Neurofeedback as ADHD therapy

Neurofeedback is a scientifically recognized treatment technique that provides the patient with immediate feedback as they attempt to change the electrical activity of the brain measured by electrodes.

ADHD is one of the main areas of application for neurofeedback,¹ along with tic disorders, migraines, tinnitus, depression and autism.²

Neurofeedback appears to improve ADHD symptoms about as effectively as cognitive behavioral therapy (⇒ Effectiveness of different forms of treatment for ADHD) and also appears to lead to lasting effects (see below). Due to the lasting effect, one can speak of a healing effect through neurofeedback. However, this healing effect is very rarely complete due to the limited symptom improvement and the fact that not all people with ADHD respond equally. Neurofeedback is usually only useful as a complementary therapy to medication.

Neurofeedback works by improving the self-controllability of your own brain activities (see below)
In the treatment, the patient learns to influence his brain frequencies while these are measured and made visible to him as immediate feedback (e.g. colors; bar graphs; airplane flying higher or lower; film that runs or stops depending on the measured values achieved).

Neurofeedback uses various neurological measurements.

1. EEG frequency bands EEG frequencies are controlled by the thalamus, whose function shows significant deviations in ADHD.³
2. SCP - slow cortical potentials SCP training of slow cortical potentials is aimed at phasic aspects of cortical excitability. Patients train potential shifts in a positive direction (“positivization”; decrease in excitability) or negative direction (“negativization”; increase in excitability).⁴
3. QEEG - Quantitative EEG - EEG with spatial assignment to specific areas of the brain⁵

In practice, frequency bands without spatial allocation or SCP are usually used, as the measurement of brain activity in certain brain areas is technically very complex.
A 1-channel EEG measuring device is sufficient for general EEG frequency band training. For a spatial resolution of brain activity in individual brain areas, a large number of EEG channels (usually 19) are required. The corresponding devices are considerably more expensive and more complex to operate.
qEEG is being investigated as a biomarker for the diagnosis of various mental disorders.⁶

Informative on neurofeedback for ADHD: Strehl.⁷⁸⁹

In our opinion, neurofeedback is the only form of therapy with a healing character, as the brain’s self-regulation mechanisms can be permanently improved. Nevertheless, its effectiveness is very limited. In our opinion, more treatment is required than is typically recommended. In particular, a combination of frequency band training and subsequent SCP training appears to produce good results. However, neurofeedback is not a substitute for treatment with medication (especially in the first few years), but a supplement to therapy that can reduce the need for medication.

• Neurofeedback training types
  • SMR training
    acts on relaxed attention, hyperactivity, impulsivity
  • Theta beta training
    affects concentration, tense attention
  • Alpha training
    acts on relaxation, sleep problems
  • SCP training
    Trauung of the slow cortical potentials
    acts on activation (for ADHD-I) and downregulation (for ADHD-HI)
    very strenuous, but also very helpful when carried out

• 1. Neurofeedback frequency band training
  • 1.1. Brain frequencies and their function
    • 1.1.1. The EEG brain frequencies
      • 1.1.1.1. Delta
      • 1.1.1.2. Theta
      • 1.1.1.3. Alpha
      • 1.1.1.4. SMR / Low Beta
      • 1.1.1.5. Beta
      • 1.1.1.6. High beta
      • 1.1.1.7. Gamma
    • 1.1.2. EEG brain frequencies according to further criteria
    • 1.1.3. Changes in EEG brain frequencies with age
    • 1.1.4. QEEG and medication response
  • 1.2. Brain wave examination methods
  • 1.3. Basics of frequency band training
• 2. Different EEG forms in subtypes of ADHD
• 3. Connectivity of brain regions in ADHD
• 4. Training models
  • 4.1. Frequency band training
    • 4.1.1. SMR training (increase of low beta, 12 - 15 Hz)
    • 4.1.2. Theta-Beta or Theta-Alpha Ratio Training
    • 4.1.3. Individual beta rhythms
  • 4.2. Training the self-regulation ability of the EEG
  • 4.3. QEEG neurofeedback training (z-score training)
  • 4.4. HeartMath - Training
  • 4.5. Training of slow cortical potentials (SCP)
  • 4.6. Infra Low Frequency Training - ILF
  • 4.7. LORETA neurofeedback
  • 4.8. EEG-BCI training
  • 4.9. Functional MRI neurofeedback (fMRI-NF)
  • 4.10. Prism EFP Neurofeedback
• 5. Effectiveness of neurofeedback
  • 5.1. Mechanism of action
  • 5.2. Treatment effectiveness
  • 5.3. Duration of effectiveness of neurofeedback
  • 5.4. Combination of neurofeedback and medication
• 6. More about neurofeedback

