Part three of a three part series on nonlinear dynamical systems, self regulation and NeurOptimal neurofeedback. Part one is Why Your Brain Is Not a Thermostat and part two is The Brain on the Edge.
NeurOptimal is a form of dynamical neurofeedback. Rather than setting a target for any particular brainwave frequency, the system monitors the electrical activity at the scalp many times a second and briefly interrupts the audio the person is listening to at moments when that activity shifts abruptly. The working model is that a self organizing dynamical system, given accurate and well timed information about its own transitions, can reorganize without being told where to go. The science of nonlinear brain dynamics that motivates this model is substantial. The research on NeurOptimal specifically is limited, and this article says so plainly.
Two articles ago we started with a thermostat. Then we spent a long time on hearts, ant colonies, neuronal avalanches and the finger wiggling experiment that reorganizes itself at a certain tempo. All of that was setup. This is the part where it either pays off or it does not.
A short recap, for anyone arriving here first
Living things are nonlinear dynamical systems: they are described not by fixed values but by how their state evolves over time, under rules where a small input at the right moment can do more than a large one at the wrong moment. Their health shows up as flexibility, meaning a wide repertoire of available states and easy movement between them. The brain appears to operate in a regime between too much synchrony and too little, where it is unusually sensitive to well timed information. And when people feel stuck, one useful description is that a particular pattern has become a deep attractor with a wide basin, easy to fall into and hard to leave.
If that is what a brain is, the interesting design question for any brain training system is not what value to aim for. It is what information to give the system about itself, and when.
What the equipment actually is
It is worth being concrete, because the mental image people arrive with is usually borrowed from science fiction and involves a helmet.
A NeurOptimal session uses five small sensors and a pair of earbuds. Two sensors sit on the scalp on either side of the head, above the midpoint between the top of the ear and the top of the head. Two clips go on each ear, one at the top and one on the lobe, and they serve as references. The sensors run to a small amplifier, roughly the size of a deck of cards, which connects to a laptop running the software.
The sensors are passive. Nothing is transmitted into the head, no current, no magnetic field, no pulse. The equipment is doing what a microphone does, not what a speaker does. This is the single most useful thing to understand about the setup, and it is the reason the approach is described as non invasive.
Then you sit in a comfortable chair for a session that runs thirty three and a half minutes by default, listening to music or watching a film with the sound on. Many people close their eyes. Some fall asleep, which does not appear to be a problem. There is no task. You are not asked to concentrate, relax, breathe in a particular way, or move a bar up a screen with your mind. Anyone hoping for a video game they control with their brain will find this either a disappointment or a considerable relief.
What the system is watching for
Here is where the dynamical systems framing earns its keep.
Traditional neurofeedback typically asks a question about magnitude. Is there too much slow wave activity at this location, too little of a faster rhythm at that one, relative to a normative database or a clinical assessment? A protocol is chosen, a target is set, and the system rewards movement toward the target. This is a reasonable approach with a substantial research base behind it, and clinics including ours offer assessment informed neurofeedback therapy, including approaches guided by EEG assessment.
NeurOptimal asks a different question, and the question is about change rather than level. The software tracks the electrical signal across many frequency ranges simultaneously, updating many times per second, and watches for moments when the pattern shifts abruptly. Not moments when a band is high or low, but moments when the system departs from what it was just doing.
In the vocabulary of part two, it is watching the trajectory rather than the coordinates. It is looking for the points where the system is leaving one transient state and heading somewhere else. Those moments are interesting precisely because that is when a dynamical system is least committed and most movable.
What it does with what it finds
At those moments, the audio stops. Very briefly, on the order of a fraction of a second, and then resumes.
That is it. That is the entire intervention.
There is no reward and no penalty. The interruption is not a buzzer for doing something wrong or a chime for doing something right, because nothing has been designated right or wrong. The clinician has not chosen a protocol based on your presenting concern, because the system does not use one. Two people sitting in the same chair on the same afternoon with entirely different histories receive the same process, and what differs is only their own activity and therefore the timing of their own interruptions.
This is genuinely strange the first time you encounter it, and it is the point at which people who like neuroscience tend to raise an eyebrow. So let us take the eyebrow seriously.
Why a brief interruption might be useful information
The argument rests on three things from the previous articles.
Brains are built to detect change, not steady states. Nearly every sensory system in the body adapts to constancy and responds to difference. You stop feeling your socks within minutes of putting them on. A sound that stops unexpectedly is one of the most reliable ways to capture the orienting response, which is why a fridge switching off makes you look up and a fridge running does not.
A system near a transition is maximally movable. This was the central claim of part two. Near a critical point, small perturbations have their largest effect, and a system in the middle of leaving one state is by definition near a transition. Timing an input to that moment is not a decorative detail; in a nonlinear dynamical system it is most of the effect.
