Classical conditioning of faciliatory paired-pulse TMS

Abstract

In this proof-of-concept study, we questioned whether the influence of TMS on cortical excitability can be applied to classical conditioning. More specifically, we investigated whether the faciliatory influence of paired-pulse TMS on the excitability of the human motor cortex can be transferred to a simultaneously presented auditory stimulus through conditioning. During the conditioning phase, 75 healthy young participants received 170 faciliatory paired TMS pulses (1st pulse at 95% resting motor threshold, 2nd at 130%, interstimulus interval 12 ms), always presented simultaneously with one out of two acoustic stimuli. In the test phase, 20 min later, we pseudorandomly applied 100 single TMS pulses (at 130% MT), 50 paired with the conditioned tone-50 paired with a control tone. Using the Wilcoxon-Signed Rank test, we found significantly enhanced median amplitudes of motor evoked potentials (MEPs) paired with the conditioned tone as compared to the control tone, suggesting successful conditioning (p = 0.031, responder rate 55%, small effect size of r = - 0.248). The same comparison in only those participants with a paired-pulse amplitude < 2 mV in the conditioning phase, increased the responder rate to 61% (n = 38) and effect size to moderate (r = - 0.389). If we considered only those participants with a median paired-pulse amplitude < 1 mV, responder rate increased further to 79% (n = 14) and effect size to r = - 0.727 (i.e., large effect). These findings suggest increasingly stronger conditioning effects for smaller MEP amplitudes during paired-pulse TMS conditioning. These proof-of-concept findings extend the scope of classical conditioning to faciliatory paired-pulse TMS.

Figure 1

Experimental design. Over the conditioning phase, we applied 170 faciliatory paired TMS pulses to the representation of the right abductor pollices breves muscle over the left primary motor cortex while recording muscle evokes potentials (MEPs) from the target muscle. The first (subthreshold) TMS pulse was adjusted to 95% passive motor threshold (MT), the second (suprathreshold) pulse to 130% passive MT. To induce a faciliatory effect, inter-stimulus interval (ISI) between both TMS pulses was set to 12 ms. Each first subthreshold pulse of each paired-pulse TMS application was paired with one out of two acoustic stimuli (830 Hz or 1480 Hz, counterbalanced across participants). The Intertrial intervals was 6 s long. The conditioning phase was 18 min long. 20 min after the conditioning phase, the test phase started. In the test phase, we applied 100 suprathreshold single TMS pulses with the same 130% MT as used for the second TMS pulse during the conditioning phase. Fifty single TMS pulses were paired with the conditioned tone (white bells) and 50 with the control tone (dark bells). Their order was chosen pseudorandomly. The tone and the single TMS pulse were presented simultaneously as during the conditioning phase. Like in the conditioning phase, inter-trial intervals were set to 6 s. The statistical comparison of single-pulse TMS MEP amplitudes paired with the conditioned tone vs. control tone was used to assess successful conditioning.

Figure 2

MEP amplitudes (in mV) over the conditioning phase (left boxes) and the test phase (right boxes). Conditioning phase: median MEP amplitudes (grouped in blocks of 20–30 trials) showed no significant shifts over the course of conditioning (Kruskal–Wallis test, chi-square = 0.67, df = 5, p = 0.123, eta² = 0.024). Test phase: MEP amplitudes of single TMS pulses paired with the conditioned tone were significantly higher than for the control tone (n = 75, Wilcoxon-Signed Rank, z = − 2.15, p = 0.031, r = − 0.248; 55% responder), suggesting successful conditioning of TMS-induced intracortical facilitation. The top of each box indicates the 1st quartile, the bottom the 3rd quartile. The horizontal line within the box marks the median, whiskers indicate minimum and maximum, dots outliers.

Figure 3

Participant groups according to their median paired-pulse MEP amplitude over the conditioning phase (i.e., < 3 mV, < 2 mV, < 1 mV). For participants with a median paired-pulse MEP amplitude during conditioning < 3 mV (n = 55), the difference between MEP amplitude for conditioned (white box-whiskers) vs. control tone (dark box-whiskers) in the test phase was not significant and the effect size was small (p = 0.06, 54% responder rate, r = − 0.249). < 2 mV (n = 38) the responder rate increased to 61% and effect size increased to moderate (r = − 0.389). < 1 mV (n = 13), responder rate increased further to 79% and effect size increased to large (r = − 0.727). These findings suggest that the effect of conditioning related on the height of the paired-pulse MEP amplitudes during the conditioning phase. The top of each box indicates the 1st quartile, the bottom the 3rd quartile. The horizontal line within the box marks the median, whiskers indicate minimum and maximum, dots outliers.

Figure 4

Three exemplary participants—one for each of the three groups (i.e., (a) < 1 mV, (b) < 2 mV, (c) < 3 mV, see Fig. 3). Shown is an overlay of single-pulse MEP responses during the test phase separated into those combined with the conditioned tone (white bells) and those combined with the control tone (dark bells). Vertical lines besides the MEPs index the average across all single-pulse MEP amplitudes during the test phase. Please note that the difference between the MEP combined with the conditioned tone and the MEP with the control tone is strongest for the participant who presented paired-pulse MEP amplitudes during the conditioning phase below 1 mV (a). For the participant with paired-pulse MEP amplitudes below 2 mV the conditioning effect became weaker (b), whereas the participant with median paired-pulse MEP amplitudes below 3 mV showed the weakest conditioning effect (c).