Bipolar affective disorder (BPAD) is:

• A serious mental illness
• Estimated to be present in as high as 6.4% of the population in Western populations
• Associated with considerable disability and high morbidity.
• Characterized by periods of both lowered and elevated mood (i.e. depression and mania/hypomania respectively).

The depressive aspect of bipolar disorder is often overlooked, possibly due to its less dramatic nature, despite its significant impact on the lives of those affected. Bipolar depression (BPAD-DP) is associated with a twenty fold increased risk of suicide, and typically lasts three to five times as long as a manic or hypomanic episode. Despite this, there has been relatively sparse investigation of treatments for BPAD-DP, with guidelines based primarily on expert judgment rather than clinical trials. In addition a significant proportion of patients with bipolar depression do not respond to the range of commonly used medications. One of the only substantially new treatments developed for unipolar depression in recent years has been the advent of repetitive transcranial magnetic stimulation (rTMS). Repetitive TMS has been evaluated in over 20 trials conducted over the last ten years, but no substantive trials have explored its use in bipolar depression. We propose to do this, conducting a large scale clinical trial. The trial will include the assessment of both high frequency left sided rTMS (as there is clearly the greatest evidence for the effectiveness of this in unipolar depression) and low frequency right sided rTMS (as this there is growing evidence of the effectiveness of this in unipolar depression and we have an excellent pilot study to suggest its potential in BPAD-DP and it has never previously been assessed in a clinical trial exclusively targeting this patient group). Our previous research strongly supports the effectiveness of rTMS paradigms including low frequency right-sided stimulation in unipolar depression and suggests these may have value in BPAD-DP. As BPAD-DP is clearly a clinical problem of significant impact and with limited treatment options, there is a pressing need for the development and definitive testing of novel treatments such as rTMS.

Bipolar affective disorder is a serious mental illness with significant mortality and morbidity [1]. Some estimates of prevalence in Western populations are as high as 6.4% [2]. The depressive aspect of bipolar disorder is often overlooked, possibly due to its less dramatic nature, despite its significant impact on the lives of those who suffer from it [3]. Bipolar depression is associated with a twenty fold increased risk of suicide [4], and typically lasts three to five times as long as a manic or hypomanic episode [5, 6].

Despite this, there has been relatively sparse investigation of treatments for bipolar depression, with guidelines based primarily on expert judgment rather than randomized controlled trials [7]. First line treatment typically involves treatment with a mood stabilizer - lithium, sodium valproate or carbamazepine [7]. Addition of an antidepressant such as an SSRI, buproprion or venlafaxine may be indicated if depression persists, although there is a possibility of inducing rapid cycling and manic or manic episodes. Alternative agents such as lamotrigine and atypical antipsychotics may also be considered [7]. However, there is little firm evidence in this area, and the lack of controlled studies in the treatment of bipolar depression was described by Thase in 2005 [7], as "an embarrassment to the field". In addition to this lack of evidence of efficacy, it is clear that a significant proportion of patients with bipolar depression do not respond to the range of commonly used treatments including medication combinations [8]. It is clear that the depressive phase of bipolar disorder is an important clinical problem and one in which there is a considerable need for the development of new treatments.

rTMS is a technique that was first developed in the mid 1980's which involves the use of a time variable magnetic field to stimulate brain activity. rTMS methods have generally been found to be well tolerated, safe and without complications such as the use of an anesthetic that are required with more invasive forms of brain stimulation such as electroconvulsive therapy.

The ability of rTMS to affect mood was first noted in normal controls in the late 1980's. Studies using focal stimulation of the dorsolateral prefrontal cortex (DLPFC) in depression first appeared in the mid 1990's [9-11]. These studies produced promising results with reduced depression severity following left prefrontal cortex (PFC) stimulation. Since that time a considerable number of trials of left DLPFC rTMS have been conducted. The majority of these studies have used 5-20 Hz stimulation frequency. Most of these have predominately been small in sample size and of short duration (i.e., 10 treatments) (CIA pub 11). They have also been confounded by concerns about the choice of sham stimulation type [12] and variation in stimulation 'dose' provided. The studies with the most robust clinical effects have used a higher number of pulses per subject and higher stimulation intensity levels, suggesting a dose - response relationship (CIA pub 2,11). They have almost exclusively focused on the treatment of patients with unipolar major depression although some of these studies have included a subset of patients with BPAD-DP.