1. Neurofeedback frequency band training¶

1.1. Brain frequencies and their function¶

The brain communicates by means of groups of nerve cells that transmit and pause in a synchronized rhythm. The different frequencies of the rhythms correlate with specific states of consciousness and can be measured in EEG frequency bands.

1.1.1. The EEG brain frequencies¶

1.1.1.1. Delta¶

• Delta - 0.4 to 3/3.5 Hz - calm, dreamless deep sleep, trance, unconsciousness; in infants and toddlers also during wakefulness

1.1.1.2. Theta¶

• Theta - 4 to 6.5/7 Hz - doziness and light sleep, waking dreams, hypnosis.
  Strong occurrence in infancy and childhood, becomes weaker with increasing age / increasing brain maturation (frequencies increase).

It could be argued that theta-beta training accelerates brain maturation in younger children.

Theta occurs in various forms:

• Cortical theta rhythm¹⁰
  In infants and small children, even when awake
  in adolescents and adults: for drowsiness, in the transition from wakefulness to sleep, during sleep, otherwise for brain pathologies
• Hippocampal theta rhythm / frontal-midline theta¹⁰
  Decrease near the apex of the head (Cz)
  Is in connection with
  • The anterior cingulate cortex (correlating with simultaneously increased glucose metabolism)
  • Theta rhythm in hippocampus, parahippocampal cortex, anterior cingulate cortex, mammillary bodies of the hypothalamus, medial dorsal nucleus of the thalamus
  • Papez cycle, which probably serves less to control emotions, but is more relevant to memory.
• Dysfunctional front-midline theta (FMT) in the 5.5-8 Hz range¹¹
  More on this at ⇒ ADHD subtypes according to EEG

1.1.1.3. Alpha¶

• Alpha - 7/8 to 12/13 Hz - alert relaxation, inward attention, closed eyes. Healthy physiological basic rhythm of the resting brain. Strongest expression moves from the center of the head (in children) towards the back of the head (puberty). Occurs predominantly when the eyes are closed and is replaced by beta waves when the eyes are opened or when thinking (e.g. mental arithmetic with eyes closed).
  Pronounced alpha waves increase creativity and reduce the risk of depression.¹²
• Alpha (and beta) are used (at least for visual stimuli) for top-down information transfer from more highly developed to less developed areas of the brain.¹³
• Methylphenidate and amphetamine drugs increase the power of alpha in the EEG (in rats), while atomoxetine and guanfacine do not.¹⁴

1.1.1.4. SMR / Low Beta¶

• SMR (Low Beta) - 12 to 15 Hz - calm concentration, relaxed/undirected outward attention.
  SMR is also described as alpha from 9 to 13 Hz, derived via C3 and C4.¹⁵
  Migraine often presents with a low SMR and an elevated beta, often accompanied by a rather high theta.
  Migraine: Training up SMR is a typical and helpful migraine treatment.
  Many GPs prescribe NFB for migraines. If this does not help either, the only option is often beta blockers (which do not block beta waves in the brain, but beta receptors in the heart), which should, however, be the last choice.

1.1.1.5. Beta¶

• Beta - 15 to 21 Hz - active attention
  Overactivation in the beta area in ADHD is often associated with underactivation in other areas of the brain, usually in the frontostriatal loop.¹⁶ This may correspond to the mechanism whereby an overactivated PFC (high dopamine level) deactivates the striatum (low dopamine level).
  Inattention correlates with low beta (12 - 30 Hz) and low gamma values (30 - 50 Hz).¹⁷ We assume that this refers to ADHD-I-typical inattention due to boredom, which is caused by an underactivated PFC, and not to ADHD-HI-typical inattention due to distractibility, which is caused by an overactivated PFC.
  In many adults with sleep disorders, beta is elevated at rest, i.e. increases when trying to relax instead of decreasing; often also in depression.

• Beta (and alpha) are used (at least for visual stimuli) for top-down information transfer from more highly developed to less developed areas of the brain.¹³

1.1.1.6. High beta¶

• High Beta - 21 to 35/38 Hz - hectic, stress, anxiety, overactivation.
  If too high: circling thoughts, not being able to keep your head still

1.1.1.7. Gamma¶

• Gamma - 35/38 to 45/70 Hz - tense attention, strong concentration and information processing
• Gamma (around 60 Hz) is used (at least for visual stimuli) for bottom-up information transfer from less developed to more developed areas of the brain.¹³

Higher scores on the Childhood Trauma Questionairy correlate (alongside higher ADHD symptoms) with

• Higher QEEG values in the areas¹⁹
  • Delta
  • Beta1
  • Beta2
  • Beta3
  • Gamma
• Significantly reduced power values in the area²⁰
  • Low alpha
• Inattention correlates with
  • Low beta (12 - 30 Hz) values¹⁷
  • Low gamma values (30 - 50 Hz)¹⁷
  • Increased high-beta²⁰
    • Which was unexpected for the authors
  • Low low alpha²⁰
    • Which was unexpected for the authors

1.1.2. EEG brain frequencies according to further criteria¶

• Rolandic waves²¹
  depending on age in the range 14 to 30 Hz. They are taken (at rest with eyes closed or open) at C3, Cz, C4 above the (eponymous) Rolando furrow (sulcus centralis).
  They differ²²²³ in
  • (rolandic) μ-rhythms or murrhythms
    The name comes from the fact that their shape is reminiscent of a μ.
    The origin of the μ-rhythms is the primary somatosensory cortex.²¹
  • Rolandic beta waves
    The origin of the rolandic beta waves is the motor cortex.²¹

1.1.3. Changes in EEG brain frequencies with age¶

In all people, the power (current strength) of the EEG changes with increasing age, especially from childhood to adolescence. With increasing age, the slow frequencies (delta and theta) decrease, while the fast frequencies (alpha and beta) become stronger. The total power (the sum of all individual frequency power values) decreases with increasing age.²⁴²⁵ Age-related changes occur in people with ADHD as well as in people without ADHD, but to varying degrees. In people with ADHD, the EEG becomes more similar to that of non-affected people as they get older, while it remains almost constant in people with ADHD-I.²⁶²⁷

When studies speak of “slowed brain activity”, they are referring to an increase in the power of slow brain frequencies (delta/theta).

1.1.4. QEEG and medication response¶

One study observed that in different attentional and affective disorders (here: ADHD, depression, bipolar according to DSM-III-R) the neurometric characteristics predicted the medication response:²⁸

• frontal alpha excess responded 87% to antidepressants
• frontal theta excess responded 100% to stimulants
• frontal alpha excess with hypercoherence responded 85% to anticonvulsants/lithium
• frontal theta excess with hypercoherence responded 80% to anticonvulsants

A study on ADHD found a good response at²⁹
(too small numbers of subjects in the groups are noted)

	ADHD-HI	ADHD-I
frontal alpha excess	(100 % AMP (n = 3))	(66.7 % AMP (n = 3))
	92.3 % MPH	(50.0 % MPH (n = 8))
frontal theta excess	76.9 % AMP	41.2 % AMP
	73.9 % MPH	20.0 % MPH
frontal beta excess	(100 % AMP (n = 3))	(50 % AMP (n = 2))
	(66.7 % MPH (n = 3))	(75 % MPH (n = 4))

1.2. Brain wave examination methods¶

• EEG
• Independent Component Analysis (ICA)
• Low Resolution Electromagnetic Tomography (LORETA)

1.3. Basics of frequency band training¶

Targeted training of the affected brain frequencies (frequencies with too low intensity are trained up, frequencies with too high intensity are trained down) can influence the symptoms correlating with the respective frequencies. Not all people with ADHD respond equally. In 50% of people with ADHD, symptoms can be reduced by at least 25%. Between 20 and 40 training sessions are required. The effect of neurofeedback remained unchanged 6 months after the end of therapy.³⁰

Procedure of the therapy session:
The patient’s EEG frequencies are measured using electrodes and visualized on the therapist’s PC.
The software now offers the patient stimuli (graphics, games, films, etc.) that they can control using the EEG frequency to be trained. For example, the patient can control an airplane by changing the flight altitude based on the measured value of the brain frequency to be trained.

Neurofeedback enables the targeted training of individual areas of the brain.

Neurofeedback is not yet generally reimbursed by health insurance companies. However, prescribed occupational therapy is reimbursable. Occupational therapists who also offer neurofeedback are authorized to use neurofeedback as a form of occupational therapy.

An in-depth presentation of neurofeedback training for home use on YouTube by Dr. Kowalski, who (together with Ute Strehl) conducts research into neurofeedback at the University of Tübingen, using software he has developed for home use: Kowalski on Youtube on neurofeedback.
According to this view, neurofeedback can generally be used to train at home, but it is essential to be accompanied/guided by a registered neurofeedback therapist. If the wrong frequencies are trained in the wrong direction, the training could be harmful. It is not yet possible to assess whether the devices and software are now sufficiently functional for home use.

2. Different EEG forms in subtypes of ADHD¶

The different subtypes of ADHD have specific different EEG patterns. This means that in ADHD - depending on the specific subtype - different brain frequencies are more active or inactive than in non-affected people.

There are more subtypes that can be classified according to EEG than there are classic subtypes classified according to behavior.

Find out more at ⇒ ADHD subtypes according to EEG in the article ⇒ The subtypes of ADHD: ADHD-HI, ADHD-I, SCT and others.

3. Connectivity of brain regions in ADHD¶

See also ⇒ Brain networks and connectivity in ADHD In the chapter ⇒ Neurological aspects.

4. Training models¶

4.1. Frequency band training¶

4.1.1. SMR training (increase of low beta, 12 - 15 Hz)¶

SMR training is a typical neurofeedback protocol for ADHD with the aim of increasing brain activity in the SMR band.³¹³²
The SMR plays an important role in motor excitability.³²
An amplification of the sensorimotor rhythm (12-15 Hz) over the motor cortex should reduce hyperactivity by inhibiting the thalamo-cortical loop.³³
SMR training reduces impulsivity and emotional outbursts and promotes calm general attention. Calm attention is different from concentration (focused attention).

In a study on SMR frequency band training, 86 children between the ages of 9 and 14 with ADHD were tested with Go/NoGo tasks. Each training session included:

• Improvement of the power ratio of 15-18 Hz in relation to the rest of the EEG (20 minutes), measurement on C3/Fz
• Improvement of the power ratio of 12-15 Hz in relation to the rest of the EEG (7 - 10 minutes), measurement on C4/Pz

After 15 to 22 neurofeedback sessions, those children who performed well in the training sessions showed an increase in P 300 (positive component within 180-420 ms latency) over frontal-central brain areas in the Go/NoGo tasks. No increase was observed in the children who performed worse in the training sessions.³⁴
Accordingly, the activation of frontal-central brain regions appears to be accessible to beta training.

4.1.2. Theta-Beta or Theta-Alpha Ratio Training¶

The theta-beta ratio proved to be stable between the ages of 7 and 11³⁵

There are several studies with different protocols:³⁶

• Theta down
• Train theta down and beta up at the same time
• Train theta down and alpha up at the same time
• Train theta down and alpha down and beta up at the same time
• Theta up and beta down
• Theta up

The two training methods, theta down - beta up and theta down - alpha up, differed only slightly in their results. Both protocols alleviated the symptoms of ADHD in general (p <0.001), the symptoms of hyperactivity (p <0.001), inattention (p <0.001) and omission errors (p <0.001), but not the oppositional and impulsive symptoms. Up-training alpha is said to be more effective in terms of omission errors.³⁷

A study comparing theta down - beta up training with the mechanisms of action of MPH concluded that theta down - beta up training primarily improves inhibition and impulse control rather than attention, but uses different mechanisms of action than methylphenidate.³⁸ The theta/beta ratio at rest appears to mediate the relationship between motor competence and inhibition in children with ADHD.³⁹
One study found different effects of theta-beta frequency band training on different subtypes (ADHD-I and ADHD-C).⁴⁰
A very small pilot study found that alpha/theta training also improved attention in students without ADHD.⁴¹

It is possible that a protocol that only upregulates theta is particularly suitable for promoting inhibition. Theta upregulation led to a specific Nogo-P3 increase, whereas beta upregulation training and the combined protocol led to less specific effects.⁴²

A protocol that only downregulated theta (for five days, 100 minutes in total) showed a significant improvement in general attention performance and executive control network efficiency in healthy children, while attention and orientation networks remained unchanged.⁴³

4.1.3. Individual beta rhythms¶

An experimental neurofeedback method used individual beta rhythms.⁴⁴

4.2. Training the self-regulation ability of the EEG¶

This approach turns away from a normalization of EEG frequencies (comparison according to QEEG databases), as studies in groups could not find uniform characteristic EEG frequency patterns in children with ADHD.³⁶
Therefore, some neurofeedback protocols work with alternating training phases to increase and decrease the target values, which also corresponds to the usual approach in SCP regulation.

4.3. QEEG neurofeedback training (z-score training)¶

The Z-Score training is based on a database with test results from around 600 healthy children and adults. The training aims to change your own brain frequency activities in the direction of these normal values.³⁶

The effectiveness of Z-score training could also possibly be achieved by training the self-control ability of the EEG. The fact that normal values of healthy people are used as target parameters does not change this.

One study reports that 70% of adults and 52% of children with ADHD improved so much after 30 sessions of a combination of Z-score neurofeedback and HRV biofeedback training that they were no longer diagnosed in the clinical range of the ASEBA, but in the normal range. Whereas the diagnosis was previously in the borderline range of the ASEBA, 80% of the adults and 63% of the children reached the normal range. 4 of 9 adults and 10 of the 44 children who had previously used ADHD medications did not use ADHD medications after treatment. No adult (out of 30) and 1 child (out of 55) who had not used ADHD medication before treatment used ADHD medication for the first time after treatment.⁴⁵
If these results can be repeated, this would be a sensation - despite all scientific caution. The results should therefore be viewed with appropriate attention and caution.

4.4. HeartMath - Training¶

Another probably trademarked training model (which always arouses a certain amount of suspicion from a scientific point of view) is reported in a study on the provider’s website on positive results with ADHD, which are said to be based on coherence improvements in brain communication.⁴⁶

4.5. Training of slow cortical potentials (SCP)¶

SCP (slow cortical potentials) are very slow frequencies of less than 0.1 Hz that lie outside the measurement range of a standard EEG.
SCP are reactions of the brain to internal or external stimuli. SCP are understood as a measure of the excitability of the cerebral cortex.

Training to control the slow cortical potentials can help to provide improved information processing in line with the situation.
The effectiveness of SCP training for ADHD has been proven several times.⁴⁷

In one study, SCP had a slightly better effect on children’s ADHD symptoms than theta/beta frequency band training.⁴⁸ However, a combination of both methods should make sense.
It would also be conceivable to train the self-regulation of interhemispheric frontal asymmetry through SCP feedback. This has already been proven in healthy people.⁴⁹ This involves specifically training the synchronization of the left and right hemispheres of the brain. This is particularly helpful in cases of depression, as depression and chronic frustration are often associated with a particularly strong asynchrony of the cerebral hemispheres.

4.6. Infra Low Frequency Training - ILF¶

ILF training was developed by Othmer and is therefore sometimes referred to as the Othmer method. It also involves training at very low frequencies. The technique, derivations and feedback are designed differently.

ILF training is not aimed at increasing or decreasing frequencies at will, but at mirroring / visualizing specific parameters of the low brain frequencies in such a way that it can be used to learn to regulate the excitation level of the central nervous system (CNS) itself.

4.7. LORETA neurofeedback¶

LORETA (Low Resolution Brain Electromagnetic Tomography) is a three-dimensional electromagnetic brain imaging technique with low resolution. LORETA is used to train specific brain networks in deeper brain structures through three-dimensional observation.

4.8. EEG-BCI training¶

One study reports on a form of training in which the test subjects do not learn to suppress or amplify certain EEG waves on the basis of preset standard values, but with an EEG pattern individualized by the system, which represents the optimal attention based on the training activities.⁵⁰
Treatment with 24 sessions resulted in a 3.2-point improvement in inattention according to the ADHD Rating Scale.

4.9. Functional MRI neurofeedback (fMRI-NF)¶

An RCT study on fMRI-NF of the right inferior frontal cortex (rIFC) showed a promising improvement of ADHD symptoms in the application group as well as in the control group.⁵¹

4.10. Prism EFP Neurofeedback¶

Prism EFP Neurofeedback is specifically designed to improve emotional dysregulation (ED) by downregulating amygdala activity.
A pilot study of 9 adults with ADHD showed a reduction in total DSM ADHD symptoms by about two-thirds after ten to fifteen sessions over five to eight weeks at the end of treatment, and improvement observed in all other clinical measures.⁵²

5. Effectiveness of neurofeedback¶

5.1. Mechanism of action¶

The fundamental effectiveness of neurofeedback is widely recognized scientifically.⁵³⁵⁴

We assume that the effectiveness of neurofeedback lies in learning to control your own brain activity. Our own brain activity is always unconsciously controlled: when we go to sleep, we calm down, we concentrate before an important task. This is nothing other than a self-regulation of (among other things) certain brain frequencies. People who react abnormally to certain brain frequencies or who have poor self-regulation can learn to improve this by specifically reinforcing successful self-regulation.

5.2. Treatment effectiveness¶

In individual cases, the degree of effectiveness (Effect size) can reach that of medication^{5556 57} , even if the average values are significantly lower.
An extensive study compared neurofeedback with stimulant treatment and found equally strong improvements in both groups.⁵⁸
One study found significant improvements in children treated with 40 neurofeedback sessions with theta-beta training compared to traditionally treated children in the following areas⁵⁹

• Sustained attention
• Verbal working memory
• Response inhibition / impulse inhibition
• Behavioral problems
• Academic achievement.

A comparative study between neurofeedback (theta/beta training) and EMG biofeedback (30 sessions each) showed a reduction in theta/beta quotients in both groups over the course of treatment. However, attention performance, intelligence level and behavior only improved in the neurofeedback group, and then only significantly. In the EMG biofeedback group, on the other hand, there was only an indication of an increase in working speed in a writing attention test, which was not confirmed in further tests.⁶⁰
A double-blind study comparing (blind) neurofeedback with sham neurofeedback and also (non-blind) with cognitive behavioral therapy, which involved considerable effort, came to the conclusion that neurofeedback can reduce ADHD symptoms just as effectively as cognitive behavioral therapy - however, the sham neurofeedback was just as effective.⁶¹

There are further reports of very significant symptom reductions through neurofeedback. Nevertheless, neurofeedback is not considered effective for all people with ADHD (responder rate 25 to 50 %) and cannot completely eliminate the symptoms even in these cases. For responders, a 50% reduction in symptoms is considered a success.

According to a meta-analysis, various factors improve the effectiveness of neurofeedback therapy:⁶²

• More intensive treatment (more treatments)
  • But not the duration of treatment (time over which the applications are spread)
• Use of high-quality EEG devices

One study found markers that can be used to better predict which people with ADHD will respond to neurofeedback and which will not.⁶³
In another placebo-controlled study, the efficacy of neurofeedback on inattention was greater in the children with the most severe or least severe previously measured deficits in the basic cognitive components of information integration.⁶⁴

Regardless of this, parents report a higher improvement compared to teachers. This is reminiscent of the differences in the evaluation of diets, where there is a suspicion that the high cost of the treatment promotes the attitude that this high cost must also result in high success (bias).⁶²

In another comparison of stimulants and neurofeedback, the drug-treated group performed significantly better, while barely any effects were shown for neurofeedback after 30 treatments.⁶⁵ Similarly, a comparative study between 25 sessions of SCP or Z-score within 5 weeks with usual treatment and working memory training found no significant benefits for neurofeedback.⁶⁶ Several meta-analyses found no efficacy of neurofeedback compared to placebo or even reduced efficacy of neurofeedback in placebo-controlled trials.⁶⁷⁶⁸⁶⁹
It remains to be seen how this can be reconciled with the multiple studies on existing efficacy in some persons with ADHD.

A review came to the following conclusion:⁷⁰

• 2 meta-analyses confirmed significant efficacy of standard neurofeedback protocols for parent- and teacher-rated symptoms with a medium effect size and sustained effects at 6-12 months.
• 4 multicenter RCT’s showed significant superiority over semi-active control groups, with medium Effect sizes at the end of treatment or follow-up and remission rates of 32-47%.
• Open studies confirm the effectiveness of neurofeedback
• No signs of publication bias were found
• No significant neurofeedback-specific side effects were reported.

