Phenotypes and Therapeutic Outcomes:

Neurofeedback and Medication

Jay Gunkelman, QEEG

Based on an interview with Nancy Faass, MSW, MPH

The DSM-5 was developed with the idea that the correct diagnosis leads to the proper treatment. The problem is that although behavior patterns tend to be reliably mirrored in the DSM, the resultant diagnosis fails to indicate how patients will respond to treatment. In the ADD population, for example, of two patients with the same DSM-based diagnosis, one may respond positively to a stimulant and the other may react negatively to that same drug, and might require a different medication or even a different therapy.

Rather than basing treatment on behavioral symptoms, in my own work I have developed treatment protocol guided by patterns in the brain’s electrical activity, reflected in the EEG (electroencephalogram). This enables targeted use of medications, supplements, and therapies. For example, only 35% of the general population has a good response to SSRIs. However, if you target antidepressant prescriptions using EEG patterns, that success rate increases to approximately 80%. This approach can also be applied in orthomolecular therapies.

The Phenotype Model

At this point in my career, I have analyzed far more than 500,000 EEGs, passing that number in the 1990s. As I reviewed trends in the EEGs over the years, I noticed patterns of EEG activity associated with common “failure modes” in clinical cases. Over time I identified eleven distinct EEG patterns, two of these having known genetic correlates. In a retrospective paper in 2005, I hypothesized that these eleven patterns were phenotypes, and that nine of these clusters had unidentified genetic correlates. Since not all genes are expressed, evaluation of genetic patterns tends to be less fruitful than the study of epigenetics and phenotypes when looking at genetic expression. These phenotypes are seen as an intermediate step between genetics and behavior, reflecting actual expression of the underlying genetics.

In a phenotype model of patients with addiction, for example, I proposed that people sharing the same EEG pattern (phenotype) were likely to respond to the same therapy, regardless of their psychological history or their “story” as reflected in the DSM categorization. Therapy could be selected based on the EEG pattern, rather than the DSM-based diagnosis alone. Applying this approach in the clinical setting, we found that therapeutic outcomes were optimized. Patients not only achieved sobriety, they also showed measurable improvements in neurocognitive function.

Today we know a great deal about patterns of brain activity through the extensive research on monozygotic twins, who have virtually identical EEGs (unless one of the twins has an acquired condition such as traumatic brain injury). A clinician can match monozygotic twins in pairs simply by looking at their EEG patterns.

Since the 2005 retrospective paper, we have prospectively tested the phenotypes’ predictive accuracy in clinical therapy. The phenotypic patterns test out extraordinarily well, to such an extent that we have been able to identify clusters for which we now have genetic correlates, previously unidentified.

Using our EEG phenotyping, Johnson & Johnson drew blood on 100 patients, from an original sample of 126 research patients identified with the DSM-based diagnosis of depression, further subdivided them using our EEG phenotyping. The researchers found an unexpectedly powerful application of the EEG phenotypes in categorizing these patients into genetic clusters and will be publishing these outcomes. We found, for example, that beta spindles seen in the EEG corresponded with a gene that regulates the enzyme that degrades serotonin (5-HT), dopamine (DA), and norepinephrine (NE). This enzyme, catechol-O-methyltransferase (COMT) has three forms: COMT-0, 1, and 2, referring to the number of allele pairs. We found COMT-2 to be associated with the beta spindle, and a negative response to SSRIs, unlike COMT-0 which predicts a strong SSRI response. As we had hypothesized, the EEG patterns and the genetic testing paralleled specific clinical behavior patterns and responses to treatment. The COMT-2 patients, for example, did not respond well to SSRIs, with a tendency to over arousal. This research was published retrospectively in 2005, and has been validated in subsequent published prospective research on ADHD, depression, addiction, and also non-clinically in peak performance applications.

Subsequent research involved more than 400 psychiatric patients using EEG and qEEG (the numerical analysis of electroencephalography data and associated behavioral correlates). We were able to successfully predict which patients were likely to fail with psychiatric medications and to determine how to treat these medication nonresponders effectively without drug therapy. Traditionally, treatment of any psychiatric condition can be an extended process involving medication trial and error, and a series of treatments or treatment combinations. That can also involve a number of side effects before the patient stabilizes. However, if we examine brain activity before we treat, we can use information gained from EEG patterns and the associated phenotypes to select optimum treatment.

Targeting Treatment

Phenotypes by definition occur in both normal and clinical populations, cutting across DSM-based categories. Each phenotype has specific underlying neurochemistry, making identification of chemical approaches to normalization quite straightforward. Within any single DSM category there are multiple clusters of EEG phenotypes and their associated variety of neurochemistry. Consequently, diagnosis alone is incapable of predicting the right pharmaceutical or nutraceutical approach.