There have been at least 6 meta-analyses of the antidepressant effects of left PFC rTMS. All but one have shown greater antidepressant effects of 2 weeks of HFL-TMS compared to sham; these included an analysis of 6 reports [13], of 12 studies [14], of 16 studies [15], of 10 studies [16] and a Cochrane review of 14 [17]. The single negative study included only 6 reports with 91 subjects and as such had less power than most of the other meta-analyses. Of the larger studies, Holtzheimer et al reported a weighted mean effect size of 0.81 [14], Burt et al reported an effect size of 0.67 [15] and Kozel at al an effect size of 0.53 [16]. These are moderate to large effects. However, the studies repeatedly report that the overall clinical results observed were not that impressive. For example, a mean overall improvement of only 23.8% on the HAMD in the blinded studies reviewed in [14] compared to 7.3% in the sham group.

Attempting to address the issue of definitely establishing the efficacy of rTMS in depression, a large international trial has been recently completed. Conducted by a private TMS manufacturer (Neuronetics Pty Ltd), this involved a randomized trial of HFL-TMS (10Hz) compared to placebo over up to 9 weeks of treatment. The results of this study show the statistically superiority of active over sham rTMS treatment [18]. Currently, the results of this trial are being utilized in support of an FDA application for the Neuronetics device for treatment of depression (M Demitrack, personal communication).

In addition to sham controlled trials, there have also been a series of trials comparing HFL-TMS to ECT [19-22]. Encouragingly, these have reported no differences in responses rates between ECT and HFL-TMS except for one study including psychotic patients (showing greater benefit with ECT in the psychotic group [22]) and a second trial which recently reported better responses with bilateral or unilateral ECT of unlimited treatment duration compared to a fixed course of 3 weeks of unilateral left PFC rTMS [23]. However, some of these studies have not necessarily had sufficient power to detect subtle differences between the groups.

More recently, several alternative rTMS paradigms have shown promise. Low frequency rTMS applied to right PFC (LFR-TMS), contrary to high frequency rTMS, is known to reduce local cortical excitability [24]. The initial trial of LFR-TMS established its efficacy in a non treatment resistant sample of patients [25]. Funded by a previous NHMRC grant, we have conducted research establishing the efficacy of LFR-TMS as compared to HFL-TMS and placebo (CIA 15):

Previous Study 1:

In this study, 60 patients were randomized to either HFL-TMS or LFR-TMS or a sham stimulation condition (n=20 in each group). All had treatment resistant depression (TRD) and had failed multiple antidepressant medication trials (mean number = 5.7 ± 3.4). There were no baseline clinical or demographic differences between the three groups. Over the double blind phase of the study there was clearly an antidepressant effect of both active groups that was superior to the response to sham stimulation. There was a continued improvement in both active groups across the 4 weeks of the study: for the group as whole, after the 4 weeks of treatment the mean percentage change in MADRS score from baseline was 48.0 ± 17.9% (range 15.1 - 87.5%). These results demonstrated that both HFL-TMS and LFR-TMS have substantial therapeutic efficacy and that robust clinical response appeared to require at least 20 sessions of treatment with the parameters used. However, only a sub-population of patients responded to treatment.

The equivalence of HFL-TMS and LFR-TMS has also been demonstrated in a more recent study [26].

Previous Study 2:

We have also conducted a large scale clinical effectiveness trial of LFR-TMS comparing 1 Hz and 2 Hz stimulation to right PFC (CIA 47). This study involved the randomization of 130 patients with treatment resistant major depression to receive right PFC stimulation in a single daily 15 minute train at either 1 or 2 Hz. 2 Hz stimulation was investigated as a potential way of doubling the number of applied stimuli within identical treatment time (to increase the treatment dose and potentially enhance treatment response). Treatment response was equivalent between the 2 groups (62.3 24.0 % change in the 1 Hz group, 64.7 21.0% change in the 2 Hz group over 4 weeks) with no difference in response. This does not support the substitution of 2 Hz as an optimal condition for LFR-TMS.

