Jump to comment:

• Control arms in brain stimulation trials and experiments
  Nick J. Davis
  Published on: 23 October 2014

• Published on: 23 October 2014

  Control arms in brain stimulation trials and experiments

  • Nick J. Davis, Lecturer in Psychology
  • Other Contributors:
    • R. Martyn Bracewell

  The JNNP recently published an interesting study which used magnetic stimulation to the spine of people with Parkinson's disease, and showed a promising improvement in camptocormia compared to a group receiving 'sham' stimulation [1]. In an Editorial Comment, Caslake [2] argued that the participants in this particular study may not have been sufficiently blinded as to the group to which they had been assigned. Although th...

  Show More

  The JNNP recently published an interesting study which used magnetic stimulation to the spine of people with Parkinson's disease, and showed a promising improvement in camptocormia compared to a group receiving 'sham' stimulation [1]. In an Editorial Comment, Caslake [2] argued that the participants in this particular study may not have been sufficiently blinded as to the group to which they had been assigned. Although the stimulation used by Arii and colleagues was delivered to the spine and not to the head (trans-spinal magnetic stimulation, TSMS, rather than transcranial magnetic stimulation, TMS), we believe there are general principles of experimental design that are of relevance to studies using these technologies.

  We have previously argued that it is almost impossible to create a convincing placebo control for experiments using stimulation such as TMS or transcranial current stimulation (tCS) [3]. TMS by its nature is a multisensory experience: driving a current through the hand-held coil creates the intended magnetic pulse, but as a side-effect creates a clicking sound and, depending on the stimulation target, a muscular twitch or an illusory flash of light in the participant's vision. One option for creating a sham condition is to use a "sham coil" which reproduces the clicking sound but does not reproduce the side-effects. If an individual, na?ve participant does not participate in both active and control arms of the study, it can usually be argued that the presence or absence of side-effects does not give the participant enough information to know which treatment they had received. However in the study by Arii et al. [1], and in many other reported in the literature, the participants crossed over so that each person experienced both conditions. This means that each participant was able to compare their second treatment to their first, so reducing the blinding. In the case of the study by Arii et al., we note first that their Table 2 suggests a tell-tale additional improvement in the group that received real stimulation after sham, and second that the authors do not report subjective comments from the participants except that two participants experienced discomfort.

  The poverty of the control arm in stimulation therefore introduces a challenge when designing experiments and trials to use these techniques. One solution, as we mentioned, might be to divide participants into groups who separately receive either real or sham stimulation. Another solution could be to introduce an alternative stimulation target which is hypothesised not to be involved in the behaviour of interest, although we note that this is a difficult and somewhat subjective alternative [3]. At a more philosophical level it could also be argued that so-called 'non- invasive' [4] techniques such as TMS and tCS are inherently imperfect for creating control and blinding conditions in awake and aware participants, and therefore to qualify any reported results by adding participants' comments and other observations so the degree of unblinding can be assessed.

  In conclusion, we agree with the comments of Caslake [2] that proper experimental science requires a proper experimental control, however it is important to acknowledge that proper blinding is very difficult to achieve in many situations. Our use of non-invasive stimulation requires an ethical balance between the need to preserve the scientific integrity of an experiment or trial, and the desire to advance treatments for debilitating conditions. We therefore advocate a culture of open reporting of all relevant circumstances in an experiment.

  Dr Nick Davis, Swansea University, UK Dr Martyn Bracewell, Bangor University, UK.

  References

  1. Arii Y, Sawada Y, Kawamura K, et al. Immediate effect of spinal magnetic stimulation on camptocormia in Parkinson's disease. Journal of Neurology Neurosurgery and Psychiatry 2014;85:1221-26 doi: 10.1136/jnnp- 2014-307651 . 2. Caslake R. Dif?culties with control arms in repetitive magnetic stimulation studies. Journal of Neurology Neurosurgery and Psychiatry 2014;85:1182 doi: 10.1136/jnnp-2014-307906.

  3. Davis N, Gold E, Pascual-Leone A, et al. Challenges of proper placebo control for noninvasive brain stimulation in clinical and experimental applications. European Journal of Neuroscience 2013;38(7):2973-77

  4. Davis N, van Koningsbruggen M. 'Non-invasive' brain stimulation is not non-invasive. Frontiers in Systems Neuroscience 2013;7:76 doi: 10.3389/fnsys.2013.00076.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top