Original ArticlesTranscranial magnetic stimulation (TMS) in controlled treatment studies: are some “sham” forms active?
Introduction
In recent years, transcranial magnetic stimulation (TMS) has shown potential as a treatment for psychiatric disorders (George et al 1996). It is a noninvasive method of brain stimulation that uses strong magnetic fields to induce small electrical currents in the cerebral cortex, depolarizing neurons. The potential of TMS to influence cerebral function has been demonstrated in a number of ways. Transcranial magnetic stimulation of the motor cortex has been reported to increase the excitability of some cortical neurons when delivered at high frequency (10 Hz, 20 Hz; Pascual-Leone et al 1994) or to depress excitability at low frequencies (1 Hz; Wassermann et al 1996). These effects lasted several minutes after a single session of stimulation. Functional imaging studies have shown changes in regional cerebral blood flow during TMS (e.g., Fox et al 1997, Paus et al 1997, Paus et al 1998). Transcranial magnetic stimulation has also been linked with downregulation of β-adrenoreceptors in the rat cortex (Fleischmann et al 1996) and effects on astroglial gene expression in mice (Fujiki and Steward 1997). Thus, there are several putative mechanisms by which TMS may exert a therapeutic effect in psychiatric disorders. Research studies have examined the efficacy of TMS in treating depression (as an alternative to electroconvulsive therapy; e.g., George et al 1997, Kolbinger et al 1995, Pascual-Leone 1996), mania (Grisaru et al 1998b), obsessive–compulsive disorder (Greenberg et al 1997), and posttraumatic stress disorder Grisaru et al 1998a, McCann et al 1998.
Controlled studies have used either stimulation at a different cerebral site Greenberg et al 1997, Grisaru et al 1998b or a “sham” form of TMS George et al 1997, Klein et al 1999, Kolbinger et al 1995, Pascual-Leone 1996 as a control condition. There are difficulties with both these types of control stimulation. Stimulation at other sites (e.g., vertex, occiput, cerebellum) may produce confounding psychiatric effects, either locally derived or by effects on the verum treatment site via indirect pathways. For example, functional imaging studies have demonstrated effects of TMS on cerebral sites far from the site of stimulation (e.g., Fox et al 1997). The alternative, sham TMS, has rarely been studied in terms of its design and cortical effects.
We propose that an “ideal” sham for TMS would have the following characteristics: 1) placement on the subject’s head of a TMS coil identical to that used for treatment to obtain visual and tactile parity with “real” treatment; 2) comparable scalp sensation arising from stimulation of superficial nerves and muscles; 3) similar acoustic artifact of TMS, time locked to the scalp sensation; and 4) no physiologic effect on the cortex. Achieving this ideal would be particularly important in crossover studies. However, this ideal is difficult to attain, as any magnetic stimulus delivered from the same coil in contact with the scalp and powerful enough to produce scalp sensation is also likely to depolarise cortical neurons.
A number of different forms of sham TMS were used in the controlled studies cited above. These mostly involved variations in the position of the stimulating coil relative to the scalp, resulting in differences in the intensity and direction of currents induced in cortical tissue (cf. different arrangements shown over the motor cortex in Figure 1). Pascual-Leone et al (1996) and George et al (1997) used a figure-eight coil placed tangential to the scalp for real treatment and with its lateral edge touching the scalp at 45° for sham treatment (as in Figure 1B; personal communication from both authors). Such an arrangement was reported not to produce evoked potentials or to cause measurable changes in cerebral glucose metabolism when used over the motor cortex (George et al 1997). However, patients may have been alert to any subjective differences with sham stimulation, given the crossover design of these trials. Klein et al (1999) used a circular coil, tangential for real TMS and “perpendicular to the scalp surface without direct contact” for sham treatment. Scalp sensation was not mentioned and was presumably absent in this arrangement. Kolbinger et al (1995) used a tangential circular coil for both treatments but reduced the machine output to a very low level (0.05 T) for sham TMS. This was sufficient to induce an acoustic artifact (which would have been much reduced in intensity) but is unlikely to have produced any scalp sensation.
All of the authors of the studies above found significant differences between the outcomes of real and sham treatment. However, in our own double-blind, controlled study in depressed patients we failed to find a significant difference between real and sham TMS treatment, with both groups demonstrating improvement (Loo et al 1999). We concluded that this improvement (about 25%) was probably because of “nonspecific” clinical factors, but we could not rule out the possibility that the sham procedure used (figure-eight coil with front edge touching the scalp at 45°, as in Figure 1A), while being weaker than real treatment, did in fact deliver a clinically meaningful stimulation. In this study, we sought to investigate this issue further.