Well timed disturbance carries information. A brief pulse delivered to a system built out of oscillators can reset the phase of a rhythm. Adding small amounts of noise to a nonlinear system can improve rather than degrade its ability to detect weak signals. Neither of these is intuitive, and both are established.
Put together, the working model is this. The interruption arrives at the instant the system is shifting, and functions as a marker of that shift. The brain, which is already in the business of predicting its own sensory stream, registers the discrepancy. Over many such moments across a session, the system receives a great deal of information about its own transitions, and, being self organizing, does something with it.
Notice what the model does not require. It does not require the software to know what a healthy brain looks like. It does not require a diagnosis, or a normative database, or a decision by the clinician about which frequency at which location needs adjusting. The information is entirely about what your own system just did, a moment ago. In dynamical systems terms it is a feedback loop with a very short delay, closed around the system’s own trajectory, which is exactly the arrangement part one identified as the stable and corrective kind.
That is the theory. It is coherent, it is consistent with a large body of well established science about how nonlinear dynamical systems behave, and it is not the same thing as proof that it works.
What the evidence does and does not support
This section matters more than the rest of the article, so we will be direct.
What is well established: that the brain behaves as a nonlinear dynamical system; that its dynamics can be characterized by scale free structure, avalanche statistics and metastable state transitions; that these measures differ, at group level, between clinical and non clinical populations; and that flexibility of state transitions appears to matter for functioning. This is mainstream neuroscience with a large literature behind it.
What has peer reviewed support in the neurofeedback literature more broadly: a framework proposing that neurofeedback can be understood as perturbing the brain’s position within its dynamical regime rather than as training a single frequency band, and a study reporting that a single neurofeedback session was followed by measurable changes in the scale free structure of spontaneous brain activity. These are interesting and directionally supportive. They are also modest in size, and they concern neurofeedback generally rather than any specific commercial system.
What is limited: research on NeurOptimal specifically. The published record consists largely of small studies, case series, uncontrolled reports, and documentation produced by the manufacturer, together with a small number of peer reviewed papers in specific applications such as a 2023 study on tinnitus perception. This body of work is not sufficient to establish efficacy for any condition, and anyone telling you otherwise is overselling it. Sample sizes are small, control conditions are often absent, and neurofeedback research in general has a well known challenge with sham controlled designs, because a credible placebo condition for feedback is genuinely hard to construct.
What we can say from the clinic: clients often report changes in sleep, in how quickly they settle after something stressful, and in general reactivity. We take those reports seriously, and we also hold them lightly, because uncontrolled clinical observation is exactly the kind of evidence that is easiest to be wrong about. NeurOptimal is best understood as a non invasive training system that may support self regulation, used alongside counselling and medical care rather than in place of either. It does not diagnose, and it does not treat conditions.
If you find that framing unsatisfying, that is a reasonable response, and it is the honest state of the field rather than modesty for its own sake.
Why change tends to arrive unevenly
People often expect the trajectory of improvement to look like a ramp. In nonlinear dynamical systems it rarely does.
Change in these systems is usually flat, then abrupt. A control parameter shifts gradually while behaviour stays put, because the existing pattern is still stable, and then at some threshold the pattern loses stability and the system reorganizes quickly. Which means a stretch of apparently nothing is not evidence that nothing is happening. It is what the approach to a threshold looks like from the inside.
Two related ideas from the dynamical systems literature are worth having.
Hysteresis is the property whereby the path out of a state is not the mirror image of the path in. Systems that have been somewhere a long time often need more to leave than it took to arrive. This is not a discouraging fact so much as a fair one, and it applies equally to sleep patterns, chronic pain, relationship dynamics and startle thresholds. A pattern that took four years to establish is not being unreasonable by taking longer than a fortnight to loosen.
Critical slowing down is a phenomenon observed across many complex systems, including ecosystems, climate records and, in at least one study, day to day mood data. As a system approaches a transition, it recovers more and more slowly from small disturbances, and its fluctuations become larger and more correlated. It is used as an early warning signal for transitions in both directions. The clinically useful version of this is that how long it takes you to come back down is often a more informative measure than how high you went.
It is also common for the first changes to show up somewhere other than where you were looking. Sleep frequently shifts before mood. People sometimes notice that a particular relative has become less irritating before they notice anything about themselves, which is a diplomatic way for a nervous system to report progress.
How to tell whether something is shifting
Given all of the above, the useful things to track are dynamical rather than static. Not how you feel today, which is noisy, but the shape of how you move.
