Financial services workers, medical students and others working in highly competitive environments are using so-called “smart drugs” to enhance cognitive performance.
Do they actually work? Some of these drugs, such as methylphenidate (aka Ritalin) and dextroamphetamine, have been used successfully as part of treatment for attention deficit hyperactivity disorder (ADHD), but less is known about their effect on people who do not have ADHD.
Users may report a subjective feeling of cognitive enhancement, but it has been less clear whether this corresponds to objective improvements in performance.
In a new study, we examined the effect of three common “smart drugs” on cognitive performance, and our results suggest the drugs are not so smart after all. Users expended more cognitive effort and showed more frantic activity, but in general the drugs made their output worse.
Complex problems
Our study looked at methylphenidate, dextroamphetamine and modafinil. The main effect of these drugs is to increase the amount of the neurotransmitter dopamine in the brain. These drugs are known for producing changes in attention, motivation, and wakefulness.
These medications have proved to be a safe and effective part of ADHD treatment. However, previous studies examining the effects of these drugs on specific cognitive tasks in people without ADHD have yielded very mixed results.
Read more: The rise of cognitive enhancers is a mass social experiment
These earlier studies into the cognitive effects of these kinds of stimulant drugs looked at simple tasks, such as memorising numbers or planning simple spatial moves.
However, modern competitive workplaces require very complex and creative integration of many different kinds of cognitive tasks. As researchers who study how people make decisions and solve complex problems, we wanted to know how these drugs might affect more complicated decisions.
The knapsack problem
For our study, we recruited 40 people aged between 18 and 35 who did not have ADHD, and invited them to take part in four testing sessions over four weeks.
At each session the participant would receive either a placebo, methylphenidate, dextroamphetamine or modafinil. The study was double-blinded, so neither the participant nor the researcher knew which drug was being given during a session.
The sessions were also balanced using a “Latin square” design, which varies the sequence of which drugs were given at which session across the whole group of participants.
After receiving the drug (or placebo), participants were given tasks to perform. The main one was a complex optimisation task called the knapsack problem, which is easy to explain but can be much more difficult to solve.
In the task, participants faced a computer game which asked them to imagine they have a bag or knapsack that can hold a certain amount of weight. Next, the game presented ten or 12 different items, each of which had a weight and a dollar value.
The task was to choose items to put in the bag, with the goal of maximising the value of the bag’s contents without going over the weight limit.
Participants were given up to four minutes to try different combinations of items and then submit their selection.
Participants had to complete eight different instances of this kind of problem, at five different levels of difficulty, each presented twice each.
The knapsack problem is an example of optimising a resource (dollar value) under a constraint (weight limit). Problems like this are found everywhere in the real world, like when you do your weekly grocery shopping.
This kind of problem has also been of great interest to computer scientists attempting to develop efficient algorithms to solve them. However, it is not obvious how humans approach these kinds of complex tasks.
More effort, worse results
Overall, after taking the drugs participants took much longer to complete the problems. They spent significantly more time and tried significantly more combinations of items before submitting their selections.
However, when we looked at how close the value of their selections were to the best possible values, we found they did less well. And on average they found the optimal combination less often.
So drug-influenced participants were expending a lot more effort in terms of time and combinations, but their actual productivity suffered significantly. The extra activity did not improve their final performance.
Read more: Mind-bending drugs and devices: can they make us smarter?
Additionally, when we looked at individual performance, we found those who did above average in their placebo session were much more likely to get worse when they took the drugs than people who performed more poorly with placebo.
What does this suggest about taking “smart drugs” to enhance performance?
People without ADHD – especially people who are already high-performing – who take these drugs to try to gain an edge in their workplace or studies may experience unintended consequences. Cognition is a complex thing, and there are no shortcuts to improving it.