Another alternative paradigm is sequential bilateral rTMS (SBrTMS), the application of HFL-TMS and LFR-TMS, one after the other in the same treatment sessions. We have recently completed the first large study of this technique and the first to extend treatment beyond 10 days (CIA 33):

Previous Study 3. In the study conducted, we combined HFL-TMS and LFR-TMS sequentially in each treatment session in a trial comparing active to sham stimulation (n=50, 25 per group). Importantly, treatment was provided for up to 6 weeks, longer than any previously published rTMS trial. In this study there was clearly a marked benefit of active over sham stimulation. There was a significant difference between the groups at 2 weeks (F(1,25)=25.5, p<0.001) which remained significant at all other trial time points (F(5,44)=3.9, p=0.005). Patients continued to respond across the 6 weeks of active treatment.

Most critically, by study end, 13 of 25 patients in the active group (>50%) and only 2 in the sham group met response criteria on the HAMD. 9 patients in the active group (36%) and no patients in the sham group met criteria for clinical remission. A further 45% of patients in the sham group who crossed over to active treatment at trial end went on to respond in a manner meeting response criteria (33% remission criteria).

rTMS in Bipolar Disorder - Depressive Phase Markedly fewer trials have been conducted investigating the effects of rTMS in BPAD. Of these, a small number have focused on the treatment of mania using high-frequency stimulation applied to the right DLPFC. Initial open studies of this technique appeared promising, although negative results have also been published [27] and there is need for a substantive definitive trial.

Even fewer studies have focused specifically on the treatment of BPAD-DP. In the first of these, Dolberg et al randomized 20 patients to HFL-TMS or placebo [28]. Unfortunately the brief report published provides few experimental details but the active group appeared to do better than the sham group after 2 weeks of treatment. In contrast, in the only other randomized trial, Nahas et al found no difference in response between active HFL-TMS and sham in a study of 23 patients [29]. In this study, there was a response rate of 36% in the active and 33% in the sham group. The authors questioned whether the type of sham coil they used was really in part 'active' due the degree of brain stimulation produced with the coil orientation they applied. The same group followed 7 patients who responded to active rTMS over 12 months providing weekly maintenance rTMS [30]. There appeared to be some degree of benefit of maintenance treatment although this was difficult to quantify due to the open nature of the follow up.

The only other source of information about the usefulness of rTMS in BPAD-DP comes from the description of the response of these patients in general trials with mixed samples. Unfortunately, some of the larger studies including the Neuronetics trial described above [18] have excluded BPAD patients and most of the studies other than our own have had too small samples to allow meaningful data to be extracted on subgroups. In studies we have conducted, none of which was designed to directly investigate rTMS responses in BPAD, a total of 50 BPAD-DP have been treated to date. These have been treated across 5 separate trials. In most of these trials the number of subjects is too limited to draw any useful inferences. However, a total of 21 patients with BPAD-DP received LFR-TMS in the clinical effectiveness trial described above over a 4 week period (study 2). In this group, there was a significant reduction in Hamilton Depression Rating Scale (HAMD) scores between baseline and week 2 (baseline = 21.6 ±5.7, week 2 = 12.0±6.3, p<0.001) and between baseline and week 4 (baseline = 21.6 ±5.7, week 4 = 9.7±6.3, p<0.001). Interestingly given that this is the first published sample with a treatment duration of greater than 2 weeks, the group showed a significant reduction in HAMD scores between week 2 and 4 (p<0.05) indicating the benefit of the longer period of treatment. 15 (71%) of these patients met response criteria (>50% reduction in HAMD scores by trial end).

Although this study did not include a sham control group, it strongly suggests that low frequency right sided rTMS has therapeutic potential in BPAD-DP which warrants the conduct of a definitive randomized controlled trial. Despite widespread communication within the international rTMS research community (the CIA of this grant is a committee member of the International Society for Transcranial Magnetic Stimulation), we are aware of no large study exploring the use of rTMS in BPAD-DP that is completed and unpublished or currently underway. In addition, we are not aware of any current or past studies exploring the role of LFR-TMS in this condition.

OTHER GAPS IN THE LITERATURE Clearly BPAD is a relapsing illness and there is no substantive data on the impact of rTMS on longer term outcomes. Long term follow-up data with rTMS treatment is very limited in general. One study has reported similar 6 month relapse rates following rTMS and ECT treatment [31] but only considered the outcome in a limited group of patients (n=21 in the rTMS group). Unpublished data from the Neuronetics trial (CIA 52) suggests that relapse rates post rTMS are similar or lower than those seen in previously published large ECT studies (M Demitrack, personal communication). We have reported relapse and re-treatment at an average time of 9 months post acute treatment but this sample was not comprehensively followed between the end of acute treatment and relapse (CIA 36). We will aim to significantly contribute to understanding the outcome of treatment by including structured assessments over a 12 month follow up period in this study.

