What's New in Neurofeedback

• Brain-wave workout may help attention troubles - Milw Jr Sentinel, Jan 24 - www.jsonline.com:80/alive/news/jan00/brain24012300.asp

• NF in "Hawaii Parent" - www.eegspectrum.com/affil/armst1.htm

• WebMD chat transcript: Tom Brod discusses NF - my.webmd.com/content/asset/chat_transcript.526881

• Alternatives to Ritalin (radio report) - http://www.winnipeg.cbc.ca/archives/ritalin/thursday.html

Abstract

The value of providing information-rich feedback in an operant conditioning paradigm is evaluated in a population of children undergoing EEG conditioning training (neurofeedback, or EEG biofeedback) for a variety of conditions, including mainly ADHD, seizures, and mood disorders. Retrospective data analysis yields the finding that some tests of cognitive function show better outcomes with more complex, information-rich feedback. Also, more absorbing feedback yields a larger number of training sessions. It is therefore likely that additional benefit may be derived by continuing in the direction of providing more immersive, information-rich feedback, leading ultimately to full virtual reality implementations.

Introduction

EEG patterns associated with pathology such as seizures, traumatic brain injury, and other conditions can be normalized with operant conditioning techniques¹. In such procedures, information about the EEG signal, typically in the spectral domain, is presented to the patient, who is rewarded whenever relevant EEG parameters approach normalcy, or otherwise exceed particular goals. This challenge, when repeated over the longer term, can achieve permanent normalization of the EEG along with remediation of the corresponding dysfunction². Many patients, particularly young children, currently have trouble identifying with, and having visceral appreciation for, the information in the EEG for which they are being rewarded. This shortcoming may be overcome by the use of more immersive feedback techniques that provide rich visual interest and may appeal to more than one sensory modality (auditory, kinesthetic). The findings of the present study support that conjecture.

Background

The EEG preferentially reflects the collective activity of neuronal ensembles. It is increasingly evident that such collective activity is not simply random noise, but reflects the explicit regulatory activity of control mechanisms that manage such collectivities³. As such, the EEG is potentially a sensitive measure of the state of the brain’s self-regulation in the time domain. Specifically, the EEG appears to reveal the activation-relaxation dynamics of the local neuronal populations being monitored. Operant conditioning on these EEG parameters can then in principle affect the robustness of the underlying self-regulatory activity. Such regulatory activity involves the interaction of cortex with sub-cortical structures, in particular with the thalamocortical loops⁴.

If the EEG is the manifestation of explicit regulatory activity of the thalamocortical loops, then operant conditioning on these EEG phenomena represents a subtle challenge to the brain, a challenge which the brain accepts on the one hand, and to which it reacts, on the other. Hence, operant conditioning sets up a subtle action/reaction dynamic on the brain’s internal control mechanisms. The persistent exercise of such an action/reaction cycle may lead to a more robust functioning of the brain’s self-regulatory capacities. This activity fits a learning model, with the potential of yielding long-term improvements in function^5.

The Disregulation Model of Psychopathologies

Work over the last several decades on thalamocortical mechanisms have yielded an increased appreciation of the generality of their role in the regulation of cortical-subcortical and cortical-cortical timing relationships⁶. Most recently, the suggestion has been made that failure modes in such timing mechanisms, termed "thalamocortical dysrythmias" by Rodolfo Llinas, could conceivably account for a variety of neurological and psychiatric pathologies, as well as chronic pain, and some of the sequelae of traumatic brain injury⁴. This emerging appreciation on the theoretical side finally allows a more respectful appraisal of the finding some 25 years ago of efficacy of EEG conditioning for the remediation of chemically induced seizures in cats, and of treatment-refractory cases of epilepsy in man.

The dysrhythmia (or disregulation) model has several virtues. First of all, it focuses attention on functionally based deficits rather than structural models of disorder. Secondly, it potentially addresses those conditions in which the brain undergoes significant, rapid changes in state: seizures, migraines, vertigo, bipolar disorder, schizophrenia, and Dissociative Identity Disorder^{7, 8}. Since these changes take place while the neurochemical milieu remains relatively unaltered, these state transitions are unlikely to yield in first order to a neurochemical model. Discontinuities of function in the time domain should rely primarily on time-domain analysis. In this regard, a full theory of neuroregulation in the time domain (i.e., the bioelectrical domain) must in principle exist, and that such models may have no particular nexus with neurochemical models of brain function, although mutual consistency must be satisfied. The emerging hypothesis of thalamocortical dysrythmias as a key to neurological, psychiatric, and (we would add) psychological pathologies puts a premium on teasing out the operative regulatory networks⁹.

