Applications of Transcranial Magnetic Stimulation to Therapy in Psychiatry

The first studies on patients with schizophrenia employed TMS administered with a large round coil to the vertex, thereby stimulating broad regions of bilateral prefrontal and parietal cortices. In 1997, Geller and colleagues reported that 60% of medicated patients with chronic schizophrenia showed some transient improvement following a single session of rTMS in an open study. Using two weeks of 1-Hz rTMS with a smaller round coil positioned on the right prefrontal cortex, 70% of treated patients with schizophrenia were moderately or markedly improved, and psychosis ratings dropped significantly (Feinsod et al., 1998). However, when the same group followed up their findings with a sham-controlled trial, rTMS did not differ from sham (Klein et al., 1999a).

Hoffman et al. (1999) had more success with 1-Hz rTMS when the coil was positioned over the left temporoparietal cortex, a region that has shown selective activation during auditory hallucinations. This trial was based upon the hypothesis that low-frequency rTMS may dampen excitability in the region implicated in this specific symptom (Chen et al., 1997). In an initial crossover study, significant reductions in auditory hallucinations were noted with active rTMS compared with sham treatment (Hoffman et al., 1999). Two of three patients experienced a near-total cessation of auditory hallucinations for at least two weeks. Significant reductions in auditory hallucinations were replicated in a larger crossover trial by the same group (Hoffman et al., 2000).

Other groups have examined the effects of high-frequency rTMS applied to the prefrontal cortex on the theory that high-frequency rTMS might be useful in reversing the hypofrontality observed in schizophrenia. In an open study of patients who received 20-Hz rTMS to the midline prefrontal cortex for at least two weeks, a significant reduction in negative symptoms was observed, but other symptoms and tests of neuropsychological function were essentially unchanged (Cohen et al., 1999). A single session of 20-Hz rTMS was administered to the left DLPFC in a sham-controlled, crossover trial (Nahas et al., 1999). Improvement in negative symptoms was noted the day following treatment. In a more recent two-week, crossover-controlled study, another group reported that active rTMS significantly decreased psychotic symptoms (Rollnik et al., 2000). More controlled studies need to be done to determine whether high-frequency rTMS will be helpful for negative symptoms.

Limited studies, mostly case series, have been conducted on the therapeutic applications of rTMS in anxiety disorders. The majority of neuroimaging studies have shown elevated right-sided activity in the frontal and hippocampal-parahippocampal regions in fear paradigms and anxiety disorders. This has led to the hypothesis that low-frequency rTMS may be helpful in dampening that lateralized hyperexcitability, similar to the rationale behind using 1-Hz rTMS for the positive symptoms of schizophrenia.

Zwanzger et al. (2002) demonstrated a reduction of panic symptoms with a marked improvement of anxiety in an open case study of a patient treated for two weeks with slow rTMS on the right DLPFC. Anxiety symptoms decreased by 78%, and panic/agoraphobia symptoms decreased by 59%. Improvements lasted at one-month follow-up. Interestingly, there was a reduction in cholecystokinin tetrapeptide (CCK-4)-induced panic attacks, associated with blunting of the CCK-4-induced elevation of serum cortisol.

In another open case series, three patients with treatment-resistant panic disorder showed modest improvement with 10 rTMS sessions (1 Hz, 110% of motor threshold, 30 trains of 60-second duration) to the right DLPFC (Garcia-Toro et al., 2002). Alternating low-frequency rTMS to the right DLPFC with 20-Hz rTMS to the left DLPFC failed to produce further benefits. There has yet to be a sham-controlled trial of rTMS in the treatment of panic.

McCann et al. (1998) treated two patients with treatment-resistant posttraumatic stress disorder, both of whom showed elevated baseline cerebral metabolism on positron emission tomography (PET). Slow rTMS to the right DLPFC reduced posttraumatic symptoms and reversed cerebral hypermetabolism, most markedly in the right prefrontal cortex.