Hypotheses / Research Questions

Primary Hypotheses

1. Treatment with high frequency left sided rTMS will result in a greater reduction in HAMD scores than sham rTMS.
2. Treatment with low frequency right sided rTMS will result in a greater reduction in HAMD scores than sham rTMS.

Secondary Hypotheses 3. Greater than 50% of treatment responders to rTMS will continue to experience clinical benefits from rTMS, as indexed by HAMD scores remaining within 25% of the HAMD score achieved at the end of the acute treatment phase during the 6 months of post acute treatment follow up.

4. Treatment with high frequency left sided rTMS and low frequency right sided rTMS will both be well tolerated as assessed by a retention rate in treatment of greater than 90%.

Methodology including project design and sequence of procedures

Experimental Design The study has been designed to allow the reporting of results in a manner consistent with the international CONSORT guidelines.

The study will involve a 4-week (20 session) randomized double-blind clinical trial with 3 treatment arms conducted at a two sites (Alfred Psychiatry Research Centre in Melbourne and the Centre for Addiction and Mental Health, Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada).

Randomization will occur via the generation of a computer number sequence with stratification by site and diagnosis of bipolar I or bipolar II disorder. Subjects will be randomized immediately prior to the commencement of the first treatment session, after the measurement of bilateral resting motor thresholds with standard means (CIA 2).

The main study phase (phase 1) will involve the 4 week randomized controlled trial conducted under strict double-blind conditions. Fidelity of the blinding process will be assessed at the end of this period with patients and raters.

Phase 2 will involve the provision of open label treatment with LFR-TMS to patients who received sham treatment and wish to receive 'active' rTMS.

Responders to active treatment from phase 1 or 2 will enter phase 3, a 6 month follow up period where participants will be reassessed at 2 weeks, 1, 3, and 6 months.

Subjects

Sample Size We aim to recruit 40 patients into each group (total n=120). The sample size calculation for the study was conducted based on the within group standard deviation of 7 (this was the upper limit of the standard deviation for the bipolar patients HAMD scores in the low frequency right sided rTMS trial data provided above) and between group difference of 5.0 points (change in 2 active groups of 12 (based on 4 week change data above) and change of 7 in the sham group allowing for a substantial placebo response). With an alpha of 0.05 the study will have a power of 0.91 to detect a difference in end scores between the 3 groups (calculation in PASS 8.0).

Clinical Measures Demographic variables and potential co-variates will be recorded at baseline following a clinical interview. These will include the duration of the current episode, years from first diagnosis, number of previous episodes, type and dose of current and previous treatment and family history of mood disorder.

Clinical measures will be performed at randomization, at 2 and 4 weeks as well as at the follow up assessments. A trained rater who is blind to treatment type will administer all measures. Raters will be required to maintain >90% reliability on the primary outcome assessments on ratings with 6 monthly assessments based on videotaped interviews.

The primary outcome variable will be scores on the 17-item Hamilton Rating Scale for Depression (HAMD) [33]. In addition we will use the Inventory of Depressive Symptomatology (IDS) - Clinician Version and Patient version [34]. In addition, we will rate patients on the Young Mania Rating Scale (YMRS) to assess for the possibility of emergent manic symptoms [35].

Assessments made at baseline to explore potential predictors of clinical response will include the CORE rating of melancholia [36], the Measure of Parental Styles [37], the Depressive Personality Inventory [38] and the Costello and Comrey trait anxiety measure [39].

To assess for the possibility that rTMS treatment may result in cognitive side effects, we will assess cognition at baseline and the end of 4 weeks of treatment with the following tests. They may also provide evidence of improved executive / frontal functioning produced with the treatment: Personal Memory Questionnaire, Block Design Test, Verbal Paired Associates Recall and Recognition, Visual Paired Associates, Digit Span (WAIS), Simple and Complex Reaction Time, Finger Tapping Test, Verbal Fluency, Trail making A&B.

TMS Treatment TMS will be administered with Medtronic Magpro30 magnetic stimulators using 70mm figure of 8 coils. Prior to the commencement of treatment TMS, single pulse TMS will be used to measure the resting motor thresholds (RMT) for the abductor pollicis brevis (APB) bilaterally in all subjects using standard published methods (CIA 2).

