One of the most popular methods in neuroscience and cognitive psychology is to correlate various behaviours, either normal or abnormal, with different patterns of brain activity. An implicit hope when it comes to abnormal behaviour is that, if we are capable of isolating the brain region which shows a dysfunction in relation to the unwanted behaviour, we can target it with various medications and therapies. Furthermore, once we manage to restore this brain activity to normal levels, behaviour will follow suit. One technique in particular takes this assumption a step further, suggesting that if we can somehow “see” our brain activity, we can voluntarily bring it under control, much like when we stop ourselves from hyperventilating and calm ourselves down by consciously focusing on taking deep breaths. The name of that technique is neurofeedback. Although first proposed in the 1960s, neurofeedback remains relatively controversial today.
However, before we jump into the evidence pro and against neurofeedback treatments, let’s have a closer look at the practical side: what exactly is neurofeedback?
The principle of neurofeedback is quite simple: first, brain activity is recorded from the participant’s brain, usually using electroencephalography (EEG) or functional magnetic resonance imaging (fMRI). Then this activity is processed by a computer so that it can be shown to the participant in a readable, understandable manner (e.g., as a bar that goes up and down depending on how high the measured activity is). Finally, the participant is instructed to focus on changing their brain activity by, for example, lowering the bar to a certain level. While neurofeedback could, in principle, use any equipment capable of recording brain activity, the most popular one is EEG. This is because EEG is relatively inexpensive (compared, for example, with fMRI), as well as quite easy to use. Therefore, for the remaining of the article, unless stated otherwise, we will direct our attention towards EEG-based neurofeedback (EEGnf).
There are about a million things to consider when trying to at least half-decently assess whether neurofeedback works and after having miserably failed in my attempt to summarize all of them in a neatly fashion here, I have decided the best approach is to dump all of them in a long list and then pick some general themes to discuss at length. So here are some of the things one should consider:
- What protocol was used? (i.e., what activity did participants try to control: a specific frequency of their oscillations from one region or from the entire brain, the synchronization between different regions, etc.)
- How many electrodes were used?
- How many are enough? Two, ten, a hundred?
- How were the electrodes placed?
- What group of people did the experimenters include?
- Did they measure changes in behaviour, in brain activity, or both?
- Did they check whether these changes were long-lasting?
- How did they define long-lasting? A week? A month? Ten years?
- Did they measure whether other brain changes happened as well? What about other behavioural changes?
- And speaking of that, how strong is the evidence linking one particular behaviour to one particular brain state?
- What is a brain state?
- Can all people really learn to control their brain activity? Why/why not?
- If not, how can you tell who will benefit from this therapy before putting them through many long and quite expensive sessions?
- Is neurofeedback better than placebo?
- Which placebo was used, by the way? (For example, it might be that true neurofeedback is not better than placebo neurofeedback, but it is better than placebo medication.)
- And what’s the exact brain mechanism through which neurofeedback works?
- And does it work the same for all disorders we want to treat?
- What disorders can we even treat with it?
- And what variables do we have to adjust when switching from one disorder to the other?
As you can see, it’s a lot. And while in this article we cannot address every single point individually, this cluster of questions without clear answers suggests two important things:
- the methodology across studies is inconsistent;
- we are missing randomized controlled trials.
Let’s begin by addressing the lack of consistency across studies. Basically, each study claims to check if neurofeedback is an effective therapy. However, some are looking just for proof of concept, i.e. they study healthy volunteers to see if EEGnf produces any measurable effect. Others take it a step further and look at a specific disorder. Then, as mentioned above, there are many details which one could alter (what measure of brain activity they derive from the raw EEG signal, from what region, from how many electrodes, how many training sessions they have, what are the demographic characteristics of the participants etc.). And even though these details might sound trivial, they could actually change the outcome of the study. Think about it like this: we could say that people have to eat to survive. This is a perfectly ok statement, but there are so many details we have left out. In fact, if an alien without the concept of “eating food” planned a roadtrip and you told them “make sure you have something to eat for the humans”, they could easily harm their human friends by giving them some random thing to eat. For example, what can humans eat? Grapes? Cereal? Bleach? How often and how much do they have to eat? Is it enough to feed them only cereal once a month? And so on.
But coming back to neurofeedback, what does this lack of consistency mean? Does it work or not? And do scientists even know what they’re doing? Cause it kind of sounds like they’re making it up as they go. To put it shortly, it means that we can’t say for sure yet whether neurofeedback works, but we do have some promising evidence which allows scientists to focus on more specific directions. In other words, it will probably work well in the future, but we do have to tinker with it some more. And while the process may seem a bit random, this is how science actually works. It would be impossible for any single scientist to plan an experiment which included all potential open questions regarding neurofeedback, let alone run it. True scientific progress takes time because it requires many scientists to tackle a problem from many different angles until their findings pooled together converge on a solution.
