If, unlike me, you don’t spend enormous amounts of time on brain stimulation forums, you might have missed last week’s announcement that transcranial brain stimulation can reduce IQ. I mean you probably didn’t. Because it’s been reported everywhere. But anyway…

Wait a minute, reduces? That’s the opposite of what we’re trying to do!

Well, yeah. But it turns out that there are two really big caveats to this, namely:





It’s not entirely clear what the study actually measured



The most common way to describe the study is that it resulted in a decrease in IQ. That makes sense because they used an IQ test, which you’d expect to measure, well, IQ. Right? Kind of.



The problem is, the IQ test they used isn’t really designed to be taken more than once. It’s especially not designed to be used twice within a week, which is how at least some of the subjects were tested. (all subjects took the test twice, once to get a baseline and once to get a measurement of the effects of tDCS).

The reason this is a problem is not obvious unless you’re familiar with how the WAIS is supposed to work. Most of the subtests depend, to some extent, on the subject being surprised by the content of the questions. The “perceptual reasoning” section, on which tDCS induces the most decrement, for instance, is designed to measure “fluid intelligence” –the capacity to think logically and solve problems in novel situations, independent of acquired knowledge. If the participant has already answered the same questions—as in this study—you’ve changed the rules—created a sort of hybrid test where performance relies not just on the ability to solve new problems, but also to benefit from memories of doing the test the last time.



This is why the scores of both the group that received tDCS and the group that didn’t had an overall increase—it’s not that the participants got smarter, it’s that how the participants did the test (and therefore the cognitive abilities that it measures) changed between sessions.



So what should we make of the fact that the tDCSed participants showed a smaller performance gain? One possibility is that the stimulation actually made them dumber, cancelling out the effects of stimulation. But the other possibility is that tDCS interacted with the myriad “nuisance” factors that caused the performance gain in the first place, perhaps impairing recall for certain types of memories or suppressing a reduction in test anxiety that occurred between sessions in the control subjects. Perhaps the groups simply differed in their ability at baseline to benefit from practice. The bottom line is that to simplify this into a “loss of IQ” ignores a lot of the confounding issues in this type of study.



Their protocol was one that nobody else uses



A lot of people reporting on this study describe how the researchers tested the “most common form of tDCS”. Which is kind of true, but only in the technically true, “it depends on what the meaning of is is” kind of way. In fact, one of the most important sections of the paper is in the discussion, where the authors explicitly point out that their experimental conditions were “not directly comparable to previous studies which administered anodal tDCS to DPFC”.



Why the confusion then? The researchers really did three different experiments: with anodal stimulation of both F3 and F4, at the same time(bilateral condition) and stimulation of just one at a time (unilateral condition). In all cases the cathode was placed at Cz (the very top of the head).



This is very, very different from the montages that have been used in studies of cognitive enhancement in the past (and the most common ones used by the DIY community), which typically use an anode placed near some site on the prefrontal cortex and another either placed on the same region on the other side of the head, above the eye on the other side of the head, or somewhere on the contralateral body below the neck (to generate a montage with only one site with high current density). The authors explain why they wanted to stimulate both sides simultaneously (complex tasks engage large regions of the frontal cortex, therefore they thought stimulating a large area would be desirable. Oddly, they even mention the more conventional way of doing this (an F3-F4 montage), but never explain why they decided against it.





This matters because the location of the electrodes is intimately connected with which regions are stimulated and the overall effects on brain physiology, much like how the chemical structure of a drug is connected with its effects. The montage that they used, while interesting, doesn’t necessarily say much about the effects of montages more commonly used for cognitive enhancement (like F3-Fp2 or F3-right shoulder), much like how the observation that acetaminophen can cause liver damage doesn’t imply the same is true of aspirin.





So why describe this as a “common” method of brain stimulation? Sellers et al. did simulate current flow and demonstrated that their montage gave the peak current flow in the middle and superior frontal gyrus, roughly the area targeted by other studies using tDCS to enhance working memory or other cognitive abilities.