Stimulation parameters rTMS will be administered on a daily basis 5 days per week for all subjects: Left: 10Hz, 110% RMT, 40 Trains, 5 second duration, 25 second inter-train interval (i.e., 2000 pulses) Right: 1 Hz, 110% RMT, 1 Train of 1200 pulses Sham: Sham patients will be successively randomized to sham stimulation on either the left or right at the same stimulation parameters but using a sham coil.

The sessions have been matched for treatment duration. Missed sessions will be 'made up' by an extension of the treatment duration but only one missed session will be allowed per week.

The two active conditions have been matched for approximate treatment duration rather than pulse number in keeping with previous research (CIA 15). Matching for pulse number would result in either a left treatment schedule with significantly fewer pulses than appears effective in other studies or an excessively long right treatment schedule.

Stimulation localization This will follow a slight modification of standard procedures. Firstly, the site for the optimal activation of the abductor pollicis brevis muscle in the contralateral hand will be located whilst stimulating the relevant motor cortical region at supra-threshold intensity. This site will be marked on the scalp. We will then measure 6 cm anteriorly on the scalp surface and mark it with ink. This point will be then used as the site of stimulation. The majority of rTMS studies have measured 5 cm anteriorly. However, this has been shown to result consistently in localization that is posterior to the dorsolateral PFC [40]. The measurement described here should result in more consistent treatment of dorsolateral PFC but still provide considerable stimulation overlap with the established '5 cm' site.

MRI scan Each subject will undergo a 3D sagittally orientated T1 weighted structural MR scan on the 1.5 Tesla MRI scanner at the Alfred or Toronto Western Hospital (128 slices). Analysis of scalp to cortex distance in prefrontal cortex and at the site of stimulation will be conducted using previously published methods [41] by a post doctoral research fellow with considerable experience in structural neuroimaging analysis using the ANALYSE software package. During scanning, the site of motor cortex and the stimulation site will be marked on the scalp with 2 vitamin E capsules. This will also allow the post hoc analysis of our actual site of stimulation in comparison to cortical landmarks of the DLPFC.

Electroencephalography (EEG) EEG before treatment will be used to examine mean absolute spectral power within 5 frequency bands. The mean absolute power (in squared microvolts per hertz) will then be computed within the following 5 frequency bands: delta (0.5-3.5 Hz), theta (3.5-7.5 Hz); alpha (7.5-12.5 Hz); beta 1 (12.5-20.5 Hz), and beta 2 (20.5-32.5 Hz). EEG activation / spectral power will be assessed at rest (eyes open and closed) and during cognitive engagement in several tasks.

Concurrent Treatment

Medications will be continued during rTMS if considered safe to do so. Patients will only enter the study if there has been no dose increase in the 4 weeks prior to randomization and there has been no significant improvement (>15% change) in depression in one week between screening and the first study assessment visit (based on HAMD ratings at these two visits). The commencement of any new psychotropic medication will not be allowed in the 4 weeks prior to study entry and during the study period. The continuation of antidepressant medication / mood stabilizers places the trial in a realistic clinical framework. We will minimize the confounding effect of medication by:

1. Ensuring the patients are stable (stable dose and not improving) prior to entry.
2. Controlling for medication status in secondary statistical analyzes. This way of dealing with medication effects has appeared to quite satisfactory in our previous trials (CIA 15, 33) and not confounded treatment effects.

Statistical Analysis The primary analysis will be conducted on HAMD and secondary analysis on the IDS scores from baseline to week 4 using the 'last observation carried forward' method in the case of subjects who exit the study prematurely. ANOVA tests and 2 tests will be used to examine differences between demographic and clinical characteristics between the groups for continuous and categorical variables following testing for normality of the data. Repeated measures general linear models will be computed to look for differences in group scores across the study time points (baseline, week 2, week 4) with group as the between subjects variable. Concurrent antidepressant use will be included in a secondary analysis as a covariate.

Analysis of the follow up phase will descriptively compare the number of patients who continue to report depressive symptoms below the HAMD threshold for clinical response.

