Deborah Simkin, MD

To understand the role of neurofeedback as an intervention to target symptoms associated with psychiatric disorders.

At the end of this CE activity, participants should be able to:

• Discuss neurofeedback mechanisms of action

• Discuss the importance of using quantitative EEG to assess which patients can benefit from specific neurofeedback protocols

• Describe various neurofeedback protocols and techniques

This continuing medical education activity is intended for psychiatrists, psychologists, primary care physicians, physician assistants, nurse practitioners, and other health care professionals who seek to improve their care for patients with mental health disorders.

This activity has been planned and implemented in accordance with the Essential Areas and policies of the Accreditation Council for Continuing Medical Education (ACCME) through the joint providership of CME Outfitters, LLC, and 

. CME Outfitters, LLC, is accredited by the ACCME to provide continuing medical education for physicians.

CME Outfitters designates this enduring material for a maximum of 

™. Physicians should claim only the credit commensurate with the extent of their participation in the activity.

Note to Nurse Practitioners and Physician Assistants: 

AANPCP and AAPA accept certificates of participation for educational activities certified for AMA PRA Category 1 Credit™.

It is the policy of CME Outfitters, LLC, to ensure independence, balance, objectivity, and scientific rigor and integrity in all of their CME/CE activities. Faculty must disclose to the participants any relationships with commercial companies whose products or devices may be mentioned in faculty presentations, or with the commercial supporter of this CME/CE activity. CME Outfitters, LLC, has evaluated, identified, and attempted to resolve any potential conflicts of interest through a rigorous content validation procedure, use of evidence-based data/research, and a multidisciplinary peer-review process.

The following information is for participant information only. It is not assumed that these relationships will have a negative impact on the presentations.

Deborah Simkin, MD, has no disclosures to report.

Joel Lubar, MD, has no disclosures to report.

James Lake, MD (peer/content reviewer), has no disclosures to report.

 staff and CME Outfitters staff have no disclosures to report.

Faculty of this CME/CE activity may include discussion of products or devices that are not currently labeled for use by the FDA. The faculty have been informed of their responsibility to disclose to the audience if they will be discussing off-label or investigational uses (any uses not approved by the FDA) of products or devices. CME Outfitters, LLC, and the faculty do not endorse the use of any product outside of the FDA-labeled indications. Medical professionals should not utilize the procedures, products, or diagnosis techniques discussed during this activity without evaluation of their patient for contraindications or dangers of use.

This article discusses the history of biofeedback, the use of quantitative EEG (QEEG) to develop protocols to target symptoms associated with psychiatric disorders, and the use of more advanced forms of neurofeedback (eg, low-resolution electromagnetic tomography [LORETA]) to target hubs and circuits. The ability to target circuits fits well with the NIH Research Domain Criteria (RDoC) framework to study mental disorders.

Pioneer work in neurofeedback began with Sterman

 In Sterman’s work, cats were connected to an EEG during operant conditioning. During traditional operant conditioning, a cat learned to push a lever when it was hungry. Then, a new element, a tone, was introduced. The cat learned to wait until the tone stopped to press the lever to receive the reward. Sterman noticed that while the cat was waiting for the tone to stop, a specific frequency rhythm of 12-15 Hz was observed. Moreover, the cat could produce the sensorimotor frequency without the tone to get the reward. Sterman called this rhythm the sensorimotor frequency (SMR).

Lubar was the first to use sensorimotor frequency training on a hyperkinetic child in 1976 by placing 2 electrodes at C3 and C4. The child learned to increase the beta rhythm of 12-14 Hz until the theta (4-8 Hz) rhythm was no longer seen. The result was an increase in attention and a decrease in hyperactivity.

 In 1991, Lubar5 used QEEG to match neurofeedback treatment (EEG biofeedback) to theta-beta ratio (TBR) abnormalities only in individuals with ADHD who were identified with a high TBR. QEEG treatment matching is used to match the abnormal EEG biomarker with the impairing symptom, thereby allowing neurofeedback protocols tailored to the individual.

