RMT determination at the hotspot of small hand muscles is considered a measure for motor cortex excitability. Thus, variability of the hotspot plays an important role for nTMS, both in motor and in language mapping, because it determines the accuracy of nTMS investigations.
In this study, still considering the low sample size, intraobserver comparison showed a higher precision than interobserver investigations in motor mapping of the hotspot for the APB (Figures
3). This difference is statistically significant when comparing the measured distances of both samples with regard to x- and z-coordinates and to the hotspot distances in the same peeling depth. In contrast, considering variations in distances between y-coordinates and between all axes, we were not able to show statistically significant differences (Figure
2). This correlates well with the fact that the peeling depth is adjusted by varying the y-coordinates. It can be modified several millimeters without having a crucial effect on the mapping procedure; thus, its setting can vary within a small range (Figure
2). Yet, we have to keep in mind that the small sample size still minors the power of statistical tests and therefore judgment of our findings.
However, we have to question why interobserver variability is larger than intraobserver variability. Previous studies have shown that there are individual differences among humans concerning not only hotspot localization but also hotspot shapes
. Unfortunately, we still do not know much about interindividual variations in measurement of hotspot size.
However, Pascual-Leone, Wassermann et al. (1995) found that the size of the motor areas in the brain for the hand muscles are able to enlarge significantly after only hours of intensive use. Thus, the timing can play an important role in plasticity investigations, especially when the enrolled subjects of our study are university students who have greater than average use of the right hand.
We also compared the RMTs of all three sessions. Previous studies have shown that the value of the individual motor threshold varies
. In our findings, the maximum individual fluctuation was 9% of the total output. Nevertheless, on average, the determined RMTs were almost stable and there was only minimal variation from mapping to mapping (Table
4). As mentioned above, no subject was under any kind of medication. Moreover, special attention was paid on comparable alertness during all mappings.
There already exist several studies on mainly non-navigated TMS that deal with reliability in hotspot determination
[6, 13–18]. Wolf et al.
, for example, achieved a distance of 8.9 mm between two sessions with a range from 0 to 22.4 mm. They used the hotspot of the extensor digitorum communis muscle instead of the APB. Nevertheless, comparison among these studies is only possible to a certain extent because previous investigations were conducted using a different muscle, a different navigation system, or non-navigated TMS.
The most important error source in non-navigated TMS is inexact coil positioning because minor movements of the coil or modifications of the angle can cause significant deviations in field strength and stimulus location
. With the development of nTMS, we are able to reduce this major error source, and respectable accuracy can be attained
[8, 21]. Yet, the nTMS system is composed of several specific units, such as MRI-registration, infrared camera system, stimulation coil, induced E-field calculation, head shape, head tracker, etc., which all add minor inaccuracies to the calculated overall error of the used system. Thus, the accuracy of the system depends on the accuracy and the interaction of such principal factors. As Ruohonen and Karhu (2010) described in detail, the E-field computation model causes the highest inaccuracy (3.8 mm) in the system, followed by shifting of the head tracker (3.1 mm), and imperfect alignment between anatomical MR images and the individual’s head (2.5 mm)
. The localization of the coil constitutes only 1.6 mm deviation.
In the end, the median overall error of the whole system was calculated to be 5.73 mm
. The mean distance of all combined axes was 8.1 mm in our intraobserver investigations and 10.3 mm in our interobserver comparisons. Taking into account that the calculated system error impairs both measurements, the inaccuracy of the whole system can increase up to 11.46 mm (2 × 5.73 mm) in the three-dimensional space. Both the mean intra- and the interobserver distances were within this 11.46 mm value, which means, on the one hand, that the two compared hotspots might actually be located at exactly the same spot, although we measured a distance of several millimeters. On the other hand, the smallest determined distance of 3.0 mm could be much larger in reality.