Neurofeedback Equipment Buyers Guide


With such a bewildering array of neurofeedback equipment out there, it can be difficult to find one that fits your need (or even to know what your needs are).

With everyone touting their own product, objective information is hard to come by. As a result, we get a number of enquiries from budding neurofeedback practitioners looking for advice on purchasing equipment

We are long familiar with the types of products on the market. We have our preferences, but no particular affiliation. Bear in mind that the following guide is only our opinion (and of course opinions vary), however ours come from experience, education, and expensive errors.

If you are looking to get started in neurofeedback, this section will help guide you through the maze and de-mystify the process.

First, let’s look at finding the right neurofeedback system for you:

What is the best Neurofeedback System?

The best neurofeedback system is the one that fits your budget and your experience level.

Your neurofeedback software and hardware are your work tools. Your choice of configuration will determine what you’re able to do with your equipment, so getting the right tools matters.

Let’s start with your software.


Your neurofeedback software is your main tool – the interface between the brainwave data and the feedback. At its core, all neurofeedback software is signal processing software. It allows you to make a series of arguments to control the feedback.

For example, the software lets you say; ‘When this brainwave frequency in this spot increases, play this video’.

Each of these logical arguments is called a ‘metric’. If you set your software to train one frequency band in one spot, is one metric. Two frequency bands in two spots are 4 metrics, and so on.

The various neurofeedback configurations are best described in terms of what they can do; how many metrics they can train, what type of metrics (power, coherence, phase etc.), and what general capabilities they have.

Your training options increase exponentially with the number of sensors you use.


With one sensor, you can train the local area under the sensor, one or two metrics at a time.

For example you could say ‘when this brainwave frequency increases in this spot, activate the feedback’. Decreasing activity is done in reverse. This is called ‘absolute’ training; training the total power (amplitude) of the brainwave.

A single sensor also lets train the ratio between two frequencies. You could say ‘when this brainwave decreases relative to that one, activate the feedback’. As you would guess, this is called ‘ratio’ training.

Prices: $100 to $500 all in.


Two sensors may sound like a small difference from single sensor neurofeedback, but it amounts to a big step up.

With two sensors, you can train 5 or 6 metrics at a time. You can train absolute and ratio (see above) in two locations at once, or train the difference between the two areas. Training the deference between two areas is called ‘coherence’ training.

Coherence training lets you say ‘when these two areas work together more, activate the feedback’. Alternatively, you can encourage the areas to disconnect, and mind their own business.

Prices: $1000 to $2000 all in.


The more sensors you have, the more areas you can train at any one time. With 8 to 10 individual sensors, you will often find yourself training 15 or 20 metrics at a time.

Multiple sensors allow you to train Alpha up over two areas at the rear, while at the same time encouraging fewer Betas in the front. You train more complex coherence and ratios, which naturally expands the possible combinations.

Multi-sensor systems often have built-in biofeedback capability, allowing you to integrate heart rate variability, galvanic skin response, or peripheral blood flow to your training.

Prices: $2500 to $5000 all in.


Nineteen sensors gives you coverage over the entire scalp, which vastly increases your training options.

Along with being able to train power, coherence, and ratio in any (or all) areas of the brain, it also makes possible phase training (the speed at which different areas connect to each other).

Using a 19 sensor system, you might train as many as 150 metrics at a time.

Prices: $10,000 to $20,000 all in.


With 19 sensors, you can move below the surface. LoRETA neurofeedback (short for Low Resolution electromagnetic Tomography) gives you a 3D image of brainwave activity, allowing you to train brain networks and deep brain structures in a wide variety of ways.

Using LoRETA software, you can say ‘when the brain’s entire default network is operating at its best, in absolute power, coherence, and phase, activate the feedback’.

Using this method, you will generally train between 50 and 1000 metrics at a time. It’s an investment that will impact your session prices, but it sure makes a difference.

Prices: $17,000 to $25,000 all in.


There are a few things to avoid when choosing your software; the major considerations are compatibility and flexibility.

If your software is compatible with several different amplifiers, you will have greater flexibility to upgrade or change later on. You don’t want to end up stuck with one type of amplifier just because your software is incompatible with any other type, and vice-versa.

Bearing in mind that much of the neurofeedback software out there is built by engineers rather than therapists, it’s important to get an interface that is flexible enough to suit any training requirement. If you want to do a certain type of training and your software can’t do it, the training is not at its best.