The ubiquitous rhythm nominally occurring at 40-Hz, together with the principle of time binding, constitutes a viable model for the coherent mapping and down-stream parallel processing of retinotopic and somatotopic inputs¹⁰. The periodicity of such primary processing in turn imposes conditions of periodicity (or, at minimum, temporal coincidence) on regulatory functions that impinge on, and modulate, sensory system activation, focused attention, and motor excitability. It is empirically observed that such regulatory functions---local activation, organismic arousal, quality of attention, and affect regulation---occurs at lower frequencies than 40 Hz (possibly at sub-harmonics of 40-60 Hz) and thus define the terrain for the EEG biofeedback intervention.

The primary variable on which EEG reinforcement impinges is local activation, as manifested in EEG spectral amplitudes. Local activation is in turn coupled to organismic arousal. Organismic arousal is mapped in the two-dimensional space of frequency and amplitude. Higher frequency correlates with higher arousal, and higher EEG amplitude correlates with lower arousal. Empirically, it is found that arousal regulation in the waking state is trainable by reinforcement in the frequency range of nominally 12-18 Hz^11,12. Lower frequencies than 12 Hz tend to be associated with passive, low-arousal, more internally focused states.

EEG biofeedback is potentially a good truth test for models of temporal or bio-electrical neuro-regulation⁶. A variety of evidence is being gathered in the clinical realm that supports the generality of the dysrythmia model: its lack of diagnostic specificity; the dynamic nature of the clinical manifestations; the functional character of the failure modes; the non-locality of deficits; the primacy of timing issues in the morphology of mental disorders^7,3. It is not the function of this paper to buttress the case either for the model or for the intervention.

It must suffice for present purposes to summarize the emerging synthesis of theory and intervention. Disorder is attributed to disregulation in key regulatory loops in the brain. These manifest---either obviously or subtly---in deviations in frequency, amplitude, and phase (coherence) characteristics of the EEG. When the amplitude and phase properties of the EEG are reflected back to the trainee in real time, long-term change in these properties can be promoted through a learning process via operant conditioning. The EEG variables that are the operants in this technique do not need to exhibit overt pathology, however. Often they don’t, at least by current criteria, because the EEG phenomenology is intrinsically highly dynamic and state-dependent. That is to say, the deviation---or variability---of the EEG attributable to dysfunction need not dominate all other sources of variability in order for the training to be effective. It is the on-going regulatory activity of the brain that is challenged to function more robustly. If that effort is successful, the brain will have a higher threshold for excursions into instabilities in the steady state, as one significant outcome. It may improve in the maintenance of homeostasis, as another.

The EEG Biofeedback Technique

The EEG training technique involves reinforcement of activity in two sub-bands of the 12-18 Hz spectral band. Typically the left hemisphere is trained at nominally 15-18 Hz, and the right hemisphere at 12-15 Hz. Electrode placements are typically on sensorimotor strip, referential to the ipsilateral ear at C3 and C4 in the International 10-20 system. Excessive amplitude activity is inhibited (by means of withholding of positive reward) in the frequency range of 2-7 Hz and 22-30 Hz. The lower frequency, or delta/theta band, monitors epileptogenic activity and elevated amplitudes often found in ADHD or traumatic brain injury. The higher band is sensitive to anxiety conditions and elevated scalp muscle tension. The higher frequency band tends to promote sympathetic arousal, and the lower frequency band in turn promotes parasympathetic dominance.

Reinforcement is by visual feedback, in which EEG amplitudes are mapped into geometric figures or into variables such as brightness or speed of movement/ Visual feedback is typically augmented by auditory reinforcement when criteria are met, and by means of continuously tracking tactile feedback, the latter at the option of the patient.

The Experimental Method

In the present study, 120 children with ADHD, epilepsy, or mood disorders underwent 20 sessions or more of EEG biofeedback using Neurocybernetics’ 2-Channel EEG biofeedback system. The amount and quality of visual interest, and the complexity of three-dimensional rendering varied across feedback displays. For each 30-min session, subjects selected one of five displays: a maze or pacman-like game (2-dimensional display); stationary boxes which mapped the amplitude in the three bands in terms of the size of the boxes against a goal (2-d); the same game with the addition of a casino-game that reinforced for maintaining continuity of state; line drawings of a receding highway (simple 3-d); and a spaceship travelling past planets (true 3-d rendering). Patient retention and cognitive performance were assessed.

