Distance-adjusted motor threshold for transcranial magnetic stimulation

Abstract

Objective

To examine the relationship between coil–cortex distance and effective cortical stimulation using transcranial magnetic stimulation (TMS) in the left and right motor cortex. We also compare the effect of coil–cortex distance using 50 and 70 mm figure-eight stimulating coils.

Methods

Coil–cortex distance was manipulated within each participant using 5 and 10 mm acrylic separators placed between the coil and scalp surface. The effect of cortical stimulation was indexed by resting motor threshold (MT).

Results

Increasing distance between the coil and underlying cortex was associated with a steep linear increase in MT. For each additional millimetre separating the stimulating coil from the scalp surface, an additional ∼2.8% of absolute stimulator output (∼0.062 T) was required to reach MT. The gradient of the observed distance effect did not differ between hemispheres, and no differences were observed between the 50 and 70 mm TMS coils.

Conclusions

Coil–cortex distance directly influences the magnitude of cortical stimulation in TMS. The relationship between TMS efficacy and coil–cortex distance is well characterised by a linear function, providing a simple and effective method for scaling stimulator output to a distance adjusted MT.

Significance

MT measured at the scalp-surface is dependent on the underlying scalp–cortex distance, and therefore does not provide an accurate index of cortical excitability. Distance-adjusted MT provides a more accurate index of cortical excitability, and improves the safety and efficacy of MT-calibrated TMS.

Introduction

Transcranial magnetic stimulation (TMS) is a non-invasive method for inducing local increases in brain activity. The unique capacity to manipulate neuronal activity within the intact human brain has established TMS as an important technique within both experimental (Walsh and Cowey, 2000, Chambers and Mattingley, 2005) and clinical (Wassermann and Lisanby, 2001, Currà et al., 2002) neuroscience. In the laboratory, TMS is typically applied as a “virtual lesion” technique (Pascual-Leone et al., 1999) to examine the behavioural consequences of neural disruption. In contrast to natural lesions, however, TMS permits experimental control of both the location and duration of cortical disruption. Increasingly, TMS is also being used in conjunction with neuroimaging methods such as functional magnetic resonance imaging (fMRI: e.g., Bestmann et al., 2005) and event-related potentials (ERP: e.g., Taylor et al., 2006). The combination of stimulation and measurement enables the neurophysiological effects of TMS to be measured as well as the behavioural consequences. Clinical research has revealed the therapeutic potential of TMS, in particular for major depression (Wassermann and Lisanby, 2001), and the technique has also emerged as a potential diagnostic tool for detecting neuropathological changes in cortical excitability (Frasson et al., 2003).

In all of these domains, safe and effective application of brain stimulation requires accurate control over induced cortical excitation. Although TMS is widely considered a safe method for inducing cortical activity, over-stimulation increases the risk of known adverse effects, including headaches, nausea and in extreme cases, seizures (Wassermann, 1998, Machii et al., 2006). Furthermore, over-stimulation reduces the focality of the induced cortical excitation (Thielscher and Kammer, 2004). Conversely, under-stimulation may reduce the efficacy of a prescribed treatment in a clinical setting (Mosimann et al., 2002), and in the laboratory, decreases the likelihood of obtaining statistically significant results. Finally, random variations in the level of cortical stimulation across experimental conditions increase experimental error, reducing statistical power; and non-random variations will confound experimental contrasts, potentially yielding artefactual results.