Package systems (pre-configured or trademarked neurofeedback packages) are notorious for inflexibility and incompatibility. Restricting your options is one of the ways that they lock you into their way of doing things.

If you are shopping around and talking to sales teams, keep a sharp eye out for any and all sales double-speak. Whatever their claims, there is nothing new under the sun. If the product is relatively expensive, is ‘uniuqe’, and the team are evasive about exactly what it does and how it does it (trade secret, patented method etc.) then look elsewhere. It’s neuro-hooey.


Most of the long-termers in the field use very fancy equipment – better tools improve the quality of their work. However, if you start out with this level of equipment you never learn the basics.

I suggest starting with an inexpensive two sensor system ($1000 to $2000), using open market software that is universally compatible. It will allow you to learn the basics and give you a better idea as to where you want to go from there.

When upgrading, be aware that the transition from individual sensors to 19 sensor / LoRETA training is seldom smooth. Investing 10 to 15 times the amount on equipment for a third the number of sessions per client takes a strong business base, and careful planning.

Once you have chosen your basic approach, it is time to move on to Brain mapping and session planning

Brain Mapping and Session Planning

When a client comes to you with an issue that they’d like to resolve, you need to determine what type of training will be the most effective for them.

There are several methods to do this. It can be done by symptoms alone, by Mini-Q, or using a QEEG.


This method is prescriptive; the protocols are pre-designed to be the most common solution for any given complaint.

The upside is that it takes almost no training to do. The downside is that is may not be the area of concern; obviously, it’s better to check and see if that is the actual problem area or not.

For example; a client walks into your office to improve his concentration. On a brain level, there can be many different reasons for the symptom of ‘poor concentration’. Perhaps it’s just dreamy distraction, perhaps it’s injury, perhaps anxiety to a level that few brain resources are left to focus with.

If you don’t look at the brain activity you won’t know which, hence symptom based protocols are not the best way to go.

Outside of a few pre-packaged systems and clinical studies (where standardisation is a requirement), symptom based protocol selection is rarely used.


A mini-Q is a way to get an idea of overall brain activity by moving one or two sensors from spot to spot on the scalp, and taking recordings from each. You compile this data into a map, which you use to design protocols to suit the client’s individual brain patterns.

The upside of mini-Q’s is that it only takes a one or 2 sensor system to do, making it quite inexpensive. The downside is that it gives you is only a fraction the information of a QEEG (below).

Some pre-packaged systems have automated this method, using software to build the maps and algorithms to suggest what training to deliver. While this kind of software assistance is helpful, it’s no substitute for the eye of a skilled neurofeedback therapist. Computers can’t make intuitive judgments, humans can.

Learning how to interpret your own brain maps has many benefits, and there is no better way to learn it than by using mini-Q’s. It generally takes six months to a year to get au-fait with them.


A QEEG is a quantitative measure (or ‘Q’) of the EEG data.

The measure that the ‘Q’ uses is called a Z-score. When you go to get your thyroid checked, they compare your hormone levels against a norm. This normative level is a Z-score.

If your client has a particular problem to solve, Z-scores are a great way to pinpoint the issue and precisely target only the problem areas. A QEEG will tell you what is happening, where it;s happening, and information on a whole range of factors.

A QEEG gives you enough information to design a fully comprehensive protocol that specifically suits the goals your client. It is also a substantially higher investment, requiring full 19 sensor equipment and professional training.

As a substitute for the training, many clinicians simply send the QEEG information to a lab, who design the protocols for them. The downside of having your training protocols designed by someone else is that it effectively makes you into a technician, rather than a therapist in your own right.

It’s a huge advantage if you can collect, process, and interpret your own brain maps.

Just to complicate things a little more, there are a few different approaches to brain mapping.


The most common type is a resting state brain map. This is when the brainwave recording is taken while the client is in a still, neutral state. This gives you a picture of the brain’s default state at rest, and a good idea of their emotional propensities or difficulties.

Ultimately, it’s these are the underlying emotional moods that you want to shift, so QEEG’s are a great way to target them accurately.

Event related potential (ERP) maps work a little differently. With ERP’s, you’re looking at what the brain is doing while at task, or under load. A series of tasks are given, recordings are taken, and a protocol is made. The trouble here is that when the neurofeedback training is delivered, the brain is not under the same load, so may not give you an accurate training target.

