It's all a sham

    As scientists we are interested in whether manipulating one variable has an effect on the other. So we try to come up with a paradigm to manipulate what we are interested in and compare the experimental condition with a control condition. After all, how else could we conclude that our results are down to whatever we manipulated? Therefore, choosing the right control condition is as important as choosing an appropriate paradigm. Ideally, the control condition should be exactly the same as the experimental condition, i.e. everything stays the same BUT the aspect you are interested in. Simple, isn’t it?  The problem is that sometimes keeping everything the same but the experimental manipulation isn’t so simple, as the experimental manipulation might come with “side effects” that are not easily achieved in a placebo condition.

 

    In studies involving transcranial direct current stimulation (tDCS), a type of non-invasive brain stimulation inducing low current, the control condition is called sham condition. It’s a condition in which the whole set-up is done as if stimulation was being administered with one difference: it isn’t being administered. In other words, the participants are made believe they are being stimulated by having strapped electrodes to their heads and the experimenter having fiddled with the machine, but no stimulation takes place. The idea is that setting up the stimulation and having electrodes fixed to one’s heads can be a potentially uncomfortable experience that might lead people to behave differently than under normal circumstances. Additionally, to achieve the physical experience of tingling or itching that is often reported especially at the beginning of the stimulation, a fade-in and fade-out period is also included in the sham condition. The fade-in/out is a gradual buildup or reduction of the stimulation at the beginning and end of the session. Fade-in and fade-out usually last for around 15 seconds and such short stimulation is not assumed to significantly affect brain activity, hence, it is commonly included in the control condition.

 

    So, in theory sham stimulation is an ideal control condition because it, similar to actual stimulation, requires participants to go through the setting-up process and assumingly induces a physical sensation comparable to full stimulation. However, it’s not as perfect a control condition as it may seem at the beginning. Kessler et al. (2012) compared the experience of active and sham tDCS in 131 participants. They found statistically significant more severe and more common adverse side effects in people undergoing active brain stimulation (such as tingling, itching, burning, pain or fatigue). Hence, if participants score more or less in a task because of a physical sensation caused by stimulation, this effect is probably higher in people undergoing active stimulation. Arguably, the difference is likely to be smaller than between active stimulation and no tDCS at all, but nevertheless should be kept in mind when interpreting the results.

 

    Another factor to consider when administering sham stimulation is the possibility of participants noticing they are in fact not undergoing stimulation. This could either happen if the experimenter isn’t careful in disguising the condition (e.g., fails to put the machine out of sight so that the participants can see it’s not turned on) or if a participant has taken part in a tDCS experiment before and, for example, notices a difference in sensation. This is even more likely to happen in a within-participants design when people experience every condition and are able to directly compare them. For example, they might notice a reduced tingling sensation or even a lack thereof. Additionally, if they come in several times, the likelihood of them reading up on the method between sessions and being aware of one condition being “pretend stimulation” increases.

 

    As soon as participants realise that they’ve been tricked they might start questioning everything else the experimenter has told them. Are they actually measuring what I’ve been told? Is there a hidden camera? Am I being fooled? Who is this person calling herself a scientist? If that’s the case, the participants’ distrust could potentially pose as much of a confounding variable as a lack of physical sensation or setting-up process. How good of a control condition is sham tDCS? It still requires measuring the participants head, preparing the skin, fixing the electrodes and occasionally fiddling around to reduce the impedance, as otherwise the machine won’t allow you to start the fade-in phase. The question arises if it is worth the effort or if it could possibly even add the variable of distrust. If it isn’t, accepting that tDCS as a method doesn’t offer a proper control condition might be the better choice.

 

    The alternative to sham as a control condition, would either be comparing anodal and cathodal stimulation only, comparing a tDCS condition to a control condition in which participants undergo the task without tDCS set-up or stimulating a brain region you assume is not involved in what you are studying. Depending on your variable of interest and how clear your predictions are on how cathodal or anodal stimulation would influence brain functioning, one would be more suitable than the other. However, if there is no clear prediction on what type of stimulation would enhance or reduce task performance, there is no brain region that can confidently assumed not to be involved and task performance might be affected by physical sensation of active stimulation alone, none of these alternative would be better suited than sham – which brings us back to our starting point. At the end of the day, as so many times in research, the key aspect of choosing an appropriate control condition in a tDCS study probably lies in being aware of the issue and keeping it in mind when interpreting one’s results.

 

 

 

Reference:

Kessler, S. K., Turkeltaub, P. E., Benson, J. G., & Hamilton, R. H. (2012). Differences in the experience of active and sham transcranial direct current stimulation. Brain stimulation, 5(2), 155-162.

How to not do something

Should experiments make sense?