Slow cortical potential (SCP) neurofeedback measures slow activity (<1 Hz). An upward shift of the negative amplitude improves attention. Rockstroh and colleagues

 used SCP in 1993 for drug-refractory seizures; Heinrich and colleagues

were the first to use SCP for ADHD in 2004.

In a meta-analysis of studies using TBR, SMR, or SCP, the within-subject effect size (ES) for hyperactivity was 0.71, for attention it was 1.0, and for impulsivity it was 0.94.

 In randomized, controlled studies, the ES was 0.80 for attention, 0.39 for hyperactivity, and 0.68 for impulsivity. However, some of the controlled studies used a semi-active control, such as cognitive training, which may account for the lower ES for impulsivity and hyperactivity. None of the included studies used QEEG to identify who would benefit most. Although most of the studies were small, the results suggest some promising aspects of neurofeedback.

Not all children with ADHD demonstrate a high TBR; therefore, it is imperative to identify the correct EEG biomarker to target during neurofeedback using QEEG. When QEEG was used to determine the most suitable protocol for neurofeedback, the ES increased substantially.

 of youths with ADHD, neurofeedback protocols were based on QEEG (Table). The ES for attention was 1.78 and for hyperactivity it was 1.22.

A meta-analysis of effects of stimulant medications in ADHD by Faraone and Buitelaar

 showed an ES of 0.84 for methylphenidate for attention and 1.01 for hyperactivity/impulsivity. Other findings indicate that neurofeedback and stimulant medication may be comparable.

Follow-up findings of parent ratings revealed that not only did the effects of neurofeedback not fade over time, but attention and particularly impulsivity and hyperactivity continued to diminish.

 50% of the children with ADHD who received neurofeedback no longer met criteria for ADHD 2 years post-treatment.

In an in-school trial, students with ADHD were randomized to neurofeedback, cognitive behavioral therapy (CBT), or control groups.

Six months post-intervention, students in the neurofeedback group had significant gains (attention ES = 0.34; executive function ES = 0.25; hyperactivity/impulsivity ES = 0.23; BRIEF [Brief Rating Inventory for Executive Function] ES = 0.31) compared with those in the CBT and control groups. These effect sizes may have been lower because the study did not use QEEG to identify the best protocol. However, at 6 months the patients who had received neurofeedback maintained the same medication dosage, but those in the CBT and the control groups had statistically significant increases in medications (9 mg, P = .002; and 13 mg, P ≤.001, respectively).

Monastra and colleagues9 used QEEG to randomize participants on medication to experimental and control groups. When the medication was removed at the end of treatment, only the participants who had completed neurofeedback were able to sustain their improvements in attention and impulsivity.

In autism studies that used the TBR protocol, significant gains were seen in nonverbal stereotypic behaviors, communication, communication skills, and social interaction.

 These changes were noted at 6-month, 1-year, and 2-year follow-up, respectively.

 used multivariate coherence neurofeedback in their study (improving connections between areas in the brain during neurofeedback). Improvements in attention/executive function, visual-motor skills, language, and parent rating scales were maintained at a 10-month follow-up.

 described positive behavioral changes and a spurt of EEG maturation with theta-alpha neurofeedback training (done 2 years earlier) in 5 learning-disabled children. After 2 years, EEG maturation continued in the children who had received neurofeedback training as well as positive behavioral changes and remission of learning disability symptoms compared with controls.

 children were randomly assigned to 33 sessions of either SCP or TBR neurofeedback. Group assignment and assessment were blinded. There was a significant increase in full-scale and performance IQ as well as improvement in variables of attention. Changes in attention were medium to high for the SCP group (ES = .77-1.19) and medium (ES = .66) for the TBR group. Changes in IQ were medium in the SCP group (ES = 0.54) and medium to high in the TRB group (0.62-0.82); 6-month follow-up scores were not significantly different from end-of-treatment scores.