Hence, ERP’s are often used in research and diagnostics is rare in clinical practice.


Most mapping techniques concentrate on the amplitude of each of the frequencies in each area, and their relationship to each other. That not the only type.

Some brain mapping techniques use relational information instead – working to balance the overall activity left to right, front to back, lobe to lobe. It is an older approach, but is still sometimes used with mini-Q’s.

Both have their place, we like a combination of the two.

Once the training plan is made, it is time to run your protocol.

Running your protocol

Now that you have collected your brain map data, you are ready to build your neurofeedback protocol.

A protocol is the actual training delivered to the client; a combination of the sensor location(s) and the brain activity trained at there. Often, you will run several different protocols in one session, each targeting different areas.

Creating the protocols is your primary role as a therapist; this is where the art and skill lies. It takes training, practice, and is well beyond the scope of this piece.

What I can do is give you an overview of the choices you have, and what they mean to your sessions.


The threshold is what activates the feedback, and tells the client when he is approaching the goal and when not. An example would be ‘when the Alpha brainwaves rise above this level, activate the feedback’. This level is the threshold.

There are two ways of setting this threshold.

The first is manually. You can watch the data, and manually set a bar to reach. Above that level, the feedback starts, below that level it stops. When the client improves, you make it a bit harder. When they tire, you make it a bit easier. You’re engaged with the training.

Automatic thresholds work a little differently. The software makes the difficulty level adjustments automatically. Auto-thresholds are a huge labour and skill saver, as you start the protocol and go hands-off from there. The main downside is that the skill level is constantly changing, so the client has no clear target to reach.

So, be sure your software has the option of manual thresholding.


There is a lot of debate around whether to use absolute training or Z-score training.

The difference comes down to this; an example of absolute is ‘reduce the activity in this area from what it is now’. With a Z-score, you say ‘reduce the activity in this area down to this level, and no more than that’.

The argument can be made that with Z-scores, you’re training towards an ‘average’ brain. And who wants to be average? On the other hand, getting back to a ‘comfortable’ position is exactly what you want if you are looking to resolve a particular symptom or get out of a mental jam.

Z-score training can be used with any number of sensors, and is almost a requisite for more complex protocols. By setting training limits, Z-score training eliminates the risk of over-training (training too far).

Not all systems are compatible with Z-score databases. It’s nice to have the option, especially if you’re working with more serious cases.


Standard Feedback
The feedback is the signal for your client, telling him when he is reaching his objective and when not.

You could visual feedback (like something happening on a screen), a game (for example a car that drives when the client reaches the target), a sound (tone, music, a change in volume), or even a physical vibration (like a sonic cushion or vibrating teddy bear). Any signal will do.

Most software platforms have a good selection of feedback options available, and there are some excellent third-party games packs available. This is another areas where software compatibility is important.

Reflection Feedback
Reflection neurofeedback is ‘non-directional’; it doesn’t give any particular ‘goal’ as standard feedback does. It ‘reflects’ or ‘mirrors’ the activity in a given part of the brain, and relies on the brain’s ability to decode that signal and adjust its activity accordingly.

For example; when certain activity in one area is high, a high note is sounded. When it is low, a low note is played, with a note scale between the two. Visual or tactile reflection feedback methods work along similar lines.

This is a favorite among package systems, who brand it under all sorts of names (‘quantum principles’ is a common description). It requires minimal training, so is good for beginners, and puts some neuro-age mystery into the sales jig.

It’s also used with ILF or infra-low neurofeedback (training at very low frequencies <.1Hz). One wave every 10 seconds is too sluggish for standard neurofeedback, though training such slow waves is pretty sketchy using refection neurofeedback as well.

Compared to standard neurofeedback, the efficacy rates of the reflection method are lower, and the results unpredictable. If your client has a particular goal for the training, reflection feedback is not the method to use.

Next, you’ll need some neurofeedback hardware.

Your EEG Hardware

Now that you have chosen your configuration and method, you’ll need hardware to suit. You’re only as good as your tools, and getting it right the first time saves expense down the line.

Several components are needed – sensors to detect the brainwave signals, an amplifier to convert it into digital data, and a computer to process it.


Your training is only as good as your detection equipment, so sensors matter.

