Applications of Transcranial Magnetic Stimulation to Therapy in Psychiatry
Applications of Transcranial Magnetic Stimulation to Therapy in Psychiatry
Transcranial magnetic stimulation (TMS) is a non-invasive means of stimulating focal regions of the brain using magnetic fields. Since its introduction in 1985, TMS has been used to study localization of brain functions, connectivity of brain regions and pathophysiology of neuropsychiatric disorders. The potential uses of TMS to treat psychiatric disorders are under active study. This article reviews the state of knowledge about the therapeutic potential of TMS in psychiatry.
The TMS Process
Transcranial magnetic stimulation is an investigational medical procedure performed by placing an electromagnetic coil on the scalp (Figure). High-intensity current is rapidly turned on and off in the coil through the discharge of a capacitor. This produces a time-varying magnetic field that lasts for about 100 to 200 microseconds. The magnetic field strength is about 1.5 to 2 tesla (about the same intensity as the static magnetic field used in clinical magnetic resonance imaging) at the surface of the coil, but the strength of the magnetic field drops off exponentially with distance from the coil. The proximity of the brain to the time-varying magnetic field results in current flow in neural tissue and in membrane depolarization. Transcranial magnetic stimulation is experimental; it is not approved by the U.S. Food and Drug Administration for the treatment of any disorder.
A striking effect of TMS occurs when one places the coil on the scalp over the primary motor cortex. A single TMS pulse of sufficient intensity causes involuntary movement in the muscle represented by that region of cortex. Thus, a TMS pulse produces a powerful but brief magnetic field that passes through the skin, soft tissue and skull. This induces electrical current in neurons, causing depolarization that then has behavioral effects. The minimum magnetic field intensity needed to produce motor movement is known as the individual motor threshold.
Repeated application of TMS pulses at regular intervals is called repetitive TMS (rTMS). The physiological effects of TMS depend upon the site and frequency of stimulation. If the stimulation occurs faster than once per second (1 Hz), it is referred to as fast rTMS and can result in excitatory physiologic changes. On the contrary, if the frequency is low, it is referred to as slow rTMS and can have an inhibitory effect on brain excitability. High-frequency rTMS carries a risk of seizure. Guidelines exist to reduce this risk by appropriate screening of participants for seizure risk factors, titrating the individual motor threshold and limiting rTMS dosage (Belmaker et al., 2003; Wassermann, 1998).
The ability to focally alter cortical excitability opens up the potential to modulate cortical circuitry for potential therapeutic benefit. The focality of the effects also presents a challenge to clinical application, because it is necessary to know the circuitry of the underlying disorder to guide where and how to stimulate to ameliorate its symptoms.
Major depression has been the focus of the bulk of TMS clinical trials to date. Initial work used nonfocal coils positioned over the vertex and found some suggestions of clinical benefit. Based on evidence for prefrontal abnormalities in depression, it was thought that rTMS over the prefrontal cortex could produce a more profound antidepressant effect than over other cortical areas. To test this hypothesis, using a within-participant crossover design, Pascual-Leone and colleagues (1996) reported that fast left dorsolateral prefrontal cortex (DLPFC) rTMS for five days had marked antidepressant effects, even for psychotic depression. Stimulation at other sites (right DLPFC, vertex) and sham treatment had no effect. This remarkable result was superior to what could be expected with any medication regimen or even electroconvulsive therapy, but most of the following studies could not replicate the same magnitude or speed of response using similar parameters and length of stimulation.
Extending the treatment to two weeks and increasing the number of pulses per day, a 50% response rate was reported in patients with medication-resistant unipolar depression in an open trial (Triggs et al., 1999). A double-blind, sham-controlled, single-crossover study of fast left DLPFC rTMS in medication-resistant outpatients found improvements with two weeks of active rTMS, but the degree of improvement was modest (George et al., 1997). Herwig and colleagues (2003) reported similar findings with a 30% response rate to real rTMS, and none responded with sham treatment.
There are suggestions that longer duration of treatment may result in more significant improvement. Pridmore et al. (1999) were the first to extend the period of treatment to over two weeks. Patients with melancholic depression referred for ECT were treated with rTMS to the left DLPFC. This resulted in remission in 88% of cases in an open study. The data regarding the application of rTMS to treat major depression suggest that a longer period of stimulation could be required to evoke a clinical response (Gershon et al., 2003). An ongoing double-blind, sham-controlled, multicenter study on the efficacy of rTMS applied to the left DLPFC for six weeks will test whether longer periods of stimulation are more effective in a controlled setting.