To guide neurofeedback treatments in dyslexia, Coben and colleagues

 identified hypocoherences in the left temporal lobe using pretreatment QEEG analysis. QEEG determined where hypocoherences occurred in the areas of the brain associated with dyslexia. After 20 neurofeedback sessions, the reading level in children with dyslexia increased by an average of 1.2 years. A long-term follow-up study is needed to determine if these results have been sustained.

In the early 1990s, 3-dimensional (3-D) locations of surface (scalp) EEG were used to identify abnormalities deep within the brain.

 Imaging modalities, including positron emission tomography, single-photon emission computed tomography, and functional MRI (fMRI), were co-registered to create a common anatomic atlas. The advent of EEG tomography (tEEG) provides coregistration of 2 imaging modalities that have similar spatial localization characteristics: fMRI measures blood flow and QEEG adds a high temporal resolution of changes in the electrical sources in the brain that are associated with changes in blood flow.

 devised accurate estimates of the deep (lower) brain sources of the EEG patterns in small regional voxels co-registered to MRI slices. He transformed these raw EEG signals into 3-D images that were then co-registered on the Talairach MRI atlas-thus LORETA was born.

 LORETA allows a clinician to translate QEEG data into a figure that corresponds with and looks like the images on fMRI that are associated with the same disease state. During neurofeedback, as waveforms are adjusted toward normal, the images generated by LORETA also become more consistent with a normal fMRI.

The field continues to advance, and 3-D images now include voxels within the interior of the brain to monitor changes before and after neurofeedback.

 shows the before and after changes of theta in a 12-year-old boy with ADHD after 11 sessions. 

 is the movement of the z score over the 11 sessions; session 1 is the first calibration, the initial z score based on the QEEG is not seen until session 2, where active training begins.

Surface QEEG neurofeedback targets amplitude frequencies or rhythms (ie, changing theta to beta frequencies). Training targets the abnormal calculated real-time QEEG z scores (standard deviations identified above or below the mean). Instead of changing a specific frequency, the focus is on changing QEEG z-score metrics toward normal or the mean (z = 0). Z score neurofeedback allows for more metrics to be targeted in a neurofeedback protocol, including:

• Active training components (up to 10 frequency bands)

• Absolute power (when too little or too much activity occurs in more bands)

• Relative power ratios (when too low or too high a ratio occurs between bands)

• The connectivity metrics of asymmetry (ie, the difference between alpha power in homologous sites that may be indicated in such things as depression)

• Coherence (when too much or too little information is shared between parts of the brain)

• Phase lag (when electrical activity is moving too fast or too slow between parts of the brain)

Three types of Z score neurofeedback utilize QEEG: surface 4 channel Z score neurofeedback (4ZNF), surface 19 channel Z score neurofeedback (19ZNF), and LORETA Z score neurofeedback (LZNF). Although similar in many respects, there is a fundamental difference between ZNF and LZNF.

First, surface ZNF involves measuring the amplitude of neurons directly beneath the electrode, where 95% of the neurons arise from a distance of 6 cm, and all frequencies are mixed together at each electrode. LZNF uses 3-D source localization applied to human QEEG in which the mixture of frequencies under each scalp electrode is unscrambled and linked to sources in the interior of the brain with accuracies of approximately 1 cm.

Second, surface ZNF calculates the z score of identified EEG metrics at various 10-20 surface electrode sites, whereas with LZNF, the z score is calculated for a particular collection of current source density voxels. This makes it possible to conduct neurofeedback with the z scores of the calculated location of deeper cortical dipole generating regions or structures (eg, Brodmann areas, cingulate gyrus, precuneus) and allows for more neuroregulation and enhanced QEEG normalization. Consequently, clinical reports of LZNF suggest that positive outcomes can be achieved with an average of 10 to 20 sessions compared with other types of neurofeedback that can require as many as 80 sessions.

In a retrospective study of children and adults with ADHD, 19ZNF was utilized for all pre- and post-comparisons.

Statistically significant differences were seen in constructs of attention, executive function, behavior, and electrocortical functioning (P = .000-.008; ES = 1.29-3.42 for attention and hyperactivity). These results were achieved with 10 to 15 sessions compared with 40 or more sessions usually required for SMR, SCP, and TBR.

Using LZNF, brain networks and hubs can be identified and targeted as regions of interest for training. By directly targeting these regions of interest, in a z-score framework, LZNF allows for specificity and localization similar to that of fMRI. As brain regions (with deviant z scores) are identified and reinforced toward the mean, the changes in the hubs and connections between hubs can be monitored and correlated to improvements in the clinical symptoms.

In a randomized, double-blind, placebo-controlled study using the LORETA phase neurofeedback for 30 minutes, significant changes were seen in the default mode network and the attention network (P < .001).

 EEG topographies were analyzed by LORETA. Feedback was strictly related to generating sources that have their center (ie, local current density maximum voxels) located within the preselected region of training. Significantly enhanced activity was seen in the left linguistic area of participants who had received neurofeedback (P < .01).

Overall safety concerns have been few. However, lack of individualization of treatment by inexperienced practitioners who merely put on electrodes without proper evaluation may be an important element in causing iatrogenic effects. Adverse effects include manic behavior; anger and irritability; increases in depression, anxiety, and agitation; fatigue; sleep disturbance; emotional lability; OCD symptoms; tics; enuresis; somatic symptoms; decline in cognitive functioning; temporary disorientation or dissociation; and incontinence.

Proper training and board certification in neurofeedback are encouraged by the Biofeedback Certification International Alliance.

 If neurofeedback is used to treat epilepsy or in patients with a history of seizures, a neurologist should co-partner with the clinician to monitor the risk of inducing seizures. Although there are no reports in the literature, many clinicians have reported that medication adverse effects may worsen with neurofeedback, which may indicate that medications are less necessary as brain functions improve. In this case, the medication dosage may be lowered as neurofeedback continues. Future studies should pay particular attention to any adverse effects that may occur during neurofeedback.

The use of QEEG and LZNF may prove an effective treatment approach to psychiatric disorders. Larsen and Sherlin

 graded the evidence for neurofeedback on a 1 to 5 efficacy scale, where 1 indicates not empirically supported evidence that consists primarily of case studies or anecdotal reports, and 5 indicates active treatment superior to placebo in randomized controlled trials conducted at a minimum of 2 independent sites. Neurofeedback was deemed efficacious (level 4) or efficacious and specific (level 5) for epilepsy, ADHD, and anxiety spectrum disorders; probably efficacious (level 3) for traumatic brain injury (TBI), alcoholism/substance abuse, insomnia, and optimal/peak performance; and insufficient (level 2) for depressive disorders, autism, PTSD, and tinnitus.

Subsequent studies by Coben and colleagues

 would raise dyslexia to level 4 and would place autism into level 3. The level of efficacy of neurofeedback seen in various studies indicates that larger, well-controlled studies of efficacy are warranted.

Perhaps one of the most important areas to target for research is PTSD and TBI, especially considering the toll these diagnoses have taken on the military. Presently, there is a $5 million grant for research at Fort Campbell, Kentucky, using LZNF to target PTSD and TBI. Preliminary results have shown promise after only 10 sessions.

Studies using fMRI neurofeedback for PTSD in adults have also been promising. An intervention involved training participants to control amygdala activity after exposure to personalized trauma scripts. Examination of changes in resting-state functional connectivity patterns revealed normalization of brain connectivity consistent with clinical improvement.

In another study, 21 adults who had a history of childhood abuse were trained to reduce the alpha rhythm (8-12 Hz) in a 30-minute neurofeedback session.

 The results were correlated with fMRI connectivity and subjective measures of state anxiety and arousal in a group of individuals with PTSD. The training was followed by a significant increase (rebound) in resting-state alpha synchronization. This synchronization was linked to increased calmness, greater salience network connectivity with the right insula, and enhanced default mode network connectivity with bilateral posterior cingulate, right middle frontal gyrus, and left medial prefrontal cortex.

