Sir,
In 2011 Jan-Jun issue of your journal, Lingeswaran[1] reported results of randomized double-blind sham-controlled trial of high-frequency repetitive transcranial stimulation (rTMS) over left dorsolateral prefrontal cortex in depression. They had found rTMS response similar to sham stimulation, which they appropriately ascribed to small sample size (n=23) leading to an underpowered study which was not adequate to identify the effect. They also asserted that there were no previous studies of rTMS in depression from India. In fact, there are two previous published reports[2,3] of rTMS in depression from India.
Ray et al.[2] in a randomized sham-controlled trial of 10 sessions of rTMS (using similar parameters) in 45 subjects with depression, conducted at Central Institute of Psychiatry, Ranchi, India, had found significant reduction in SIGH-D scores in those receiving active rTMS as compared with sham stimulation (Pillai's Trace F [1/38]=56.75, P<0.001), with high effect size (η2=0.60). Furthermore, in their study those with psychotic depression, there was a significant reduction in SIGH-D scores (Pillai's Trace F [1/25]=43.04, P<0.001) and BPRS scores (Pillai's Trace F [1/25]=42.17, P<0.001) in those receiving active rTMS as compared with sham stimulation, with high effect sizes. Larger sample size (n=45) as well as higher number of sessions (ten sessions, as compared with six in Lingeswaran[1] study) might explain the positive results in Ray et al.'s[2] study.
In another study conducted at Central Institute of Psychiatry, Ranchi, India, Nongpiur et al.[3] had found add-on, active-priming rTMS (4-8 Hz; 400 pulses, at 90% of motor threshold) followed by low-frequency rTMS (1 Hz; 900 pulses at 110% of motor threshold) over the right dorsolateral prefrontal cortex to be more effective than sham-priming stimulation in 40 patients with moderate-to-severe depression as shown by reduction in SIGH-D scores over time (F=3.53, df=2.12, 74.23, P=0.032, Greenhouse-Geisser corrected) with small effect size (η2=0.092).
Lingeswaran[1] conducted a post hoc power analysis in their study and detected very low power because of smaller sample size. Such post hoc power analysis has been criticized as there is a one-to-one correspondence between the P value and statistical power; thus, any non-significant result will always be associated with lower statistical power.[4] The best option is to calculate sample size a priori, specifically when an estimate of effect size is available from previous studies.
REFERENCES
- 1.Lingeswaran A. Repetitive transcranial magnetic stimulation in the treatment of depression: A randomized double-blind placebo-controlled trial. Indian J Psychol Med. 2011;33:35–44. doi: 10.4103/0253-7176.85393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ray S, Nizamie SH, Akhtar S, Praharaj SK, Mishra BR, Zia-ul-Haq M. Efficacy of adjunctive high frequency repetitive transcranial magnetic stimulation of left prefrontal cortex in depression: A randomized sham controlled study. J Affect Disord. 2011;128:153–9. doi: 10.1016/j.jad.2010.06.027. [DOI] [PubMed] [Google Scholar]
- 3.Nongpiur A, Sinha VK, Praharaj SK, Goyal N. Theta-patterned, frequency-modulated priming stimulation enhances low-frequency, right prefrontal cortex repetitive transcranial magnetic stimulation (rTMS) in depression: A randomized, sham-controlled study. J Neuropsychiatry Clin Neurosci. 2011;23:348–57. doi: 10.1176/jnp.23.3.jnp348. [DOI] [PubMed] [Google Scholar]
- 4.Ellis PD. Cambridge: Cambridge University Press; 2010. The essential guide to effect sizes: Statistical power, meta-analysis, and the interpretation of research results. [Google Scholar]