Medications.

EEG patterns can be used to effectively predict which medications are most likely to be effective (or not) for any given patient. In psychiatry, treatment failure is common, as seen in the STAR-D study on depression. This large study found that patients had a 25% chance of having the right drug prescribed the first time, with success rates dropping precipitously with each added trial. For patients who did not respond well by the third drug trial, there was a high likelihood that they would never be effectively treated with medication. Among people suffering from depression, approximately 36% either had a negative response to drugs or never responded appropriately.

Supplements.

Of interest to orthomolecular providers, the eleven EEG phenotypes also suggest which neurotransmitters are imbalanced. For example, one cluster of the ADD population has a genetic marker involving dopamine transporter genetics. Their dopamine levels are insufficient at the level of the striatum (deep basal areas of the frontal lobe), and as a result, elevated theta levels appear in the EEG, an indication that the patient needs more dopamine. This particular pattern can be improved by giving dopamine reuptake inhibitors such as Ritalin, a form of stimulant. Not all stimulants have this effect, as amphetamines have a different mechanism involving increased norepinephrine, which speeds up the alpha frequencies and is not related to the theta rhythm. Practitioners expert in orthomolecular therapy who are knowledgeable regarding amino acid supplementation and nutrition can use their approach to alter the level of the appropriate precursors and support resultant neurotransmitter levels if they know which systems to target.

Neurofeedback.

The phenotype approach can also be efficacious in predicting what type of neurofeedback is needed, based on peer-reviewed and published classic neurofeedback protocols. This approach increases the efficacy of the therapy by indicating protocols that match the client’s physiological pattern. Neurofeedback protocols which match the phenotype patterns were published in 2005.

[

Please insert Neurofeedback table about here or almost anywhere in this section, “Targeting Treatment.”]

Research Funding

The challenge in this field is the fact that the U.S. has not funded neurofeedback research since the 1970s. The E.U., Germany, and Korea have ongoing international studies that are well funded. There is a large consortium research project that was just funded by the NIH, primarily due to the positive outcomes reported from the European research. International neuroscience is re-leveraging interest in neurofeedback research.

The Treatment of Pain

There is no EEG signature specifically associated with pain. There are no patterns that would indicate pain being experienced in one patient as compared with the next, nor any discernable pattern that might suggest the presence or absence of pain in a particular patient. However, this lack of specificity does not mean the brain is uninvolved in the perception of pain, or that changing brain function won’t improve the perception of pain for an individual

Neurofeedback.

On the EEG “tracings” the baseline indicates the level of direct current (DC) in the brain. This direct current (DC) system was first described In the British Medical Journal in 1875 (issue #2) by Richard Caton, decades before the human EEG was developed in 1924, based on the work of Hans Berger.

When an area of the brain is in use, it shifts to an electronegative state. (In the vocabulary of electronics, negative current is referred to as “down.”  However, in describing an EEG, negative voltage is “up.”) In an EEG, when the baseline goes up (negative), that brain is more active in that particular area. To reduce pain, the somato-sensory strip can be shifted from an electronegative state in which the cortex is active, and the patient is perceiving the pain, to an electropositive state in which the brain is “off,” dialing down pain sensitivity. This type of change can be taught with neurofeedback or induced through a variety of treatments.

I experienced this directly when I had the unfortunate occurrence of a severe hand injury. Being fully immersed in the field of biofeedback, I learned how to dial down pain working on myself. At the time, I was in charge of the first state hospital-based biofeedback lab in the world, which gave me access to an exceptional resource. I used neurofeedback to interrupt the pain peripherally, turning off the generation of the pain at the source, and also in the central nervous system, to change my perception of the pain.

This concept of turning off the perception of pain can be seen in the work of Kowakami, a Japanese Kundalini expert (noted for inserting large metal skewers in his neck and tongue, apparently oblivious to pain). I had the opportunity to track his responses on an EEG at one of the demonstrations he gave in the U.S. for a professional society. The data showed that he shifted into an electropositive state, turning off sensation, inserted the skewers, and then turned back on brain activity so that he was fully conscious, but pain-free.

Neuromodulation.

In addition to neurofeedback training, the brain can be treated with DC current stimulation, the treatment essentially turning the brain area being treated on or off, depending on the polarity used for treatment. This is an ancient technique used by the Greeks at the time of Christ. Claudius Galen’s writings report that Pliny the Elder treated patients with epilepsy or migraine headaches using torpedo fish (electric eel). He would place the eel on the head of the patient and the creature would “shock” the patient, knocking them unconscious. When the patient awoke, their migraine headache or seizure was effectively treated.

Although our technology is clearly more sophisticated, treatment today still involves applying electrical stimulation to the cortex through the skin, skull, and meninges. We use neuromodulation techniques that include transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS). Approximately 10% of the current applied to the scalp goes to the cortex, the rest absorbed by skin, bone, and cerebrospinal fluid. This technique changes the excitability of the cortex, promoting a 30% to 40% increase or decrease from a baseline level, which represents 60% to 80% of the total signal. Using this approach, patients can be trained to dial down the pain.

Professionals who wish to refer patients with pain issues will want to seek out neurofeedback experts working in the area of neuromodulation, using techniques such as mag stim or DC stim. This can be a viable approach to reduce excitability in the cortex, particularly for patients who have shown an adverse reaction to medication.

Acupuncture.

Techniques such as acupuncture also have neuromodulatory effects, but are beyond the scope of this paper. The electromagnetic nature of acupuncture points is well-documented. If you measure the body with a DC microvolt meter/null detector and look for areas that have higher electronegative charge, those areas turn out to be acupuncture points. The points are highly specific: moving the detector merely a millimeter or two off the point, the current source drops by orders of magnitude. This DC field system does not follow the structure of the nervous system: it is an energetic system with a distinct structure unto itself.

Therapeutic Outcomes

There are a number of different protocols for neurofeedback. Some practitioners take a highly statistical approach, such as the use of univariate scores from databases of an age-matched healthy reference population that guide their training, though these emerging protocol have yet to validate their efficacy and remain experimental. Earlier methods are well-validated in neurofeedback efficacy literature, having passed the field’s own standards, based on well-designed outcome studies and also on meta-analysis review of published outcomes. The newer, statistically driven techniques may have efficacy, but it will take time for the peer review and publication process to show which approaches have appropriate levels of support.

ADD/ADHD

Pediatricians recently published a consensus position that neurofeedback has achieved level-1 evidence for efficacy in the treatment of ADD/ADHD. (Note that this is a fairly unbiased group of providers who neither suggested superiority over medication nor lesser effect than medication.) The vote of confidence was based on their review of multiple, well-controlled studies and meta-analyses. The meta-analysis literature includes a recent evaluation which reported a greater effect size for neurofeedback than for medication. The majority of the funded studies reviewed in this recent meta-analysis were done outside the USA at academic centers in Europe.

Addictions

The use of neurofeedback in addiction treatment dates back to the 1970s: I used it in my first lab and subsequently wrote an NIH grant for alpha training (which was not funded at the time). There is a robust efficacy literature for neurofeedback applied to addiction, with a history of innovative therapies and positive outcomes, reflecting the work of researchers such as Eugene Peniston and more recently Bill Scott, as the field developed the so-called “alpha/theta” protocol to treat addiction.  The level of efficacy has been judged as “probably efficacious” according the hierarchy of standards in the field of applied psychophysiology.  

Peniston reported a 30% recidivism rate. To provide a context for this outcome statistic, in the field of addiction treatment, Twelve-Step Programs have an 83% to 87% recidivism rate. Another challenge in judging efficacy in addiction is the fact that outcomes are generally only judged based on sobriety. Although people may be clean and sober, they still may not be fully functional.

In 2008, we published research using a standard model of addictions treatment that included group and one-on-one therapy, as generally seen in modern treatment programs, with the addition of neurofeedback. Our goal was to determine which of the eleven identified phenotypes occurred most frequently in the addictions population and to document the effect of NF guided by the phenotypes. We found that two-thirds of the addicted population had an over-arousal drive mechanism. There are three phenotypes associated with over-arousal, with the other one third having cingulate dysfunction, suggesting an obsessive/compulsive drive for those individuals. The over-arousal individuals received classical alpha/theta training, but the other third needed a different approach. Peniston showed in his study that 70% of patients would be successfully treated using this approach, but the intervention was not effective for 30% of participants. Our findings may explain these failures as well as explaining the mechanism for the alpha/theta training efficacy.

The other third of the study population had a disorder involving a different neurological mechanism: a cingulate dysfunction located in the area of the midbrain, which manifests as an obsessive-compulsive drive, rather than as over arousal. If the cingulate is not addressed, the patient’s addiction to drugs can be resolved, but they will still have a tendency to addictive behavior and typically will seek another form of addiction such as the internet, sex, or gambling. Functionally, these patients continue to have a dysfunction, even when they are not drinking or using drugs. Sobriety does not equal health… though it helps.