Regrettably, little is known about the relative intensity of cortical stimulation resulting from various coil positions. Lisanby et al (1998) measured the electrical currents induced in cerebral electrodes implanted in live, nonhuman primates while stimulation was applied with a figure-eight coil in various positions. Variations from a tangential position reduced the magnitude of induced currents by 25–73%. How this translates into physiologic effects is not readily apparent, as neuronal depolarization is determined by a complex set of factors apart from the magnitude of induced currents. These include the direction of currents relative to neuronal orientation (Amassian et al 1992) and the conductivity of surrounding tissue (Barker 1991). The effects of coil orientation and position on cortical stimulation therefore require study in human subjects. In the current investigation, seven coil positions were studied (Figure 1) to encompass the common variations reported in published studies and further the search for an ideal sham. While the prefrontal cortex remains the focus of interest in psychiatric disorders, our study was limited to the motor cortex because of the practical advantage the latter offers in the ability to measure evoked responses in peripheral muscles.
Since subjective sensation is an important component of any sham treatment, we also administered single TMS pulses over the left dorsolateral prefrontal area, a common treatment site in psychiatric studies (e.g., Pascual-Leone et al 1996), to determine the relative degree of scalp sensation produced by different coil positions.
Section snippets
Subjects
Nine normal adult subjects (four male, two left-handed, aged 25–45 years) participated in the study with informed consent and institutional ethics committee approval. Subjects were seated comfortably throughout the study with the right hand resting on a support. The electromyogram (EMG) was recorded from the right first dorsal interosseus (FDI) muscle through surface electrodes. The EMG was amplified and filtered (20 Hz–1 kHz) and recorded to disk through a laboratory interface (CED 1902, Micro
Results
Motor thresholds (T) for the subjects ranged between 38% and 66% of maximal stimulator output and did not change significantly during the course of the experiment (paired t test, p = .29). Two subjects had thresholds above 60% of maximal stimulator output and could therefore only be tested at levels up to 1.5 T.
Discussion
Stimulation over the motor cortex was used as a test for the functional efficacy of TMS with a figure-eight stimulating coil held in various positions. Stimulation conditions were tested with the target muscle at rest and during voluntary contraction, as the latter increases both the likelihood that a stimulus will activate neurons in the cortex and that any descending volley evoked by the stimulus will activate enough spinal motoneurons to provide a detectable response in the muscle (e.g.,
Acknowledgements
This study was supported by the Australian National Health and Medical Research Council Mood Disorders Unit Program Grant No. 993208, private donations from Mr. Peter Joseph (Director, Bankers Trust, Australia) and Mr. William Loewenthal (Company Director), and a part-time Research Fellowship from the New South Wales Institute of Psychiatry (CKL).
References (26)
- et al.
Modelling magnetic coil excitation of human cerebral cortex with a peripheral nerve immersed in a brain-shaped volume conductorThe significance of fiber bending in excitation
Electroencephalogr Clin Neurophysiol
(1992) - et al.
High frequency transcranial magnetic stimulation mimics the effects of ECS in upregulating astroglial gene expression in the murine CNS
Mol Brain Res
(1997) - et al.
Effect of transcranial magnetic stimulation in posttraumatic stress disorderA preliminary study
Biol Psychiatry
(1998) - et al.
Magnetic brain stimulation with a double coilThe importance of coil orientation
Electroencephalogr Clin Neurophysiol
(1992) - et al.
Intracortical facilitation and inhibition after paired magnetic stimulation in humans under anaesthesia
Neurosci Lett
(1995) - et al.
Rapid-rate transcranial magnetic stimulation of left dorsolateral prefrontal cortex in drug-resistant depression
Lancet
(1996) - et al.
Use and safety of a new repetitive transcranial magnetic stimulator
Electroencephalogr Clin Neurophysiol
(1996) An introduction to the basic principles of magnetic nerve stimulation
J Clin Neurophysiol
(1991)- et al.
Transcranial magnetic stimulation downregulates β-adrenoreceptors in rat cortex
J Neural Transm
(1996) - et al.
Imaging human intra-cerebral connectivity by PET during TMS
Neuroreport
(1997)