There are several different types of sensors available.

Sintered Electrodes 
The most common electrodes (sensors) are either sintered Silver-Silver Chloride (Ag/AgCl) or Gold. Individual sensors attach to the scalp using a conductive paste (EEG paste), or a full sensor cap connects to the scalp using a conductive gel (Electrogel).

They make an excellent contact with the scalp, have low environmental interference, and give a strong, reliable signal. That is why they are the clinical standard. The downside is the ‘goo-factor’ – the residual gel or paste left in the client’s hair.

Silver Chloride sensors generally sell for $10 each, or a full 19 sensor cap sells for about for $350.

Wet Sensors
Wet sensors use a pressure-pad soaked in a saline solution to make a contact against the scalp.

They are cheap as chips and leave no mess, but are highly sensitive to interference, making it difficult (usually impossible) to get a reading that is good enough to use.

Dry Sensors
Dry sensor technology is developing quickly, but it’s not there yet.

Comfort is a concern, but practically there are two main barriers; interference, and price. Their sensitivity to interference makes them rather ‘jumpy’ compared to gel sensors, but most of all the prices run between $20,000 and $40,000. It’s new technology, so nobody knows their life expectancy.

Given the impact on the session price, I would suggest waiting until prices drop.

Now that you have an idea as to what configuration you’d like, it’s time t look for an amplifier to fit your need (and your budget).

A single sensor system with a wet sensor (above) could go for as little as $120. A two sensor amp sells for between $600 and $2000, while a 19ch amplifier will run between $5500 and $10,000.


You want an EEG amplifier with good shielding against environmental noise (to keep your signal clean), one that is durable, and most of all has good customer service in case of a problem. The latter can be the hardest to find.

Also, it’s really nice if it has standard sensor ports, and doesn’t need a special type.

A choice of bitrates (bps, or bits per second) is handy. The minimum clinical spec is 250bps, though some can reach up to 8000bps. Counter-intuitively, the faster your bitrate, the better resolution you have on slow wave signals. The only reason to have bitrates above 2048bps is for infra-low training.

Another thing to check is whether it uses sequential or simultaneous sampling; whether it takes its samples one sensor at a time, or all together in one stream. For example, it might take a bit of data from sensor 1, then sensor 2, and so on, or it may collect all at once. Most therapists in-the=know prefer simultaneous sampling, though it does cost a little more.

A good precaution is to get an amplifier that is compatible with many software platforms. Some amplifiers only work with the software provided with it, so if it doesn’t fit you need or you need to upgrade, you will need to buy another amp. It’s an expensive error.


Once you decide on your software, you will know the required specs for your computer.

It is always a good idea to get a computer that’s a bit faster than the minimum requirement. If you are spending thousands on your hardware, don’t skimp on the last bit.

A computer with an i7 quad core processor, 8GB of ram and at least 2GB of video memory will be fine for just about any software platform. These run between $1000 and $1500.

If you use a mac, then it’s time to buy a PC.

Now that you have chosen your methods, software, and hardware, it’s time to get some training.

Neurofeedback Courses

Once you have decided which software, methods, and equipment suits your need, you will need to find traiing in that method.

There are a number of great neurofeedback courses out there, in whatever method you want to learn. A variety of courses are regularly held in the States, Netherlands, and Germany. A beginners course usually runs about $1200, plus travel expenses.

Trademark neurofeedback systems include basic training as part of the package (a few days to a couple of weeks). They teach you enough to place the sensors and operate the software, but never enough to become independent. It’s an ‘affiliate’ business model under a different name.

Often, your neurofeedback equipment manufacturer or software provider run training courses. This is usually a great way to get started, and you can branch out from there. Taking a variety of courses gives you a wide skill base.

A teacher that gives ongoing support and mentoring is important. Like any profession, don’t count on being completely independent for the first three to five years. All in all, budget in a good $15k to $30k over the term.

At risk of raining on the parade, there is one final thing to consider. Neurofeedback is a poorly recognised, expensive, time consuming, and education-intensive alternative therapy. It’s not easy – the drop-out rate for new practitioners in their first 18 months is 80%.

My suggestion is to first dip your toe in, get a feel for it, and see if it’s for you. Go for some sessions. Experience it. Find what methods you are drawn to. Learn and grow with it.

You’ll be glad you did.

Frequently Asked Questions