The different feedback options appeal variously to different nervous systems. The maze game involves movement toward a goal, with the speed of movement encoding the reward amplitude. The brightness of the pacman-like object also encodes the reward amplitude. Variations in brightness are appraised more rapidly than variations in speed, and hence the two variables accentuate different aspects of the reward dynamic. Inhibits stop the pacman-like object from moving, but there is no distinguishing whether the object was stopped by virtue of the high or low frequency activity.

The Boxes game reflects the amplitudes of all three bands continuously to the patient, irrespective of whether thresholds are exceeded. Hence this game downplays the importance of threshold-crossing. The addition of the casino element adds reward for maintaining continuity of state above threshold.

The receding Highway game adds the element of history, displaying the reward amplitude over the last 15-20 seconds between the foreground and the vanishing point. In other respects, it resembles the Boxes game, above.

The full 3-d rocket game makes the patient the pilot of a central rocket in outer space, one which is competing with two rockets representing the two inhibit bands. There are also casino game elements that reward continuity of state. Anecdotal data from clinicians indicate that this more complex game is much more engaging to the client than the available alternatives.

Findings

Preliminary analysis was limited to 2d and simple 3d displays only, because of limited data on the complex 3-d display due to novelty. Because of this limitation, only 46 subjects are currently included in this analysis. The data revealed greater improvement on the Symbol-Digit Modality test for subjects who trained primarily on 3d displays, p<.05¹³. As shown in Figures 1 and 2, trends in patient retention, word fluency, and impulse control were also found, however with only marginal significance, p=.10. Subjects who trained primarily on simple-3d displays completed 38.9 sessions compared to 32.3 for 2d training, p=.10.

Figure 1. Effect of game dimensions in neurofeedback training on cognitive performance and sessions completed.

Figure 2. Effect of game dimensions in neurofeedback training on attentional performance.

Conclusions

Since clients were given the option of training with whatever game they like, there is an issue of binning of the populations, since few of them trained exclusively with one game. Moreover, the better results in cognitive training could be partly accounted for by the longer training time. On the other hand, the objective is to achieve results. It is the unambiguous finding of several studies that more sessions lead to better outcomes, hence a commitment to a larger number of sessions has a payoff^14,15. Whereas the objective of this study is worthwhile, more data clearly need to be acquired to firm up the conclusions one is tempted to draw.

The use of 3d displays improved the outcome of EEG biofeedback on one cognitive test, and possibly others, and increased patient retention. Patients agreed to 20 percent more sessions when simple 3d displays were used. As additional subjects and the effect of complex 3d displays are analyzed, we expect the trends in retention and cognitive performance on additional measures to attain statistical significance as well, as there is consistency in trends throughout. It is therefore reasonable to project that implementation of more immersive and multi-modal feedback, leading ultimately to full virtual reality implementations, including realistic portrayals of physiological activity, should enhance patient commitment, comprehension, task engagement, and training efficiency.

NEW BOOKS

Attention Deficit Disorder : A Different Perception by Thom Hartmann

Newest information on this promising approach to treating a variety of psychiatry disorders.

Handbook of Psychological Treatment Protocols for Children and Adolescents
By Vincent B. Van Hasselt & Michel Hersen

Up to date information on treatments of a broad array of psychological disorders seen in children and adolescents.

Handbook of Psychological Assessment
by Gary Groth-Marnat

Step-by-step guide to the classic psychological assessment tests, including Wechsler Intelligence Scales, Bender-Gestalt, Rorschach, MMPI, and projective drawings.

The Prehistory of the Mind : The Cognitive Origins of Art, Religion and Science
by Steven Mithen

How the world of our ancient ancestors shaped our modern modular mind. Were our brains hard-wired in the Pleistocene Era by the needs of hunter-gatherers? When did religious beliefs first emerge? Why were the first paintings made by humankind so technically accomplished and expressive? These and other questions explained.

Handbook of Attachment Interventions
by Terry M. Levy

Glenn Weiner, Ph.D.
Dominion Behavioral Healthcare
703 N. Courthouse Rd, #101
Richmond, VA 23236
(804) 794-4482; fax- 379-7578
 glennweiner@att.net


Jeannine L. Calaba, Psy. D.
4820 Gentry Ave
Valley Village, CA 91607
(818) 763-3361
(moved)