In line with symptom provocation studies that have demonstrated significantly greater activity in patients with PTSD in brain regions associated with motor preparedness in response to threat, Grisaru et al. (1998) applied 0.3-Hz rTMS bilaterally to the motor cortex of patients with PTSD. Posttraumatic stress symptom scores improved transiently, but this may underestimate the value of such an approach because the stimulation parameters were overly conservative (only 30 pulses per day).

Rosenberg et al. (2002) hypothesized that left frontal rTMS (either 1 Hz or 5 Hz, 90% of motor threshold, total 6000 stimuli) could mimic the beneficial effect of antidepressant medications in patients with combat PTSD and comorbid major depression. They found that 75% of patients had a clinically significant antidepressant response, but just minimal improvements in PTSD symptomatology. At two-month follow-up, the antidepressant effects were maintained in half the patients.

Cohen et al. (2004), in a double-blind, sham-controlled trial, found beneficial effects of rTMS at 10 Hz, but not at 1 Hz or with sham rTMS, to the right DLPFC in patients with PTSD core symptoms (re-experiencing, avoidance) and other anxiety symptoms. The frequency specificity of these effects, demonstrated in the context of a well-controlled study, offered strong support that the use of rTMS in this disorder is worth exploring further.

Neurophysiological data converge in indicating that cognitive impairment and motor "intrusive" and repetitive behaviors in obsessive-compulsive disorder may be a consequence of a reduction of cortico-cortical inhibitory phenomena and a higher than normal level of cortical excitability. Greenberg et al. (2000, 1998) used the technique of paired-pulse TMS to test whether deficient intracortical inhibition exists in OCD. They found that patients with OCD, like those with Tourette's disorder, had markedly decreased intracortical inhibition. Those with tic-related OCD showed the most profound deficit in intracortical inhibition. Additionally, patients with OCD had lower resting and active motor thresholds than did normal volunteers.

Imaging studies of OCD implicate hyperactivity in a circuit involving orbitofrontal cortex and basal ganglia. To test whether modulating activity in this network could influence OCD symptoms, Greenberg et al. (1997) administered rTMS to the right lateral prefrontal, left lateral prefrontal and a midoccipital (control) site on separate days in a blinded trial. Patients' compulsive urges decreased significantly for eight hours after right lateral prefrontal rTMS. A short-lasting, modest and nonsignificant reduction in compulsive urges occurred after left lateral prefrontal rTMS.

Two other studies have examined possible therapeutic effects of rTMS in OCD. A double-blind study using right prefrontal 1-Hz rTMS and a less focal coil failed to find statistically significant effects greater than sham treatment (Alonso et al., 2001). In contrast, an open study in a group of patients with OCD refractory to standard treatments who were randomly assigned to right or left prefrontal fast rTMS found that clinically significant and sustained improvement was observed in a quarter of the patients (Sachdev et al., 2001). More work will be needed to clarify whether rTMS will be helpful in OCD, but the availability of a defined circuitry should guide the design of such trials.

Routine clinical use of rTMS in psychiatric disorders is far from certain at the present time, but that may change as the results of larger well-controlled trials become available. None of the key effects has been rigorously replicated, and most of the positive findings are based on small samples in short-duration trials. Depression is the condition with the most consistent evidence, but there are discrepancies among the initial studies in the magnitude and nature of the effects. In addition to the usual concerns about sample comparability and the reliability of assessment, the therapeutic application of rTMS has particular methodological issues involving sham application and the parameters used.

In any case, the initial studies suggest that rTMS can exert a variety of both short- and long-term behavioral effects. On the optimistic side, they raise the possibility that focal modulation of cortical excitability can have therapeutic properties in psychiatric disorders and that TMS may prove informative about the anatomy and physiology of the neural systems involved in achieving therapeutic effects. At the clinical level, because its adverse-effect profile is so benign, rTMS may ultimately offer a less-invasive alternative to already established somatic interventions for severe or treatment-resistant illnesses.