We all want to be smarter. For a while, it looked like science was ready to deliver. Transcranial electrical stimulation (TES) is a type of noninvasive brain stimulation that works by placing electrodes over a small area of the brain. Direct current administered through these electrodes can cause the targeted brain area to generate electrical currents, which changes neuronal excitability and leads to alterations in brain function. Originally developed to help stroke patients rehabilitate, studies show that TES, when applied while learning a task, has the bonus of enhancing cognitive function in healthy adults. Depending on the brain area stimulated, TES can boost language and numerical learning, increase attention span, memory and even motor coordination. So is this the “magic pill” that will boost our brainpower to the max?
What’s the catch?
Iuculano T and Kadosh RC. 2013. The Mental Cost of Cognitive Enhancement. Journal of Neuroscience. 33(10):4482-4486.
This is the question the authors astutely asked. Is there a cognitive price we need to pay for cognitive enhancement after TES? To test this, they designed a task to simulate numerical and mathematical learning. They assigned arbitrary symbols (picture Egyptian calligraphy like) to the numbers 1 to 9, and told the subjects to learn what number each symbol represents.
To see how well they’ve learned, subjects had to compare the numerical value of two symbols, and identify the larger one. Their reaction times were recorded as an indicator to how well they’ve learned. As they performed the task, some subjects received TES in the posterior parietal cortex (PPC), a brain area central to numerical understanding; some were zapped in the dorsal lateral prefrontal cortex (DLPFC), which is widely implicated in learning. A control group was mock stimulated in a way that they felt a tingly sensation, but didn’t get the brain-changing effects.
So who did better?
In terms of learning the task, the PPC group learned significantly faster than the control group, while the DLPFC group learned the slowest. By the end of the training blocks, PPC still outperformed both control and DLPFC groups, with DLPFC group taking the longest to get to the right answer. Notably, each group did similarly at the beginning of the training, meaning motivation or attention factors were probably not the cause of the effect.
So far so good! However, the researchers went a step further, to look at how well the subjects could fluently and effortlessly process the new information once it’s learned (“automaticity”). To do this, they administered the infamously difficult “stroop task”. Here’s how it goes: subjects saw two different symbols that represent two numbers. One symbol will physically be larger than the other (in terms of font size). If the physically larger symbol is also larger in its numerical value, then it’s a congruent trial. If the larger symbol actually represents a smaller value, then it’s called an incongruent trial. Subjects were asked to pick the PHYSICALLY smaller symbol regardless of what number they represent, and their response time to both congruent and incongruent trails was measured. Automaticity was measured by subtracting the reaction time of incongruent trails to the reaction time of congruent trails. So if the subjects learned the numerical value of each symbol REALLY well, so well that it’s ingrained to the point they cannot think about symbol without processing the meaning, they would show a MUCH longer processing time for the incongruent trails than the congruent trails. On the other hand, if the subjects could still dissociate the symbol from its numerical value, it would be much easier to process the incongruent trails (since all you need to do is pick the physically larger symbol and not think about what they represent), hence they would have a smaller reaction time difference between congruent and incongruent trails.
(I suggest giving the stroop task a shot yourself: make a bunch of flashcards with two different numbers written on each one, one bigger than the other. Say, write 3 and 8. Make 3 bigger than 8. Now when you see the flashcard, you’ll have to pick the physically larger number – in this case 3, instead of the numerically larger one, 8. As you probably guessed, it takes longer to pick the 3 when it’s written larger than 8).
Here’s the zinger. The PPC stimulated group, which learned the fastest and the best, showed the lowest level of automaticity in the stroop task, while the DLPFC group showed the best. This indicates that while DLPFC group took longer to learn, they could retrieve and process what they learned more automatically in a new situation, resulting in a larger lag between incongruent and congruent tasks. This seems to only apply to the new symbol system, as all groups performed similarly in a stroop test using normal everyday numbers.
Two central aspects of human cognitive expertise involve learning new information and using it automatically in new situations. In this case, the authors found that PPC stimulation increased the rate of numerical/symbol learning, but decreased the learners ability to “automate” – that is, to quickly and effortlessly perform the task without much thought. On the other hand, although DLPFC stimulation decreased the initial learning rate, it increased the learner’s ability to automatically retrieve and use the information later on in a game of mental gymnastics.
In this sense, “cognitive enhancement” by TES does come with a cost – “cognitive impairment” in some other mental faculty. It is difficult to say whether this is true for other types of learning or alternative ways of noninvasive brain stimulation. However, until scientists find out more about the hidden cost of cognitive enhancement, zapping yourself may not be the best way to give your brainpower a boost.
Note: This post was updated to better explain the numerical stroop task in response to No’s comment. I realized I never outlined what “better” meant and could have led to confusion. Thanks No for the heads up!
Iuculano, T., & Cohen Kadosh, R. (2013). The Mental Cost of Cognitive Enhancement Journal of Neuroscience, 33 (10), 4482-4486 DOI: 10.1523/JNEUROSCI.4927-12.2013