Recovery time is the best single measure. Something stressful happens; how long until you are back? Sleep is worth watching in terms of onset and continuity rather than total hours. Reactivity is worth noticing at the threshold: what does it now take to set off the response that used to be set off by very little? And range matters. Can you get properly activated when it is warranted, and also come back? A system that has only become calmer has not necessarily become more flexible, and flexibility is the actual goal.
A very short weekly note on those four things will tell you more than trying to evaluate each session on its own. Sessions are a poor unit of analysis for a system that reorganizes on its own schedule.
Neurofeedback alongside counselling
From a systems view, the various things that help are different kinds of information delivered to the same system.
Counselling supplies relational and reflective information, and it works with meaning, which nothing else on the list does. It is also where the actual difficulty gets named, which matters, because a nervous system that has become more flexible is not the same as a life that has become more workable. Neurofeedback supplies information about the system’s own electrical dynamics on a timescale far below conscious awareness. Sleep, movement, breathing and connection change the parameters of the system itself.
They are not substitutes for one another, and there is no requirement to do all of them. Plenty of people do very well with counselling alone. Some people find the idea of talking exhausting and would rather start somewhere that asks less of them, and starting there is a legitimate choice rather than an avoidance of the real work.
How Sojourn can help
NeurOptimal neurofeedback sessions at Sojourn run in our Cloverdale office in Surrey, and several members of the team provide them alongside their counselling work. Because the process is the same regardless of what brought you in, there is no assessment requirement before starting, though we are happy to talk through whether it makes sense for you.
For people who want to train more frequently, or who live too far away for regular visits, we also offer home neurofeedback rentals, where the system goes home with you for a set period and can be used by more than one person in the household. It is the same system used in the office.
If you are still weighing whether any of this is for you, we have written a plainer piece on whether neurofeedback is a good fit for you, and a broader overview of our neurofeedback services.
And if what you actually need is to talk to someone, that is a perfectly good conclusion to reach at the end of three articles about attractors.
Glossary additions for this article
Dynamical neurofeedback: an approach that provides feedback based on moment to moment changes in brain activity rather than training activity toward a predetermined target.
Hysteresis: the property whereby the conditions required to leave a state differ from those that produced it, so systems can be slow to reverse out of long standing patterns.
Critical slowing down: the tendency of a system near a transition to recover more slowly from disturbances, used as an early warning signal for change in complex systems.
Orienting response: the automatic shift of attention toward an unexpected change in the environment.
Frequently asked questions
What are the pauses in the music actually doing?
They mark moments when the software has detected an abrupt shift in the electrical activity at the scalp. They are not rewards, corrections or signals that something is wrong, and they are not designed for you to notice or respond to consciously. The working model is that the interruption arrives at a point when the system is already in transition and therefore unusually responsive, and that the accumulated information across a session gives a self organizing system something to work with. Most people stop noticing the pauses within a few minutes.
Do I have to do anything during a session?
No, and that is by design. There is no task to perform, no state to hold, and nothing to concentrate on. You can listen to music, watch a film, let your mind wander or fall asleep. Because the feedback is driven by your own activity rather than by performance against a target, there is nothing to do well or badly.
Is dynamical neurofeedback better than traditional neurofeedback?
They are different approaches, and the honest answer is that the comparative research needed to rank them does not exist. Assessment guided approaches set specific targets and have a longer research history for particular applications. Dynamical approaches set no targets and require no assessment, which makes them simpler to deliver and easier to start. Which is a better fit depends on your goals, your history, whether you want an assessment, and practical questions of cost and scheduling. Anyone claiming one is definitively superior is going beyond the evidence.
How many sessions do people usually do?
It varies a great deal, and there is no strong consensus in the literature. Many people try a block of sessions and reassess, and change tends to appear unevenly rather than accumulating in a straight line, so evaluating over a block is more informative than evaluating session by session. We would rather talk with you about what you are hoping for than quote a number that will be wrong for half the people who read it.
Can neurofeedback replace counselling or medical care?
No. It is a training system that may support self regulation, and it is used alongside therapy and medical care rather than instead of them. It does not diagnose or treat conditions, and it is not appropriate as a response to a crisis. If you are in immediate danger or thinking about harming yourself, please contact emergency services or a crisis line rather than booking a training session.
Where to start
If you have read all three parts of this series, the practical summary is short. Your nervous system is a nonlinear dynamical system whose health lies in flexibility rather than steadiness, it reorganizes in response to accurate and well timed information rather than force, and there are several different ways to give it that information, of which neurofeedback is one and a good conversation is another.
Curious whether counselling, neurofeedback or both might fit what you are dealing with? View therapist profiles, approaches and availability, and book online through JaneApp. If you are not sure where to start, contact us and we can help you find a good fit.