1. Fitzgerald, P.B., Is it time to introduce repetitive transcranial magnetic stimulation into standard clinical practice for the treatment of depressive disorders? Aust N Z J Psychiatry, 2003. 37: 5-11.
2. Fitzgerald, P.B., et al., Transcranial magnetic stimulation in the treatment of depression: a double-blind, placebo-controlled trial. Archives of General Psychiatry, 2003. 60: 1002-8.
3. Fitzgerald, P.B., et al., A Randomized, Controlled Trial of Sequential Bilateral Repetitive Transcranial Magnetic Stimulation for Treatment-Resistant Depression. American Journal of Psychiatry, 2006. 163: 88-94.
4. Loo, C.K., et al., A review of the efficacy of transcranial magnetic stimulation (TMS) treatment for depression, and current and future strategies to optimize efficacy. J Affect Disord, 2005. 88: 255-67.
5. George, M.S., et al., Daily repetitive transcranial magnetic stimulation (rTMS) improves mood in depression. Neuroreport, 1995. 6: 1853-6.
6. George, M.S., et al., Mood improvement following daily left prefrontal repetitive transcranial magnetic stimulation in patients with depression: A placebo-controlled crossover trial. Am J Psychiatry, 1997. 154: 1752-6.
7. Pascual-Leone, A., et al., Rapid-rate transcranial magnetic stimulation of the left dorsolateral prefrontal cortex in drug-resistant depression. Lancet, 1996. 348: 233-7.
8. Lisanby, S.H., et al., Sham TMS: intracerebral measurement of the induced electrical field and the induction of motor-evoked potentials. Biol Psychiatry, 2001. 49: 460-3.
9. McNamara, B., et al., Transcranial magnetic stimulation for depression and other psychiatric disorders. Psychol Med, 2001. 31: 1141-6.
10. Holtzheimer, P.E., 3rd, et al., A meta-analysis of repetitive transcranial magnetic stimulation in the treatment of depression. Psychopharmacological Bulletin, 2001. 35: 149-69.
11. Burt, T., et al., Neuropsychiatric applications of transcranial magnetic stimulation: a meta analysis. International Journal of Neuropsychopharmacology, 2002. 5: 73-103.
12. Kozel, A., et al., Meta-analysis of left prefrontal repetitive transcranial magnetic stimulation (rTMS) to treat depression. Journal of Clinical Practice, 2002. 8: 270-5.
13. Martin, J.L., et al., Repetitive transcranial magnetic stimulation for the treatment of depression. Systematic review and meta-analysis. Br J Psychiatry, 2003. 182: 480-91.
14. Pridmore, S., et al., Comparison of unlimited numbers of rapid transcranial magnetic stimulation (rTMS) and ECT treatment sessions in major depressive episode. Int J Neuropsychopharmacol, 2000. 3: 129-34.
15. Janicak, P.G., et al., Repetitive transcranial magnetic stimulation versus electroconvulsive therapy for major depression: preliminary results of a randomised trial. Biological Psychiatry, 2002. 51: 659-67.
16. Grunhaus, L., et al., Repetitive transcranial magnetic stimulation is as effective as electroconvulsive therapy in the treatment of nondelusional major depressive disorder: an open study. Biological Psychiatry, 2000. 47: 314-24.
17. Grunhaus, L., et al., Effects of transcranial magnetic stimulation on severe depression. Similarities with ECT. Biological Psychiatry, 1998. 43: 76S.
18. Chen, R., et al., Depression of motor cortex excitability by low-frequency transcranial magnetic stimulation. Neurology, 1997. 48: 1398-403.
19. Klein, E., et al., Therapeutic efficiency of right prefrontal slow repetitive transcranial magnetic stimulation in major depression: a double blind controlled trial. Archives of General Psychiatry, 1999. 56: 315-20.
20. Isenberg, K., et al., Low frequency rTMS stimulation of the right frontal cortex is as effective as high frequency rTMS stimulation of the left frontal cortex for antidepressant-free, treatment-resistant depressed patients. Ann Clin Psychiatry, 2005. 17: 153-9.
21. Loo, C.K., et al., Double-blind controlled investigation of bilateral prefrontal transcranial magnetic stimulation for the treatment of resistant major depression. Psychol Med, 2003. 33: 33-40.
22. Cohen, C.I., et al., The efficacy and safety of bilateral rTMS in medication-resistant depression. J Clin Psychiatry, 2003. 64: 613-4.
23. Conca, A., et al., Combining high and low frequencies in rTMS antidepressive treatment: preliminary results. Hum Psychopharmacol, 2002. 17: 353-6.
24. Hausmann, A., et al., No benefit derived from repetitive transcranial magnetic stimulation in depression: a prospective, single centre, randomised, double blind, sham controlled "add on" trial. J Neurol Neurosurg Psychiatry, 2004. 75: 320-2.
25. Rybak, M., et al., An attempt to increase the rate and magnitude of the antidepressant effect of transcranial magentic stimulation. German Journal of Psychiatry, 2005. 8: 59-65.
26. Fitzgerald, P.B., et al., A Naturalistic Study of the Use of Transcranial Magnetic Stimulation in the Treatment of Depressive Relapse. Aust N Z J Psychiatry, 2006. in press.
27. Dannon, P.N., et al., Three and six-month outcome following courses of either ECT or rTMS in a population of severely depressed individuals--preliminary report. Biol Psychiatry, 2002. 51: 687-90.
28. Holtzheimer, P.E., 3rd, et al., Shorter duration of depressive episode may predict response to repetitive transcranial magnetic stimulation. Depress Anxiety, 2004. 19: 24-30.
29. Nahas, Z., et al., Safety and benefits of distance-adjusted prefrontal transcranial magnetic stimulation in depressed patients 55-75 years of age: a pilot study. Depress Anxiety, 2004. 19: 249-56.
30. Fitzgerald, P.B., et al., The application of transcranial magnetic stimulation in psychiatry and neurosciences research. Acta Psychiatr Scand, 2002. 105: 324-40.
31. Fava, M., Diagnosis and definition of treatment-resistant depression. Biol Psychiatry, 2003. 53: 649-59.
32. Khan, A., et al., Severity of depression and response to antidepressants and placebo: an analysis of the Food and Drug Administration database. J Clin Psychopharmacol, 2002. 22: 40-5.
33. Hadzi-Pavlovic, D., et al., Inter-rater reliability of a refined index of melancholia: the CORE system. J Affect Disord, 1993. 27: 155-62.
34. Parker, G., et al., The development of a refined measure of dysfunctional parenting and assessment of its relevance in patients with affective disorders. Psychol Med, 1997. 27: 1193-203.
35. Huprich, S.K., et al., The Depressive Personality Disorder Inventory: an initial examination of its psychometric properties. J Clin Psychol, 1996. 52: 153-9.
36. Costello, C.G., et al., Scales for measuring depression and anxiety. J Psychol, 1967. 66: 303-13.
37. Herwig, U., et al., Transcranial magnetic stimulation in therapy studies: examination of the reliability of "standard" coil positioning by neuronavigation. Biological Psychiatry, 2001. 50: 58-61.
38. Nahas, Z., et al., Brain effects of TMS delivered over prefrontal cortex in depressed adults: role of stimulation frequency and coil-cortex distance. J Neuropsychiatry Clin Neurosci, 2001. 13: 459-70.