Randomized controlled trials
While most of the inconsistencies in methodology could iron themselves out over time, there is one point which scientists have to address head-on. I’m talking about double-blinded randomized controlled trials or RCTs. Now you might be thinking: huh? RCTs are scientific experiments which try to reduce as many biases as possible when testing the effectiveness of a treatment. They do so by randomly allocating participants either into the treatment or the placebo/no treatment group. Furthermore, such RCTs can be double-blinded, which means that neither the experimenter, nor the participant know to which group they belong until after the experiment is over. On the participant’s side, this is necessary in order to eliminate placebo/nocebo effects caused by expectations. On the experimenter’s side, this helps prevent them from acting in a way that might reveal to the participant from which group they’re part of.
Double-blinded RCTs are the best way of determining the efficacy of a treatment. If the treatment group shows improvements greater than the placebo group, then we can conclude that this particular treatment with these particular settings works for this particular group of people. As you can imagine, however, RCTs are expensive and time-consuming, and they usually require solid evidence to be funded. Very few proper, double-blinded neurofeedback RCTs have been conducted so far. Additionally, these studies have found sometimes conflicting results. For example, one RCT published in 2016 showed that neurofeedback training did lead to changes in brain activity in children with ADHD, while another one found no behavioural improvements in children with ADHD who underwent neurofeedback training compared to those who received medication. For chronic pain treatment though, one meta-analysis published this year shows that several RCTs have reported pain reduction following neurofeedback treatment.
Again, it looks like neurofeedback has potential as a treatment, but we need more and better studies before we can say that with certainty and start recommending it as an alternative to already-established ones.
What this means for you
Even though researchers still have a lot of work to do, neurofeedback definitely looks promising. What’s more, unlike other more experimental techniques, this one is a lot easier to bring to the consumer market. So it should come as no surprise that there are a lot of “neurofeedback” companies popping out nowadays, and they all promise to offer you neurofeedback therapy either for stress or for a wide range of neuropsychiatric disorders. Now, neurofeedback itself is not harmful. The worst thing that can happen is simply that it doesn’t work, not even as a placebo. But…these therapies are usually quite pricey and not covered by insurance. And of course, a big concern is that people might give up on proven treatment methods in favour of promises on which the neurofeedback technology cannot deliver.
So if you do decide to opt for commercial neurofeedback therapy, here are a couple of things you might want to keep in mind:
- neurofeedback is promising from a research perspective, and it has been shown to be more useful in treating some disorders compared to others, but a lot of evidence is still needed before it becomes a legitimate treatment;
- there should be some form of feedback in neurofeedback therapy. If someone just sticks a little piece of plastic to your head for an hour and that’s it, there’s a good chance that’s not a legitimate neurofeedback provider;
- you might feel comfier purchasing a home neurofeedback toolkit (after all, you don’t have to put on pants and travel), but the chances of neurofeedback actually working are higher when you go to a center where you can interact with a professional who has been specifically trained in this;
- this goes without saying, but don’t just blindly buy into what you see on the Internet. If what someone promises you sounds too good to be true, it probably is.
What did you think about this article? Would you like us to talk even more in-depth about the different aspects of neurofeedback research? Let us know in the comments.
Geladé, K., Janssen, T. W., Bink, M., van Mourik, R., Maras, A., & Oosterlaan, J. (2016). Behavioral effects of neurofeedback compared to stimulants and physical activity in attention-deficit/hyperactivity disorder: a randomized controlled trial. The Journal of clinical psychiatry, 77(10), e1270-e1277.
Janssen, T. W., Bink, M., Geladé, K., van Mourik, R., Maras, A., & Oosterlaan, J. (2016). A randomized controlled trial into the effects of neurofeedback, methylphenidate, and physical activity on EEG power spectra in children with ADHD. Journal of Child Psychology and Psychiatry, 57(5), 633-644.
Omejc, N., Rojc, B., Battaglini, P. P., & Marusic, U. (2019). Review of the therapeutic neurofeedback method using electroencephalography: EEG Neurofeedback. Bosnian journal of basic medical sciences, 19(3), 213.
Patel, K., Sutherland, H., Henshaw, J., Taylor, J. R., Brown, C. A., Casson, A., … & Sivan, M. (2020). Effects of neurofeedback in the management of chronic pain: A systematic review and meta‐analysis of clinical trials. European Journal of Pain.
Rogala, J., Jurewicz, K., Paluch, K., Kublik, E., Cetnarski, R., & Wróbel, A. (2016). The do’s and don’ts of neurofeedback training: A review of the controlled studies using healthy adults. Frontiers in human neuroscience, 10, 301.
You can check out more of our posts here: