The Promise of Neurofeedback
By Siegfried Othmer, Ph.D.
November, 2001

Introduction

In the last few years there has been increasing interest in a technique called EEG Biofeedback, also known as Neurofeedback, particularly in connection with Attention Deficit Disorder. This treatment modality involves operant conditioning of the brain’s electrical activity, as reflected in the EEG. This brief article reviews the history and the current state of the field, discusses some possible theoretical models for neurofeedback, and suggests what may lie in store for the future.

EEG neurofeedback falls into the general domain of biofeedback, in which physiological cues are used in a feedback paradigm to change, and ultimately to normalize, certain regulatory functions.

The most commonly known form of biofeedback has been peripheral biofeedback, which generally involves training hand temperature, galvanic skin response (GSR) or muscle activity (electromyographic, or EMG biofeedback). Peripheral biofeedback has been used mostly to treat people with overly reactive or over-aroused nervous systems, with generalized anxiety, with hypervigilance, or with symptoms involving disregulation of the autonomic nervous system. There are also specific applications for incontinence and neuromuscular reeducation.

The common element among the vast majority of people who benefit from biofeedback is a nervous system in need of profound calming, which is often beyond the reach of psychotherapy or of meditative techniques. Neurofeedback is a specialized kind of biofeedback, in which physiological cues from the brain itself are used in a feedback configuration to improve brain function. Since we are usually unaware of our brain’s ongoing electrical activity, the signal level has to be measured and amplified in order for us to be able to observe it. We also have to understand what the signal represents. In the case of peripheral biofeedback, it is easy to see that hand-warming can help an anxious person with cold and clammy hands, or that learning to relax muscles with EMG training can help a person who carries his stress as chronic tension in his neck and shoulders. Due to the much greater complexities of the brain, the potential and implications of working directly with our brain wave activity are both fascinating and profound.

When we look at the EEG we see a lot of rhythmic activity, for example in the well-known alpha rhythm or in the sleep spindles of Stage II sleep. This rhythmic activity is obvious during meditative states and is even more apparent when the brain fails to function properly, as in seizures, where the EEG is often prominently periodic. Rhythmicity appears to be the way the brain organizes continuity of state. The EEG directly reflects the brain’s activity as it regulates its own state. With the EEG we are able to observe how the brain manages arousal, attention, and other functions that involve cortex. The EEG, in fact, directly reflects arousal levels, with higher EEG frequencies corresponding to higher arousal, and lower frequencies corresponding to lower arousal. If EEG frequencies can be conditioned like hand temperature, arousal regulation can be achieved by directly training the brain. At a minimum, this would increase the options available to the clinician.

Over time, it was recognized that a variety of functions that are managed with cortical involvement would yield to neurofeedback training. We are, in fact, training brain functional networks in some generality. We have come to see that structural problems in the brain result in functional deficits that range beyond the locus of the lesion or injury. These functional deficits can often be remediated with a functionally based approach, even if the underlying structural deficit is untreatable. We are therefore looking at a new frontier of trying to understand the brain’s operating system, if you will, along with strategies of improving the functioning of that operating system using learning (i.e., conditioning) strategies.

The History of Operant Conditioning of the EEG

Back in the sixties, sleep research was still entirely in the domain of psychology. M.B. Sterman, then at the UCLA School of Medicine and the Sepulveda VA, was interested in the process of voluntary inhibition related to the initiation of sleep. He trained cats with operant conditioning techniques to obtain food and then trained them to withhold their natural response to go for the food whenever a tone was sounded. During this anticipatory and motorically quiescent period, the EEG showed spindle-burst activity that was later identified as the resting rhythm of the motor system (much as the alpha spindle is the resting rhythm of the visual system). He called this the sensorimotor rhythm (SMR, nominally 12-15 Hz). In further research he rewarded the cats with food for the direct production of these spindles, rather than indirectly by rewarding the behavior of motor stillness. The direct training turned out to be more efficient. It was then discovered that the quality of sleep had also changed in the trained cats. Their sleep was calmer, and briefer, and showed a higher density of sleep spindles. For the first time, animal behavior had been changed in a lasting way by means of EEG conditioning. Later it was serendipitously discovered that the cats that had been trained to enhance their SMR rhythm also had a higher threshold for the onset of chemically induced seizures. This was subsequently replicated in primates.