But talking about the peak current flow runs the risk of understating exactly how rough this equivalence is. Saying a montage targets “the middle and superior frontal gyrus” is somewhat akin to giving directions by saying “the restaurant is in California”–you’re talking about a huge brain region,









Above:California. Below:Middle and superior frontal gyri.



At the same time, the inherent low spatial resolution of tDCS makes talking about “targeting” somewhat tricky. All methods of targeting parts of the frontal cortex have off target effects, current flowing through parts of the brain other than the targeted one. Talking solely about the location of the peak current, therefore, neglects the potential effects of the “off-target” stimulation.



Both of these factors mean that two montages that supposedly target the same brain “region” can differ dramatically from each other in what parts of the brain they actually activate. We can demonstrate this using a simple spherical-shell model, (which doesn’t provide details on what specific brain regions are targeted but does show the overall “footprint” of current flow through the cortex) for conditions like the Sellers et al. “bilateral” condition (left) compared to more typical bilateral condition (f3/f4,right)

Spheres 2 reconstruction of the Sellers montage on the left and a F3-F4 montage on the right, looking at the top of the head. Arrows indicate direction of current flow.



Clearly despite technically activating similar brain regions, the way that current actually impinges on the brain is quite different between the two conditions. A similar result can be seen in the Sellers unilateral condition (F3,Cz) versus a more typical unilateral condition (F3/Fp2).





There’s another wrinkle to this not shown in current models, which is the difference between anodal and cathodal stimulation.. Cathodal stimulation inhibits neurons from firing which is often thought to be detrimental to performance, which is why some studies place the cathode in a location where it will have minimal effect on the brain. The cathode placement in this study (approximately over motor, premotor, and sensory regions) almost certainly has some effect, and is virtually unprecedented in studies exploring the effects of tDCS on cognitive function. However, Sellers et al. do point out that the effects of cathodal stimulation on cognitive tasks are less well defined than the effects of anodal stimulation, and may not be always inhibitory (possibly true, but not something I would rely on to design a study) No. And any way you dice it, the parameters used in this study are quite different from what’s used in previous studies and, in most cases, for cognitive enhancement.





So what should we do?

It’s probably obvious at this point that I don’t think this study says much about the cognitive effects of using brain stimulation, at least in the ways that people typically use it. That’s not to say that the study is bad—it’s not, and in fact it raises some very interesting questions about the role of different parts of the cortex in these tasks– but the evidence is very, very, preliminary, and I’m skeptical of how much it translates to the real world, especially given the existence of other studies using more typical protocols which show improvement of specific tasks (i.e.Fregni et al. (2005)) or evidence of cognitive enhancement across severla tasks (i.e. Martin et al (2013))





That said, if you use tDCS there’s also a very practical safety step you should take: tracking yourself! There are all sorts of tools for assessing cognitive function (my favorite is Cambridge Brain Sciences, and Quantified Mind is also good). As a general principle, while formal studies of tDCS montages can be useful for picking protocols, self-tracking is essential to confirm that the stimulation is having the effects that you want, especially when adverse effects are at least theoretically possible.



In a more general sense, there’s a real research need to do exact (as opposed to conceptual) replications of protocols claimed to produce cognitive enhancement or impairment. The purpose of this is twofold: first, it’s important to see whether previously reported effects can be replicated (a pressing issues in the biomedical sciences), and second, it’s important to know the global picture of how these protocols affected multiple cognitive functions. Along with this, there’s a need to develop evidence-based ways for categorizing tDCS protocols—to answer, for instance, what counts as an exact replication (does the current level need to be the same? What about the electrode size? Voltage?). Finally, we must develop better unifying theories of how tDCS affects cognition—for instance, whether enhancement of one function will necessarily cause impairment of another, or whether “trade-offs” are really just properties of a specific task or experiment.





All of these are serious scientific issues, and it; s important to keep up on published research and self tracking. But to these most recent results mean your brain stimulator is making you dumber? Probably not.