The purpose of neurofeedback

You might have noticed that there are an increasing number of articles showcasing diverse applications of neurofeedback.  As an area of interest within the field of biofeedback it's not exactly new to the world, but of course it often takes decades before something becomes "mainstream".  In the UK we seem to be at last waking up to recognise the uses of neurofeedback following a much wider acceptance in North America and the rest of Europe.

Whilst it might appear that the purpose of neurofeedback is to alleviate a disorder or change some particular brain rhythm or connection, as Collura (2014) points out, at the lowest level, the purpose of neurofeedback is to "produce an artificially constructed reality that will affect the choices made by the brain in relation to it's own self-regulation".

How Neurofeedback works

Typically we like a dual screen setup with a separate trainee and trainer screen with Biotrace software

Neurofeedback works when the brain is presented wth information (by a feedback process) that is related to it's own function in a way that can facilitate beneficial change. The mechanism depends on differentiating or learning to distinguish one state from another - a beneficial state from an undesirable one.

Briefly, neurofeedback starts with the EEG signal (brain's electrical activity) detected at the client's scalp. There a number of ways that this can be captured and these have been explored over the past decades. It's important to distinguish neurofeedback (or EEG biofeedback as it is sometimes called) from conventional EEG and quantitative EEG which have a different purpose even though they use the same fundamental signal. 

The EEG is not there for any physiological reason and it doesn't really reflect the brain's "business" in any direct sense. It's an indirect and yet useful pointer to the underlying brain's activity that we have learned to take advantage of.  In a typical neurofeedback situation

  • EEG is detected at the scalp
  • these signals from the brain are processed and revealed (fed back) to the trainee as a sound, an image, an animation a video or a combination of these things
  • conditioning and change occurs as the trainee is rewarded by the feedback for success in finding the desired state

This sounds a little mysterious but in fact from the trainee point of view the system must be intuitive, relatively simple and engaging.  The trainee needs to be engaged but should not find that training involves a lot of effort. It's more like creating a state of "allowing" - forcing isn't the way.

The EEG signal is typically processed according to its characteristic frequency content via a frequency transform that imagines the source signal broken down into bands of frequency components. These bands are given names such as alpha, beta, theta and so on.


Biotrace software - The green bar signifies the changing score compared with a threshold value

Biotrace software - The green bar signifies the changing score compared with a threshold value

For any neurofeedback application there has to be a protocol. This could be quite simple such as the intention to "reduce the magnitude of theta frequencies at the electrode location".  In current practice it is possible to design quite detailed protocols to address a brain state that will have clinical significance.

However, there is the basic and yet difficult question about what to tell the trainee?

What is the trainee expected to do?

As we hinted at above, neurofeedback can work quite well in the absence of effort on the part of the trainee

The trainee can generally allow training to occur without forcing anything or trying too hard. Neurofeedback is a method of allowing the brain to learn a new state and find it's own way to implement the learning. This is a bit different from other biofeedback applications that require some instruction or coaching to get them to work

Generally Neurofeedback is a more automatic process and the trainer's job is more to encourage the trainee's conscious brain to get out of the way. Having said that the brain must be primed by making the feedback appealing in some way.


If a protocol implies "reduce" or "increase" frequencies in a region it implies that we have to use some form of threshold to sense whether the state is desirable or not. Hence we have the concept of Thesholds which is an important topic and a controversial one because

  • thresholds determine what behaviour (brain activity) is reinforced (rewarded)
  • a threshold is comparing a signal amplitude with some set value which could be constant or it could be varying.

For example, in the Mind Media NeXus system, you can choose to set a fixed threshold, a manual threshold which the trainer will vary during a session or choose to let the software automatically vary the threshold. A threshold value might be the magnitude of a particular component but could be a composite value of a number of factors as suits the situation. 

And - what do you do if the threshold is exceeded? Well thresholds can be used to

  • Reward being over threshold with positive feedback
  • Inhibit - when being over threshold inhibits feedback. In other words success means keeping the value below threshold.

It's beyond the scope of the article to cover the applications of neurofeedback in detail. A good starting point is the book

"Technical Foundations of Neurofeedback" by Thomas F Collura, Published by Routledge 2014