A study involving 16 cases of servicemen with PTSD utilized LZNF. Based on Cohen’s analysis, large effect size was seen for the current source density in the region of interest ranging from 0.5 to 4.6 with an average of 1.4. A negative correlation between ES and psychotropic medication was found along with a trend toward requiring less medication as training progressed.

Although well-intentioned, many double-blind placebo-controlled studies did not find differences between sham and neurofeedback. These studies used flawed methodologies that are not recognized as valid interventions for neurofeedback, including unconventional protocols; auto-thresholding (where a child is always rewarded even if there is no active learning); reinforcement that was set too high so that no learning occurred because it was too easy; and complicated neurofeedback (where it was difficult to determine whether feedback occurred due to entertainment or treatment).

 Unfortunately, such studies have helped to marginalize neurofeedback as a beneficial intervention for psychiatric disorders.

In response, a collaborative group of researchers was formed to develop a more precise double-blind placebo-controlled study. The NIH-funded study, led by Dr. Eugene L. Arnold, involves the use of a TBR greater than 4.5 in children with ADHD (normal TBR = 2.5).

 The high TBR was selected to determine if it can serve as a biomarker responsive to treatment regardless of ADHD subtype. This biomarker fits the RDoC and should show less variability than medication trials, which do not select for a biomarker.

Localized deep brain changes monitored by LORETA will be used to identify possible moderators and non-specific predictors of neurofeedback outcome. Improvement in event-related potential will also be examined as a possible mediator of neurofeedback therapeutic effect in those in whom TBR is decreased. Event-related potentials will measure attention; pre- and post-QEEGs will monitor progress.

Although more and larger studies are needed, neurofeedback shows promise as a tool to be utilized by psychiatrists. LZNF would provide a more accessible and affordable intervention than fMRI neurofeedback. And, LZNF and 19ZNF seem to be effective with fewer sessions than surface neurofeedback, making them more affordable to patients.

Post-tests, credit request forms, and activity evaluations must be completed online at 

 (requires free account activation), and participants can print their certificate or statement of credit immediately (80% pass rate required). This Web site supports all browsers except Internet Explorer for Mac. For complete technical requirements and privacy policy, visit 

PLEASE NOTE THAT THE POST-TEST IS AVAILABLE ONLINE ONLY ON THE 20TH OF THE MONTH OF ACTIVITY ISSUE AND FOR A YEAR AFTER.

Dr. Simkin is Clinical Assistant Professor, Department of Psychiatry, Emory University School of Medicine, Atlanta, GA. Dr. Lubar is Professor Emeritus, Department of Psychology, University of Tennessee, Knoxville, TN; and Affiliate Scientist for the Center of Complex Systems and Brain Sciences, Charles E. Schmidt College of Science, Florida Atlantic University, Boca Raton, FL.

 Howe RC, Sterman MB. Cortical-subcortical EEG correlates of suppressed motor behavior during sleep and waking in the cat. 

 Sterman MB, MacDonald LT, Stone RK. Biofeedback training of the sensorimotor electroencephalographic rhythm in man: effects on epilepsy. 

 Lubar JF, Shouse MN. EEG and behavioural changes in a hyperkinetic child concurrent with training of the sensorimotor rhythm (SMR): a preliminary report. 

 Shouse MN, Lubar JF. Operant conditioning of EEG rhythms and Ritalin in the treatment of hyperkinesis. 

 Lubar JF. Discourse on the development of EEG diagnostics and biofeedback for attention-deficit/hyperactivity disorders. 

 Rockstroh B, Elbert, Birbaumer N, Wolf P, et al. Cortical self-regulation in patients with epilepsies. 

 Heinrich H, Gevensleben H, Freisleder FJ, et al. Training of slow cortical potentials in attention-deficit/hyperactivity disorder: evidence for positive behavioral and neurophysiological effects. 

Articles

This article provides an overview of the role of neurofeedback as an intervention to target symptoms associated with psychiatric disorders.

Deborah Simkin, MD

Articles

Neurofeedback: Significance for Psychiatry