Threshold curves for transcranial magnetic stimulation to improve reliability of motor pathway status assessment

Abstract

Objective

To provide a new protocol for a simple determination of resting motor threshold (MT) and assessment of excitation-inhibition balance in motor cortex and pathways.

Methods

Navigated TMS was used to map cortical representation area of the FDI muscle bilaterally in ten healthy subjects. Reference MTs were determined using a threshold hunting paradigm. Subsequently, a novel stimulation protocol was applied which included 70 stimuli (7 intensities, sub- and suprathreshold). The “MT-curve” was constructed by computing the MTs with several threshold amplitudes with the novel protocol. The measurements were repeated. Sensitivity of the MT-curve to stimulus location was also tested.

Results

The reference MTs agreed with those determined with the novel protocol (R = 0.96–0.99, p < 0.001). Based on coefficient of repeatability derived from non-parametric one-way ANOVA, the repeatability was good (ρ_AO = 0.929, p < 0.05). Generally, the mean difference between the repeated MT-curves was <3% of the maximum stimulator output. Coil movement 10 mm medially from the optimal stimulus location increased that difference to >7%.

Conclusions

The MTs derived using the MT-curve protocol concurred with the reference MTs. The MT-curve is highly reproducible and sensitive to the exact cortical location of stimulation.

Significance

The MT-curves provide a simple way to assess motor pathway status using a single stimulation train. This may be useful in the follow-up and monitoring of motor pathway recovery e.g. from stroke or trauma.

Introduction

Transcranial magnetic stimulation (TMS) pulses targeted to the motor cortex probe the state of the cortex, cortico-cortical and corticospinal connections, finally leading to muscle twitches quantified by electromyography (EMG). The observed strength of motor evoked potentials (MEPs) reflects the capacity of the cortex and pyramidal tract to generate and conduct signals necessary to perform movements. Several variables derived from the MEPs have been found to characterize clinical and pathological conditions in the nervous system (Chen et al., 2008). The most prominent MEP derivations are the recruitment, or input–output (I/O) curves, motor thresholds (MTs), and MEPs’ amplitude and latency. For example, post-stroke MEPs originating from the lesioned hemisphere are sometimes absent and practically always slower and of significantly diminished amplitude, and the normalization of MEPs is correlated with the recovery of motor function; see Chen et al. (2008) for review.

In practice, however, MEPs and in particular the amplitudes have been notoriously variable both within subject and between subjects (Kiers et al., 1993, Säisänen et al., 2008). When entering clinics, the demand for repeatability becomes critical. A typical TMS experiment will take 10–20 min when it includes only the basic measurement of MT and recruitment curves for one muscle. This excludes the mapping of the motor cortex to locate the cortical origin of the most viable pathway to individual muscles. For the monitoring of changes in the patient’s motor conditions during recovery/rehabilitation or for diagnostic measurement, 20 min per muscle may be too long for most of the patients. A long duration of recording adds notable variability in the data because the brain will respond to the stimuli differently depending on varying factors such as physical restlessness, alertness and level of consciousness. The reasons for long exams include the great intrinsic variability of MEPs, which reduces the repeatability and accuracy of the results, demanding more pulses. On the other hand, lengthy recordings do not allow the stabilization of the measurement due to variability of patient’s overall state leading to even more pulses.

The MT describes the “dose” of stimulation, which is able to induce a muscle response of agreed size with a 50% probability. The most commonly used threshold amplitude of MEP for an accepted muscle response is 50 μV (Rossini et al., 1994, Rothwell et al., 1999). Several techniques and protocols have been proposed for determining the MT, most used of which are: the threshold hunting paradigm (Awiszus, 2003), the Mills–Nithi method (Mills and Nithi, 1997) and the Rossini–Rothwell method (Rossini et al., 1994, Rothwell et al., 1999). All of these differ in principle, but provide a similar outcome (Tranulis et al., 2006), none superior over the other. The MT is still a rather primitive measure for the stimulation dose, since it is very much dependent on the used stimulator as well as the coil size, shape and (pulse) type (Kammer et al., 2001). However, MT is rather easy to apply within a lab or clinic when using the same system over and over.

The I/O-curves refer to an increase in MEP-amplitude as a function of stimulation intensity. Instead of assessing purely localized excitability of certain area in the cortex, I/O-curve reflects neurons’ excitability and spread from the centre of activation. The I/O-curves have relation to MTs, as they have been shown to be steeper in muscles with low MT, such as hand muscles (Chen et al., 1998). The slope of I/O-curve is affected by certain drugs (Boroojerdi et al., 2001), and among other physiological disorders it has been shown to be altered by the lesion location and the affected area of the brain in stroke (Liepert et al., 2005).

Navigated TMS has been shown to reduce the variation in MEP size (Julkunen et al., 2009a) although conflicting results have been reported (Jung et al., 2010). It also provides the possibility to localize motor representation areas of certain muscles in relation to individual brain anatomy (Niskanen et al., 2010, Säisänen et al., 2008) or to sensitively modulate brain function based on individual somatotopy (Hannula et al., 2005). Further, navigated TMS has been shown to enhance the reliability in determining MT (Danner et al., 2008, Julkunen et al., 2009a). Consequently, the cortical representation and motor pathway status of specific musculatures can be investigated. This is especially useful during presurgical cortical mapping of motor function in epilepsy or tumour patients (Picht et al., 2009, Vitikainen et al., 2009). The MEP-amplitudes are maximized when the stimulation is targeted specifically to recorded muscles representation area (Rossini et al., 1994). This has been clearly demonstrated by deliberately targeting stimuli off-target leading to reduced MEP-amplitude and prolonged latency (Julkunen et al., 2009a). Recently, navigated TMS has also been used to target repetitive TMS (rTMS) for therapeutic purposes to certain anatomical areas based on anatomical MRI of the patient (Ahdab et al., 2010, Lefaucheur, 2010).

The MT, I/O curve, and MEP measurements today require separate examinations. In the present study, we introduce a repeatable measurement procedure that provides the results as a graph and requires only a small fraction of the total examination time. This allows us to combine the measurements of multiple muscles into a single, quick examination and gives us a way to present the results as an informative graphical curve that is closely related to the I/O-curve.