Initial trials with human subjects with medically uncontrolled epilepsy were published in the early seventies. This was followed by controlled studies using reversal designs and sham training. The history of this research was recently reviewed by Sterman. (Ref.1) Unfortunately, the thrust in neurology and psychiatry toward pharmacology largely orphaned behavioral methods, and funding for large-scale studies was not available. The claims of biofeedback were generally accepted within medicine, but they were marginalized as being tangential to health care. The whole thrust toward self-regulation of physiological function was trivialized by use of such terms as “relaxation training.” Relaxation in this context should be considered as one end of the activation-relaxation dynamic, and the possibility of bringing that under conscious control is far from trivial. Self-regulation is the way our system works and if that process can be improved by conditioning techniques, it should be closer to our first resort in clinical practice rather than our last. This whole issue needs to be revisited.

After 1985, when sponsorship of major studies by the NIH was halted, the field progressed only fitfully at the hands of a few tenacious and somewhat single-minded clinicians who were persuaded of the power of the technique. The applications were primarily to ADHD, traumatic brain injury, the anxiety/depression spectrum, and sleep disorders.

During this entire history, a parallel thrust in the field involved reinforcement of the lower alpha and theta frequencies. This caught the public fancy during the psychedelic age of the late sixties and early seventies and unfortunately tarnished the field to the extent that serious academic inquiry of the technique became almost impossible. Again it was left to a few tenacious clinicians to carry forward until the field was recently rediscovered. In 1989 Eugene Peniston and Paul Kulkosky published stunning research findings using alpha/theta reinforcement successfully with intractable alcoholics in the VA system. (Ref.2) Peniston’s procedures were based on EEG training strategies developed at the Menninger Foundation by Elmer and Alyce Green, Steve Fahrion and Pat Norris, and Dale Walters. Publication of this work revived interest within the biofeedback community in a field that had been thought moribund. Over the last ten years, EEG biofeedback has enjoyed a renaissance of sorts. There are now some 2000 practitioners nationwide, and EEG biofeedback has become the largest interest group within the national professional organization, the Association for Applied Psychophysiology and Biofeedback (www.aapb.org) . A new organization has also sprung up around neurofeedback, the Society for Neuronal Regulation (www.snr-jnt.org). Certification in neurofeedback can be arranged through the Biofeedback Certification Institute of America (www.bcia.org)

An Emerging Model for EEG Biofeedback

The two traditions in the field, one focused on higher frequency training (SMR and beta, covering roughly the frequency range of 12-21Hz), and the other concerned with low-frequency training (alpha and theta, covering roughly the frequency range of 4-12Hz) have continued somewhat in their own separate realms, and have historically attracted distinct populations of clinicians. This is only now changing, as a synthesis between the two approaches is being achieved. The higher frequency training can be thought of as actual training of brain function, producing long-term changes in the operating characteristics of the brain, such as self-regulation of arousal and of attention. The lower frequency training, by contrast, appears to facilitate psychological integration and recovery from trauma.

The SMR/beta training has continued with a dominant focus on ADHD and its comorbidities, the more disruptive behavior disorders, oppositionality and conduct disorder. In its application to ADHD, progress can be documented via continuous performance tests. These tests confirm that fundamental improvements can be obtained in sustained vigilance and impulsivity. Other behavioral observations confirm improvements. Moreover, studies that have evaluated IQ measures in these children have in all cases shown significant improvement. This is presumably because of impact on attentional variables involved in the IQ testing challenges, but specific improvements in working memory, short-term memory, fine motor control, and other functions are also indicated. Medications targeting ADHD in children can in many cases be reduced or eliminated entirely. In a recent survey of outcomes in their ADD Clinic, Michael and Lynda Thompson found that more than 80% of children on Ritalin no longer required the medication after training. (Ref. 3)

The most appealing model to explain these results is that of Malone, Kershner, and Swanson, who attribute ADHD to a bi-hemispheric failure in which a left-hemisphere, dopamine-dominated regulatory system is under-activated, and a right-hemisphere, norepinephrine-dominated regulatory system is over-aroused. (Ref. 4) EEG biofeedback is unique in that training can be tailored to the needs of each hemisphere, and these hemisphere-specific deficits are directly addressed. Left-hemisphere training aids vigilance, whereas right-hemisphere training addresses impulsivity and distractibility. Finally, the matter of inter-hemispheric functional integration can be specifically addressed. The failure modes of both hemispheres appear to be relevant to hyperactivity.

Michael Posner has postulated the existence of networks that mediate attention. These must be coordinated in the time or frequency domain in order to accomplish their function. (Ref. 5) ADHD can be seen as a breakdown in the internal communication of these networks, and EEG biofeedback as a remedy that restores the proper activation, accessibility, and general “attunement” of these communication pathways.

To illustrate the role played by these networks, let us conduct the following Gedanken- or thought experiment. Imagine being in a state of broadly receptive focus, in which equal attention is given to all the incoming sensory stimuli of which you are aware. Then imagine shifting into extreme narrow focus, perhaps onto the detail of the font in which this article is written. What happened in the brain during this transition in the quality of your attention? The shift in external attention was accompanied by an equivalent shift in the brain’s “internal attention.” Both broadly receptive and narrowly focused attention are mediated by linkages of different brain regions, a process which is sustained in the timing or frequency domains. This whole realm of brain function is just beginning to be an area of great research interest in the neurosciences, driven on the one hand by the cornucopia from the new imaging technologies, and by improvements in EEG analysis on the other.

If we put our scientific instead of clinical hat on for a moment, we can argue that the external observables of attention can serve as a mirror of the brain’s internal mechanisms of attention, of these ephemeral internal linkages. The EEG training zeroes in on the latter, and challenges the brain to learn to subtly change its state under these controlled conditions. A repeat of this exercise over many sessions leads to newly honed skills of state management---of the state of arousal, of the state of vigilance, and of the quality of attention. In this model, distractibility can be seen as a discontinuity in the established attentional networks, a condition in which these are too easily disrupted by extraneous stimuli. There may be a deficiency in parallel processing, which would normally allow the intrusive stimulus to be appraised in background. Likewise, impulsivity is a case of too low a threshold for the maintenance of state, but in this case involving the output functions. And then there is hyperactivity, which can emerge either out of too low or too high a setpoint of motor excitability. All these functions are mediated by a myriad of transient linkages in the brain, the activity of which are revealed to us in the EEG.

If this model is valid, then EEG biofeedback addresses the core issues in ADHD—the very brain mechanisms that underlie the phenomenology. And if that is the case, somebody is sooner or later going to ask the question of whether we shouldn’t be calling this a cure. We would answer that this is the wrong question. If a person who suffered a stroke is re-taught language skills, we have not effected a cure for the stroke. We have however, manifestly restored function. Thus it is with EEG biofeedback. Whatever the organic cause of ADHD, it probably remains. But the brain apparently has sufficient plasticity that the organic cause need not be troublesome for the lifetime of the individual. This clearly goes beyond how biofeedback therapists normally think about their craft. At one level this is pure psychology: the technique is operant conditioning, and the appraisal of whether a remedy has been effected is done by the usual psychodynamic and neuropsychological means. At another level, we are clearly impacting on what is increasingly thought of as a medical condition. A reappraisal may be needed as to whether these conditions are more adequately understood in terms of structural or functional models.

With this model in mind, it is now possible to address the subsidiary question of how biofeedback remediates such conditions as oppositionality and conduct disorder. This issue is of great practical import, because the EEG practitioner is likely to see the more difficult end of the ADHD spectrum, where 10mG of Ritalin® is not the answer. The first question is whether such resolution is the secondary consequence of the remediation of ADHD symptomotology, or the primary consequence of the EEG training. Clinical observation suggests that the impact on conduct is too rapid to be attributable to the second-order effects, but is rather the direct result, or a primary effect, of the training.

It appears that the deficient emotional regulation of the conduct-disordered child is also mediated by linkages and networks, just as attention itself, and can be similarly re-ordered through EEG training. In fact, we have reason to think that disregulation of affect is an important part of the ADHD spectrum. Whereas it rises to clinical significance in Oppositional Defiant Disorder and Conduct Disorder, it may be present sub-clinically throughout the ADHD population. Our attention and our actions are guided by inputs from our emotional centers. In line with this thinking, Russell Barkley has in recent years focused on the issue of “intention” even more than “attention” as core issues in ADHD, with the suggestion that intention may be the more fundamental issue. We now have accumulating evidence that neurofeedback can impact the self-regulation of arousal, attention, and affect. This should clearly be of clinical interest to the mental health practitioner, as we are dealing with phenomena that lie at the heart of our psychotherapeutic interventions.

At the beginning of the work with ADHD in the mid-seventies, the use of EEG biofeedback was based on an understanding of ADHD as a neurobiological disorder, which provided a rationale for a physiologically-based remedy. However, Prozac and many similar medications are now routinely administered quite irrespective of whether the depression is endogenous or reactive. And with EEG biofeedback it has been shown that attentional function can be enhanced quite irrespective of whether clinical diagnostic criteria for ADHD are met. Hence, EEG biofeedback obliterates the (already indistinct) boundary between the realms of function and dysfunction. EEG biofeedback now also obscures the distinction between the psychodynamic and the physiological realms. We now know from imaging studies that psychological states have physiological correlates. This should not have come as a surprise. It is not reductionist to assert that the crucial regulatory functions addressed psychodynamically all have their physiological mechanization. One no longer needs an objective evidence for pathology in order to invoke a physiologically-based technique. EEG biofeedback gives therapists a choice of whether to intervene with psychotherapy or with a physiologically based technique such as EEG biofeedback.

Intimations of the Future

Although positive preliminary results have been published, much of the current clinical work with neurofeedback has not yet been fully validated in published research. However, Frank H. Duffy, M.D., Associate Editor for Neurology of the journal Clinical Electroencephalography states: “The literature, which lacks any negative study of substance, suggests that EBT (EEG Biofeedback Therapy) should play a major therapeutic role in many difficult areas. In my opinion, if any medication had demonstrated such a wide spectrum of efficacy it would be universally accepted and widely used.” (Ref.6) Promising applications are extending beyond seizure management and ADHD to the anxiety/depression spectrum, traumatic brain injury, pain management, sleep disorders, trauma recovery, pervasive developmental delay and the autistic spectrum, cerebral palsy, chronic fatigue and fibromyalgia, and addictions, among other conditions. The technique appears to be quite robust in the hands of skilled and knowledgeable practitioners.

It is not premature to speculate that much of psychopathology may have its physiological basis in disregulation of brain function in the bio-electrical domain. We call this the disregulation model. Consider two of the most extreme mental disorders we encounter: Bipolar Disorder and Dissociative Identity Disorder (DID). Both of these conditions are characterized by temporal discontinuities in brain function. Causal mechanisms must be sought in the bio-electrical organization of the brain, since the state transitions cannot be attributed to deliberate changes in the ambient neurochemical environment. In DID, entire personalities can be organized without awareness of one another, on the same cortical real estate, and in the same neurochemical milieu. The pathology does not appear to lie so much within each personality as it does in the instability between them. Each personality requires its distinct bio-electrical network. The discontinuity must be described in the bio-electrical domain.

