Re-Test-Reliabilität von TMS-Parametern bei gesunden Probanden

Transcranial magnetic stimulation (TMS) is a useful non-invasive technique for the experimental and clinical study of the human brain’s function and dysfunction. Due to the rising popularity of TMS, a closer examination of the reliability is necessary. In this study, the re-test-reliability of the individual TMS parameters was determined for a large number of healthy subjects. Here, additional factors, such as the examiner, an exercise effect, gender and the length of the re-test interval were considered separately. For evaluation of the CSP, a manual and automated analysis method were compared to each other. The data of the present study were collected as part of a larger project, in which the influence of a polymorphism in the gene SCN1A (IVS5N + 5 G -> A; dbSNP: rs3812718) was examined on cortical excitability. The subjects with homozygous genotype were included in the study. The two accomplished baseline measurements were evaluated. Four different examiners measured the 93 subjects. By means of single and double pulse TMS paradigms, the various TMS parameters, such as teststimulus (TS), motor resting threshold (RMT), intracortical inhibition (SICI) and facilitation (ICF) as well as the cortical silent period (CSP), were determined. For statistical analysis, outliers were filtered out. In order to determine the investigated correlation, parameters from session one were compared to the respective parameters of session 2 and correlation was calculated using Pearson's product moment correlations coefficient. Although the underlying concept of CSP is widely accepted, different definitions for the beginning and the end are used. Since the manual analysis method is a subjective process, an additional automated method (CSPDuration, C. Bauer, Schopp, Germany) was used based on the method published bv Garvey et al. (2001). The correlations of the individual TMS parameters show that RMT (r=0.851, p<0.001) had the strongest retest reliabilities. The CSP had average correlations. However, there was a difference with an even better correlation using the automated CSP analysis compared with manual method (r=0.550, p<0.001). For SICI (r=0.383, p<0.001) and ICF (r=0.034, p=0.756) the correlation between session 1 and 2 was weak to non-existent. As already shown in previous studies, it was confirmed that the reproducibility of the cortical inhibition and facilitation is problematic. While the correlations were poor, a stronger connection between the two sessions for SICI could be found. This coincides with the results of previous studies. If one compares the different investigators, differences can be found. The TR had, regardless of the investigator, consistently strong correlations (all r≥0.9). The RMT reached a correlation of r>0.8 with three out of four investigators. Only one investigator had a slightly lower value of r=0.768. These results show that there is no examiner dependence, neither the change of the examiner between the two sessions nor the investigator himself nor his experience led to changes in the correlations. The results of both CSP (manual as well as automated), however, were conspicuous. The correlations ranged from strong (examiner 3, r=0.861) to very low (examiner 2, r=0.397). The comparison of the two methods of evaluation for the CSP suggests that the automated CSP analysis with r=0.697 seems to be slightly superior to the manual method with r=0.55. Overall, the investigator seems to have the greatest effect on the examined parameters. He shows a greater impact on the re-test-reliability than the evaluation method to determine the CSP period. Therefore, results based on these data support the conclusion that a better re-test reliability can be achieved if repeated tests are performed by the same examiner. The use of an automated CSP analysis provides more subordinated a way to facilitate comparison of different studies. With regard to that, the general use of the automated CSP analysis should be given a thought.