Inclusion Criteria:

Patients will be included if they:

1. Have a DSM-IV diagnosis of a bipolar disorder (type I or II) and currently meet criteria for a major depressive episode (SCID 11).
2. Be aged 18-70.
3. Have the persistence of depressive symptoms for at least one month at sufficient severity to warrant a diagnosis of major depressive episode
4. Have a Hamilton Depression Rating Scale Score of > 20 (moderate - severe depression). Including only a more severely ill group of subjects limits the placebo response rate [32]. Moreover, this will allow us to address the application of rTMS methods in the most clinically relevant subgroup of patients (in addition helping to constrain group heterogeneity, a major issue in depression research).
5. Have had no increase or initiation of new antidepressant (or other psychoactive) therapy in the 4 weeks prior to screening.

Exclusion Criteria:

1. Patients who have an unstable medical condition, neurological disorder or any history of a seizure disorder or are currently pregnant or lactating.
2. In the opinion of the investigator, are a sufficient suicidal risk to require immediate electro-convulsive therapy.
3. Have a current DSM IV diagnosis of substance abuse or dependence disorder, a diagnosis of a personality disorder (SCID II) or another axis 1 disorder.

Please note: several of these criteria (e.g. inclusion criteria 1 & 2, exclusion criteria 3) have been selected to explicitly constrain the heterogeneity of the sample to increase the likely power of the study to detect differences between the groups given the potentially subtle difference between the treatment methods.