Commercial stimulators express TMS intensity according to the percent of maximum stimulator output. However, the actual intensity of the generated electromagnetic field is determined by a number of stimulator-specific parameters such as the waveform and duration of the magnetic pulse, and induction coil properties such as size, shape and number of copper windings. Therefore, to provide a measure of stimulation intensity that can be generalised across different TMS-configurations, stimulation output is typically calibrated according to an overt and reliable physiological index of cortical excitation, such as an evoked motor response following stimulation of the contralateral primary motor cortex (M1). To provide a discrete index, the threshold of the motor response, or motor threshold (MT), is generally determined using a staircase method of titration (Rossini et al., 1994). TMS intensity expressed as a percentage of MT provides a measure of applied stimulation that can be generalised across coil geometry and stimulator types, and is used to define standard safety guidelines (Wassermann, 1998). However, MT does not provide a direct measure of intrinsic cortical excitability; several recent studies have demonstrated that MT is also strongly influenced by individual differences in the distance between the scalp and underlying motor cortex (Kozel et al., 2000, McConnell et al., 2001, Stokes et al., 2005).

The physical principles of electromagnetic induction state that the magnetic flux density decreases with distance; therefore, the effect of TMS is inversely proportional to the distance from the stimulating coil (Jalinous, 1991). In practice, scalp–cortex distance critically determines the minimum distance between the stimulating coil and the underlying cortical tissue. Consequently, individual variations in scalp and skull thickness directly influence the magnitude of the induced secondary current within underlying cortex. Previous studies have found that scalp–cortex distance influences the effect of TMS indexed by MT (Kozel et al., 2000, McConnell et al., 2001, Stokes et al., 2005), response to treatment (Mosimann et al., 2002), metabolic response indexed by fMRI (Nahas et al., 2000), and also neurophysiological response measured via intra-cranial recordings (Wagner et al., 2004)

Previously, we developed a novel technique to isolate the effect of distance from individual variations in cortical excitability (Stokes et al., 2005). By systematically varying the distance between the scalp surface and stimulating coil, we demonstrated a steep linear relationship between the scalp–coil distance and the percent of stimulator output required to reach MT. Specifically, using a Magstim Rapid system connected to a 70 mm figure-eight stimulating coil, we found that for each millimetre separating the scalp and coil surface, an additional ∼2.9% of stimulator output (∼0.064 T) was required to reach MT. The effect of distance has clear implications for studies that use MT as an index of cortical excitability, and in particular, for studies that use stimulation protocols calibrated to MT for stimulating cortical regions beyond M1. Anatomical analysis reveals substantial scalp–cortex distance variations between different cortical sites (Okamoto et al., 2004, Knecht et al., 2005, Stokes et al., 2005). Consequently, MT-calibrated TMS of non-motor regions with greater scalp–cortex distances than M1 will result in under-stimulation; whereas MT-base TMS of non-motor cortical regions that lie closer to the scalp will result in over-stimulation.

To account for the strong and quantifiable effect of distance, we derived a simple linear correction to calculate a distance-adjusted MT (Stokes et al., 2005):

$$\text{AdjMT}=\text{MT}+m\times (D_{\text{Site}X}-D_{\text{M1}})$$

where AdjMT is the adjusted MT in % stimulator output, MT is the unadjusted MT in % stimulator output, D_M1 is the distance between the scalp and M1, D_SiteX is the distance between the scalp and a second cortical region (SiteX), and m is the distance-effect gradient. We argue that distance-adjusted MT provides a more direct index of cortical excitability than conventional MT, reducing the risk of over-stimulation (e.g., Chambers et al., 2006b), and increasing the validity of quantitative comparisons between cortical sites (e.g., Chambers et al., 2006a).

In the present study, we applied our previously validated 3-step method (Stokes et al., 2005) to further explore the application of AdjMT. In the first experiment, we examined the distance-effect characteristics in the right and left hemisphere, within participants using a standard 70 mm figure-eight induction coil. In the second experiment, we explored the effect of distance on MT using a 50 mm coil, which is becoming increasingly popular for inducing focal stimulation with reduced superficial artefacts (e.g., Muggleton et al., 2003). The results were consistent with our previous finding: distance significantly influences MT, and the relationship between MT and distance is well characterised by a steep linear function. In both hemispheres, and using both induction coils, we reveal an average distance-effect gradient of ∼2.8% per mm. These results confirm and extend our proposal that AdjMT provides an accurate method to correct for the effect of distance on TMS efficacy.