Neurofeedback can be understood as enhancing the stability of the networks that manage local activation, organismic arousal, attention, and affect. However, the discussion of both the EEG and neurofeedback will progressively become unmoored from concerns with specific diagnoses to consideration of brain function in some generality. It is important to recognize that this model is not inconsistent with the implications of pharmacological approaches to stabilizing brain function. The stabilizing effect of most psychoactive medications is not immediate, but rather takes days to weeks. If we were dealing with simple neurotransmitter deficits, then the drug action would be immediate, as it is with stimulant medications. The fact that efficacious drug action generally takes time is proof that it is dependent on long-term adaptations of the brain to the presence of the drug. This more realistic model of drug action is fully consistent with what we are doing with EEG biofeedback—namely stimulating long-term adaptations. Medications are increasingly seen in terms of affecting neuro-regulatory networks rather than specific conditions. Likewise, EEG biofeedback must be seen as impinging on bio-electrical networks in general rather than on specific diagnostic conditions.

We have a need therefore to describe brain function in terms of the conventional theory of control systems, in that the brain must satisfy stability criteria just as any other feedback control system. The EEG then yields an index of the quality of self-regulation, and many of the exotic features of the EEG which have to date been deemed not to be of “clinical significance” by neurologists may yet emerge to have significance for us in discerning the general quality of brain self-regulation.

Since we are using a learning paradigm, and impinging on the brain’s intrinsic regulatory mechanisms in a fairly natural way, there is reason to believe (and cumulative clinical confirmation) that the effects of the training can be lasting. And since training can be done over the entire spectrum of EEG frequencies, and over all scalp locations, it appears neurofeedback training has the potential of being much more nuanced than pharmacotherapy. Neurofeedback can allow the practitioner to navigate the frontier of physiologically-based management of mental disorders in an organic way that complements and takes advantage of available assessment and psychotherapeutic skills. It offers a very promising third major treatment modality to complement psychotherapy and pharmacotherapy. Readers interested in a more detailed treatment of this model for EEG biofeedback will find it in the new Academic Press book “Introduction to Quantitative EEG and Neurofeedback,” (Ref. 7).


References:

1. Sterman, M.B., Basic Concepts and Clinical Findings in the Treatment of Seizure Disorders with EEG Operant Conditioning, Clinical Electroencephalography, 31, #1, 45-55 (2000)

2. Peniston, E.G., Kulkosky, P.J., Alpha-Theta brainwave training and beta endorphin levels in alcoholics. Alcoholism Clin. Exp. Res. 13,271-279 (1989)

3. Thompson, L., and Thompson, M., Neurofeedback Combined with Training in Metacognitive Strategies: Effectiveness in Students with ADD, Applied Psychophysiology and Biofeedback, 23, #4, 243-263 (1998)

4. Malone, M.A., Kershner, J.R., and Swanson, J.M., Hemispheric processing and methylphenidate effects in attention-deficit hyperactivity disorder. Journal of Child Neurology, 9, 181-189 (1994)

5. Posner, M.I., Petersen, S.E., Fox, P.T., and Raichle, M.E., Localization of cognitive operations in the human brain, Science, 240, 1627-1631 (1988)
6. Duffy, F.H., Editorial, The State of EEG Biofeedback Therapy (EEG Operant
Conditioning)in 2000: An Editor’s Opinion, Clinical Electroencephalography, 31 #1, v-viii (2000)
7. Othmer, S., Othmer, S.F., and Kaiser, D.A., EEG Biofeedback: An emerging model
for its global efficacy. Introduction to Quantitative EEG and Neurofeedback, James
R. Evans and Andrew Abarbanel, editors, Academic Press, San Diego, 244-310
(1999) (Available as a monograph from EEG Spectrum International, www.eegspectrum.com)

For additional information, see other areas of this website (www.eegspectrum.com) or those of the organizations listed in the text. A popular treatment of the history of this field is to be found in “A Symphony in the Brain”, by New York Times Science writer Jim Robbins.