A meta-analysis found no improvement in executive functions in ADHD with neurofeedback.⁷¹
A double-blind, randomized and controlled study found efficacy in ADHD on aggressiveness and impulsivity after just 12 one-hour sessions over 6 weeks⁷²
A study of adults with ADHD reports improvements in attention and unclear results on hyperactivity.⁷³ Neurofeedback in combination with cognitive computer games (CCGs) can improve time perception, attention and working memory in children with ADHD.⁷⁴
A meta-analysis of 14 studies on EEG neurofeedback with n = 429 subjects with ADHD found:⁷⁵

• significant improvement in attention performance
• significant EEG-NF-associated positive effect on sustained attention
• limited improvement in selective attention
• limited improvement in working memory
• Beta enhancement protocol worked better than theta/beta ratio reduction or SCP modulation protocols
• three sessions per week showed the best effect with EEG-NF
• Effectiveness of surface EEG-NF was significantly lower
• Improvement in attentional performance persisted unchanged 6-12 months after the EEG-NF intervention (3 studies)

In a small study, neurofeedback improved visual memory, acoustic short-term memory and auditory working memory, but not perceptual organization.⁷⁶

In conclusion, standard neurofeedback protocols were considered a well-established treatment for ADHD with medium to large effect sizes.

5.3. Duration of effectiveness of neurofeedback¶

Studies suggest that the positive effects of neurofeedback treatment are long-lasting. A meta-analysis of 165 studies found a long-lasting significant improvement in ADHD symptoms through neurofeedback.⁷⁷
An extensive study comparing neurofeedback with stimulant treatment found persistent symptom improvements only in the neurofeedback group after discontinuation of the stimulants and cessation of neurofeedback training, which was consistent with the persistent changes in the EEG in these participants.⁷⁸
In another study, the effect of neurofeedback was also unchanged 6 months after the end of therapy.³⁰

One study found that one year after beta/theta training, the significant improvement in working memory achieved by the training was fully maintained.⁷⁹ A randomized double-blind placebo-controlled study on theta/beta training found slight advantages for the treatment group after 25 months, in particular a relatively lower need for medication.⁸⁰
After 40 neurofeedback sessions with theta-beta training in the following areas⁵⁹

• Sustained attention
• Verbal working memory
• Response inhibition / impulse inhibition
• Behavioral problems
• Academic achievement.

the improvements were still present 6 months after the end of treatment.

5.4. Combination of neurofeedback and medication¶

A combination of neurofeedback and methylphenidate proved to be more effective than neurofeedback alone.⁸¹⁸²

6. More about neurofeedback¶

According to this view, research on ADHD suffers fundamentally from the fact that many studies only differentiate their subjects according to ADHD/non-affected individuals. Since the subtypes of ADHD-HI and ADHD-I differ not only in their stress response phenotype (= ADHD symptoms) but also in terms of different amplitude strengths of brain frequencies, a more precise separation would be helpful in studies.

Neurofeedback frequency band training is usually only carried out with 1- or 2-channel systems.⁸³ This means that the brain frequencies are measured at 1 or 2 points on the skull. Medically approved 1 to 2-channel devices cost from €3,500.
With just a few EEG channels, it is not possible to spatially localize the measured brain frequencies to specific areas of the brain. Since very specific brain areas show very specific changes in brain frequencies in ADHD, a spatial assignment would be desirable for some therapies. Frequency band training with 1 or 2 derivations hopes quasi “blindly” that the values measured at these 1 or 2 locations “somehow” originate from the brain areas affected. For this purpose, empirical values are used to determine at which point of the skull the activity of the relevant brain areas is most likely to be measurable. This does not ensure a reliable assignment. However, for some brain frequencies (such as SMR = low beta), a single-channel recording (for SMR at Cz = center of the skull) has already proven to be quite reliable and practicable.
It is to be hoped that as neurological research progresses, multi-channel EEG measuring devices will be used much more frequently and thus become cheaper. It remains to be seen whether the recommended 100 recording channels² will actually be required in therapy for precise spatial allocation of the measured brain frequencies to individual areas of the brain, or whether the 19-channel devices (which cost around €20,000 as of 2017) described by other researchers as sufficiently reliable will be sufficient for reliable diagnostics and therapy.