We found that once the brain is actually functional, whether that originally involved over arousal or cingulate dysfunction, not only are the clients sober/abstinent, they have better brain function . In our study of 30 patients with a history of mixed addiction to drugs and alcohol

,

all 30 of these patients were clean and sober at the study’s completion (identified by blood work), and also at one-year follow-up. They were all still clean at three years, which is a speciously high success rate. At the time of the one-year follow-up, we also administered the Woodcock-Johnson III and found on average a 21 standard-point increase in GIA (an IQ ‘equivalent’). The mean group standard score increased from 98 to 120. There was an average 20-point increase on all neuro-cognitive measures, and no scores degraded. Delayed recall increased from 60 to above 100. Not only were these patients clean and sober, they were more functional.

My conclusion: “When you help the brain work better, it works better.” These gains and sobriety persisted because the patients no longer had the drive toward addiction and their brain function was optimized based on their personal phenotype. Personalized medicine aspires to this approach with the provision of effective evaluation and individualized therapy.

This methodology is based on the use of data from the EEG, applying the phenotype to define the nature of the dysfunction and pairing that with the appropriate treatment—in contrast with a psychodynamic approach to addiction that relies on the individual’s “story.” Given what we now know about neurochemistry and the electrochemical nature of the brain, neurofeedback provides a deeper layer of information relevant to the treatment of addictions based on reliable, repeatable measures. The effect of the therapy is pervasive, with an impact on every area of life requiring brain function.  The changes are seen in the functioning of the patient, reflecting the intertwining of physiology and psychology, body, mind, and consciousness.

Autism

Our work with autism includes patients who have both affective and language issues, from the earlier classification of “Asperger’s autism” (no longer in the DSM) to those with a fully mute presentation. Although it can take one to two years of training, we typically help patients progress to the point where they can no longer be diagnosed as autistic. These are not patients merely experiencing a little social awkwardness. Often they are individuals who essentially have no hope—clients fully on the spectra. In many cases, we are able to get them off the spectra. When we use this approach with “just Asperger’s” clients we generally have much less difficulty resolving their symptoms. Treatment for autism is not fast: it is a learning approach which takes time and effort to slowly climb the EEG learning curve, step by step. Behaviors improve gradually, fading away or occasionally improving in a series of breakthrough experiences. We definitely see strong positive outcomes for our autistic clients.

Epilepsy

Only about one-third of those diagnosed with epilepsy have effective medication control of their clinical presentation, and one-third are considered “intractable.” In the treatment of epilepsy, neurofeedback offers an important and viable adjunctive treatment along with medication or as an alternative to surgery. Today, brain surgery is an increasingly common form of therapy for these conditions. At three-year follow-up, meta-analysis of neurosurgery for epilepsy found a 50% chance of a 50% or greater reduction in seizure rates. Neurofeedback has an 82% chance of a 50% or greater reduction in seizures for patients with intractable epilepsy. In our centers, we have seen excellent results for patients with epilepsy and for those with autism.

[Possible boxed sidebar>]

Dual Diagnosis: Intractable Epilepsy and Autism

In our experience, individuals with the dual diagnosis of epilepsy and autism spectrum disorder are treatable. Consider the case of a severely impaired child whose family contacted one of our professional trainees in Israel. This was essentially a young person with no hope. She was experiencing multiple seizures every day and was placed on anticonvulsant medications to no effect, with an extreme level of disability. This was not a mild case of Asperger’s, but rather full-blown autism.

The little girl, eight years old, could not speak and was unable even to tie her shoes. The child suffered from multiple seizures and she experienced more than 250 electrographic seizure episodes within the first 10 minutes of the baseline EEG. We applied an algorhythm in recording the data to assure consistency across repeated measures, and the reports were also confirmed by visual analysis.

At session 20, there were 115 spikes, reflecting less seizure activity. Clearly no one would describe this as improvement. However, at session 42, there were no spikes, confirmed using both the algorhythm and visual evaluation.

When the treatments began, the child was totally mute. That is the nature of intractable epilepsy and autism, a life essentially without trajectory. These are children who never really find their wings and take off. There is no altitude gained. The expectancy is that adulthood will be more of the same. However, at session 42, the child is riding a bicycle, speaking fluently, seizure-free, and no longer on anticonvulsant medications.

