brain areas outside the SI and, although the latter is involved in the two-point discrimination task, other areas play important roles. The results do, however, suggest that the rules governing the effect of plasticity-inducing interventions, and especially interactions between them, are complex, and depend on what type of data is considered to be indicative of plasticity (e.g. behavioural vs. neurophysiological). A similar dissociation between changes of excitability and behavioural measures has been described for the SI following PAS (Litvak et al., 2007). In these experiments, a gain in tactile acuity depended on whether TMS applied to the SI was near-synchronous to afferent signals containing either mechanoreceptive or proprioceptive information. In the latter case, acuity remained unchanged despite changes in excitability, which questions a simple relation between enhancement of synaptic efficacy and behavioural gain. In another study, facilitative PAS has been reported to inhibit motor learning (homeostatic interaction), only if 90 min were allowed to elapse between PAS and motor practice (Jung & Ziemann, 2009). If motor practice was carried out immediately after PAS, then PAS actually improved learning (non-homeostatic interaction). In contrast, studies that explore homeostatic plasticity using MEPs as an indicator often find that such effects develop immediately. Furthermore, the time window during which homeostatic plasticity can be demonstrated using this paradigm appears to be relatively short, as revealed by studies in which short priming interventions were used. In such cases, even a 5- or 10-min interval between interventions is sufficient to abolish homeostatic interaction (Huang et al., 2010; Iezzi et al., 2011).
The lack of significant influence of iHFS on tactile acuity when applied after rTMS contrasts with the results previously reported by Ragert et al. (2003), in which the two types of stimuli produced an additive effect. This shows that the manner in which the two interventions interact might be dependent on their timing. In a previous study (Nitsche et al., 2007), it was shown that the same two plasticity-inducing techniques (tDCS and PAS) interact homeostatically when applied simultaneously and synergistically when applied in succession. This, as the authors point out, contradicts previous results combining tDCS and rTMS (Lang et al., 2004; Siebner et al., 2004), which showed a homeostatic interaction after sequential application. This indicates that the mode of interaction between two interventions (i.e. homeostatic or synergistic) may also depend on the specific form of stimulation used. However, once a certain plasticity process is underway, it may exhibit a degree of immunity to further changes induced by additional interventions. Such an explanation has been put forward by Jung & Ziemann (2009) in connection with the above-mentioned finding that was based, perhaps significantly, on functional parameters (motor learning) and not on neurophysiological parameters (e.g. MEPs).
Pearson's analysis showed that there was no correlation between the changes in two-point discrimination and changes in the PPR after either rTMS (Groups 1 and 29) or iHFS (Group 3). Significant correlations between perceptual changes and neural changes have been robustly demonstrated for blood oxygenation level dependent signals and dipole changes (Pleger et al., 2001, 2003; Dinse et al., 2003a,b), whereas a correlation with excitability measures has so far been described only once (Höffken et al., 2007), offering a greater dynamic range of changes, which facilitates the detection of correlations. We therefore assume that, in the present study, because of the large observed fluctuation in the PPR, together with smaller acuity effects, a correlation between the two parameters did not emerge.
Site of effect and interaction
The fact that sequentially applied rTMS and iHFS showed an interaction can be regarded as an indication that the two interventions probably affect, at least in part, the same population of neurones. When one intervention affects the outcome of a second intervention, this is taken to indicate changes in the threshold for the induction of plasticity induced by the first intervention (see e.g. Sale et al., 2011). This is particularly interesting in view of the fact that rTMS and iHFS represent completely different methods of stimulation, with the former activating cortical networks directly, and the latter making use of the sensory pathway to reach the cortex.
The rTMS has the advantage of allowing for localized stimulation of the brain tissue that lies directly under the coil. Although it is not clear exactly which cell populations are predominantly activated during TMS, modelling studies suggest that the induced electric fields are particularly strong around the gyral crowns and lips, and are less likely to extend deep into the sulcal walls (Opitz et al., 2011; Thielscher et al., 2011). In terms of the primary SI in the post-central gyrus, this corresponds broadly with Brodmann area 1. This is, furthermore, the proposed origin of the N20-P25 component of the median nerve SEP, according to many studies (Arezzo et al., 1979; Allison et al., 1989; McCarthy et al., 1991;.) It is thus highly probable that the homeostatic interaction occurred in a neuronal population located on the crown of the post-central gyrus as a result of the two interventions used, rTMS and tactile coactivation, as the latter has been previously shown to effect changes in the same SEP component (N20-P25) that originates in this area (Höffken et al., 2007). However, from our experimental design it cannot be ruled out that interactions between iHFS and rTMS can also occur outside the primary SI. For example, recent data showed that inter-regional interactions can be induced via premotor-to-motor inputs (Pötter-Nerger et al., 2009).
Separate analysis of the raw amplitudes of P1 and P2 for both groups showed that all changes in the PPR were mediated primarily by an increase in the amplitude of P2, whereas P1 underwent no significant change. A similar finding was reported by Ragert et al. (2004) after a similar application of rTMS over the S1. In fact, of numerous studies that have used rTMS applied directly over the primary SI, none has found changes in the early components of the SEP when measured as single pulses (single-pulse SEPs), that could be considered as analogous to the first peak of a paired-pulse paradigm (Enomoto et al., 2001; Restuccia et al., 2007; Nakatani-Enomoto et al., 2012). This indicates that the effect of rTMS is focused on the mechanism responsible for paired-pulse suppression, rather than the excitability of thalamocortical afferents. In contrast, the related technique of PAS applied over the S1 has proven capable of modulating the amplitude of single-pulse SEPs (Wolters et al., 2005; Pellicciari et al., 2009), although this effect has not been consistently reproducible (Bliem et al., 2008; Tamura et al., 2009).