It is still an open question as to whether the antidepressant effects of rTMS are region- or frequency-dependent. Klein et al. (1999b) randomized outpatients with depression to two weeks of active slow (1 Hz) rTMS or sham treatment over the right prefrontal cortex using a round, nonfocal coil. In the active group, 49% of patients responded compared to 25% in the sham group. This study suggested that right-sided stimulation might be effective. A much smaller double-blind study replicated these results with the active group, showing a better clinical response compared to the sham treatment (Kauffmann et al., 2004).
If lower frequencies are indeed as effective as higher frequencies, this would have significant safety implications, as lower frequencies carry less risk of seizure. When low and high frequency have been compared in the same study, there are suggestions that participants who receive rTMS at lower frequencies may actually fare better. Kimbrell et al. (1999) found a trend toward better improvement following two weeks of left prefrontal rTMS at 1 Hz compared to 20 Hz in a crossover design. There were suggestions that baseline metabolic activity in the DLPFC correlated with response. Similarly, Padberg et al. (1999) tested patients with nonpsychotic depression with sham treatment, slow rTMS or fast rTMS, all over the left DLPFC. After five days, 83% in the slow group and 50% in the fast group improved, with no change in the sham group. A parallel-design, blinded study suggested that slower (5 Hz) left prefrontal TMS may be as effective as faster (20 Hz) stimulation (George et al., 2000). Fitzgerald et al. (2003) demonstrated that both high-frequency left rTMS and low-frequency right rTMS have benefits in patients with medication-resistant major depression. They concluded that treatment for at least four weeks is necessary for clinically meaningful benefits to be achieved.
There has been great interest in determining whether rTMS could offer an alternative to ECT for severe or treatment-resistant depression, particularly because the adverse-effect profile of rTMS is relatively benign. Using a parallel-group, nonblinded design, Grunhaus and colleagues (2000) randomly assigned inpatients to treatment with fast left DLPFC rTMS or ECT. Among patients who were nonpsychotic, up to four weeks of daily rTMS was not different in efficacy to ECT, but ECT was superior among patients with psychotic depression. Patients treated with rTMS or ECT showed the same percentage of clinical stabilization at three and six months of follow-up (Dannon et al., 2002). Janicak et al. (2002) randomized patients with severe depression to be treated with rTMS or ECT. No significant difference in efficacy was found between the two treatments for improvement on the Hamilton Rating Scale for Depression (HAM-D; 55% improved with rTMS, and 64% improved with ECT).
All these studies had the limitation that the patients were not blinded to the form of treatment, and some have questioned whether the ECT comparison group represented optimal ECT practice. Nevertheless, it would be impossible to blind the patient to the treatment modality in this case (because sham ECT would not be considered ethically acceptable), and all studies have found the side-effect profile to favor rTMS.
Several meta-analyses have examined the antidepressant efficacy of rTMS (Holtzheimer et al., 2001; Martin et al., 2003, 2001) and concluded that there is evidence for statistical benefit of rTMS. However, the effect size could be described as moderate and in some cases of limited clinical significance. For example, Burt et al. (2002) found an average percent improvement with active TMS of 28.94% (SD=23.19) and with sham treatment of 6.63% (SD=25.56). Relatively few patients met standard criteria for response or remission.
It is also true, however, that the meta-analyses are heavily weighted towards the earlier studies that used what may now be considered to be adequate dosages and durations of rTMS.
Further work using controlled designs is needed to determine whether the antidepressant effects of rTMS are region-, frequency- or intensity-dependent, and to test the efficacy of more robust parameters in a sample large enough to provide adequate statistical power.
Such work is presently underway, with multicenter trials sponsored both by industry and the National Institute of Mental Health.
Electroconvulsive therapy is effective in depression as well as in mania, so it was hypothesized that rTMS might likewise have similar efficacy. It was also thought from imaging studies and post-stroke mood effects that the laterality of mania may be opposite to that of depression and involve more heavily the right prefrontal cortex.
In the first study of rTMS for acute mania, Belmaker and Grisaru (1999) tested that hypothesis by randomly assigning patients with mania to fast rTMS to the left or right DLPFC as an add-on to standard pharmacotherapy. After two weeks, the group treated with right DLPFC rTMS did better, consistent with the theory that the laterality of fast rTMS necessary for antimanic effects is opposite to that needed for antidepressant effects. However, they could not replicate the antimanic effect in a subsequent follow-up study of right active rTMS versus right sham rTMS, wherein right rTMS was no more effective than sham rTMS. It is possible that the previous results were due to a worsening effect of left rTMS on mania. Additionally, this patient group had much more psychosis than the previous study of rTMS in mania, and depression studies have reported that psychosis is a poor prognostic sign for rTMS response (Kaptsan et al., 2003). More recently, a study of a four-week course of rTMS showed a sustained reduction of manic symptoms in all patients (Michael and Erfurth, 2004). Due to the open, add-on design of the study, a clear causal relationship between rTMS treatment and reduction of manic symptoms could not be established.