This is not cherry-picked data: this is one of the therapist’s initial dual-diagnosis clients. We flew the therapist to the States to present her work a t a meeting on Catalina Island of SABA (Society for the Advancement of Brain Analysis). An epileptologist and a neurologist specializing in autism were present, and they were amazed by the results. The therapist has video tapes of these children at various stages of treatment. What you see at the beginning is a child who has no hope and at the end, a young person who is fully functional.

[<End of sidebar]

Peak Performance

EEG patterns that mirror brain chemistry are characteristic of both normal subjects and patients. Consequently the EEG can be used not only to treat psychiatric issues, but also to optimize normal functioning. To enhance performance, we treat normal subjects with the same phenotype-driven approach that we would use for clinical patients with a similar EEG pattern. At our treatment centers, we see not only patients with clinical issues, but also athletes and business professionals focused on performance. Another interesting client group is researchers studying consciousness. We do not change the treatment model when a different person walks through the door.

Ultimately, peak performance is not about achieving at the mean. It is a unique outlier state. How many Olympic athletes would be described as “average”?

Goals of Treatment

To assume that treatment takes the patient from an outlier state to the mean is a theoretical model. When you are dealing with a multivariate system such as the brain, you cannot characterize it using univariant measures. Yet today we apply Gaussian statistics from the 1950s to describe brain function. The brain has skew, kurtosis, and a level of complexity that simply cannot be described with a normal distribution.

In a realistic model of healing, we are not moving patients toward the mean, we are identifying their divergence and optimizing their function within their genetic pattern. Healing occurs within the patient’s divergent genetic cluster.

Phenotype research provides the opportunity to define subsets of patients based on the genetic, neurological, neuroelectrical, and neurochemical characteristics within each patient population. Phenotypic modeling gives us the information we need to treat the patient before us, by providing data that indicates which medications, supplements, and therapies will be most effect. For patients with a complex presentation such as autism, this information is essential, given the enormous number of causal factors that can be involved.

We cannot currently alter our patients’ genotypes. However, we can change behavior, genetic expression, and the severity of that expression. This work reflects the convergence of evidence-based medicine and outcomes research, applied in the practice of personalized medicine.

Jay Gunkelman, QEEG

Jay Gunkelman, QEEG Diplomate, is recognized as one of the top leaders in the field of EEG and QEEG, and has processed over 500,000 EEGs since 1972. He has served as president of The International Society for Neurofeedback and Research, as well as a board member and treasurer of the Association for Applied Psychophysiology and Biofeedback and is past-president of the Biofeedback Society of California. Jay was the first EEG technologist to be certified in QEEG (1996) and was granted Diplomate status in 2002. He has conducted, published, or participated in hundreds of research papers, articles, and books, including seminal work on EEG endophenotypes. Jay is co-founder and Chief Science Officer of Brain Science International and is a popular lecturer worldwide on the topic of QEEG and the phenotype identification of neurological disorders.

Brain Science International

BSI provides EEG training, as well as analysis and consultations on EEG recordings for clients in the U.S. and around the world on issues ranging from autism to consciousness research. We focus on EEG and phenotype analysis. Although we do not use neurochemical biomarkers in our own work, we have clients who do use this approach in working with the EEG data.

Brain Science International
2410 San Ramon Blvd. Ste 140
San Ramon, CA 94583
(925) 837-1100

www.BrainsInternational.com

Resources

Certification.

Biofeedback Certification International Alliance (BCIA) offers training, examination, and certification in biofeedback and neurofeedback. States such as Washington require BCIA certification for neurofeedback providers, whereas in other states, licensed psychologists may perform neurofeedback without additional training (Note: buyer beware).

Website: BCIA.org

Pain Therapy.

When seeking a health professional who works with chronic pain, it is helpful to know about the training program at Harvard, the Berenson-Allen Center for Noninvasive Brain Stimulation, which offers coursework in the use of mag stim and DC stim. It is important to make referrals with care. This is emerging technology across the board: just because a provider is good with one area of application does not mean that they are good in another. Not only do they need to know the stimulation techniques, they need to be knowledgeable in the application of these techniques to the diagnostic area of interest.

Website: tmslab.org

References

Gunkelman J. Medication prediction with electroencephalography, phenotypes, and biomarkers. Biofeedback. 2014; 42(2):68-73.

Johnstone J, Gunkelman J, Lunt J.Clinical database development: characterization of EEG phenotypes. Clinical EEG and Neuroscience. 2005: 36(2):99-107.

Editorial.

Nancy Faass, MSW, MPH, provides support for authors in the development of articles, books, manuals, white papers, and writing for the Web via her company and can be reached at info@HealthWritersGroup.com or by phone at 415-922-6234.