EEG Biofeedback: A Generalized Approach to Neuroregulation

By Siegfried Othmer, Susan F. Othmer, and David A. Kaiser

To appear in "APPLIED NEUROPHYSIOLOGY
& BRAIN BIOFEEDBACK"
Edited by Rob Kall, Joe Kamiya, and Gary Schwartz

Page 9 of 13

Premenstrual Syndrome

Another indication for which EEG biofeedback is very helpful is Premenstrual Syndrome (PMS). This condition is not recognized as a distinct disorder in the DSM-IV, but that is probably at least partially in recognition of societal sensibilities. In its severe form, it is known as Premenstrual Dysphoric Disorder, which is conditionally listed in the Appendix of the DSM-IV (DSM-IV, p. 715). The difficulties with such a listing are, among others, that the symptoms of PMS are so diverse, so highly variable, so subject to 'psychosomatic" influences, so frequently seen simply as an exacerbation of other existing disorders, and so devoid of discernible organic basis. One wishes to blame hormonal shifts, but these are not usually out of line in those suffering PMS symptoms.

The weight of evidence is that PMS is a matter of brain sensitivity to ordinary shifts in hormonal levels. PMS can even be considered as the defining condition for the functionally based "brain disregulation model" of psychopathology. That is, disregulation is the defining characteristic of PMS, and the remedy offered by EEG training is to return brain function to homeostasis and to stability, i.e. to a restored capacity for neuroregulation. Almost no condition remediates as completely and consistently as does PMS with EEG training, and few conditions entail such a breadth of symptomatology. Yet PMS in all its clinical variety is successfully addressed with little more than this straight- forward training. PMS symptoms which have been identified are shown in Table 6 (O'Brien, 1987), and the symptoms which have been observed in our practice, and which have been subject to remediation, are shown with an asterisk. We have no relevant experience with the symptoms that are not marked.

The above results are also non-trivial. PMS symptoms can be disabling in their severity for a significant fraction of women. Yet is has been possible to remediate even the most severe cases encountered. Individual case histories cannot not be reviewed within the prevailing limitations in space, but one example may be given for concreteness: It has been observed that a woman who had a lifetime history of severe PMS, with frequent episodes of suicidality, and with a litany of failed interventions, was able to reach a point within forty sessions of training where she was unaware when her period approached. At the initiation of training, the woman was scheduled for surgery for fibroid tumors. The surgery was never performed. The failure rate in training is on the order of five percent or less for those who follow through with the training until meaningful milestones (20 or 40 sessions) are reached. Some of these failures probably relate to ongoing emotional issues that compromise or sabotage the training. Other cases of PMS are likely sustained by histories of early sexual or physical abuse, and might not remediate with high- frequency training alone, but rather would require alpha-theta training as well.

Medically, PMS is typically managed with anti-depressants such as Wellbutrin. Such pharmacological approaches remain deficient, since the condition is so volatile and variable that no unilateral, long-acting shift in neuromodulator function can offer remedy. It is noteworthy, however, that the EEG biofeedback protocol most commonly employed is also used for depression, and is probably the closest EEG training analog to an antidepressant. Clearly, the EEG training can not only shift the "operating point" of the nervous system in terms of arousal, but also increase the "operating range" over the continuum of behavioral states.

The training does not have to be done during the symptomatic phase of the cycle. This makes it apparent that the training promotes a nervous system capability rather than a particular state. On the other hand, if the training is done during the symptomatic phase, the trainee may experience changes in symptoms literally from session to session, or even during a single session. If at least six sessions of training are accomplished between periods, then substantial relief will typically already be experienced by the time of the subsequent period.

Headaches

Finally, in the category of the conditions most readily remediated we have what are colloquially referred to as tension headaches. Such headaches typically subside within thirty minutes of the appropriate training. Conversely, they can get worse with the wrong protocol selection. With repeated training sessions, susceptibility toward tension headaches can be abated and a person rendered essentially headache-free. Curiously, tension headaches tend to respond to the higher frequency training, as opposed to the lower frequency training that is thought to be more calming. The training in this instance is probably best thought of in terms of increased control of brain states, as opposed to a "relaxation" model.

Attention Deficit Disorder

Next in the order of difficulty and complexity we have ADD, migraines, panic attacks, bruxism, and hypoglycemia. Within the diagnostic category of ADD we also have to distinguish the combined type from the inattentive and impulsive subtypes. The combined type is slightly more complex to deal with, and therefore is placed in the next level of complexity.

Although the dominant application of EEG training is to ADD, it is by no means the easiest to deal with. Perhaps this is due to the fact that the condition is clearly not a unitary phenomenon. ADD is a "dirty" diagnosis. It is so riven with comorbidities that its essence can be obscure. (This is particularly true in the children likely to be referred for EEG training,
who have typically already failed to respond to conventional remedies.) The case has even been made that ADD is a composite of more fundamental disorders, including affective disorders, specific learning disabilities, and a primary disorder of vigilance. (Weinberg, 1992, 1993). In an explicit investigation of comorbidities, less than half of ADD was found to be uncomplicated by diagnoses of major depressive disorder, anxiety disorder, or conduct disorder. (Biederman, 1991) Oppositional-defiant disorder alone overlaps 60% with ADD. And when one also considers Tourette's Syndrome, dysthymia, bipolar disorder (Biederman, 1996), specific learning disabilities, elimination disorders, pain syndromes, sleep disorders, and PTSD, then there remains very little which is not compromised in a significant way by comorbidities which have their own specific implications for EEG training.

Consistent with our model that much of the phenomenology of ADD and its comorbidities is traceable to a modest set of underlying failure modes, it is appropriate to assess the remedy by a means of an evaluation tool which focuses attention at that level. This caused us to eschew the conventional behavior rating scales. Instead, we relied upon a continuous performance test, a computerized test which assesses sustained attention, vigilance, and impulsivity. We chose the Test of Variables of Attention, or TOVA(r) (Greenberg, 1987). This test was favored because it had a demonstrated lack of practice effect, and it has been in common practice for titration of fast-acting stimulant medications for ADHD because of its sensitivity. Thus, if the evidence surfaced by this test was accepted for assessing medications, then it would clearly have to be accepted as a measure of EEG biofeedback as well.

The test is a go-no go challenge that requires only up-down discrimination (which even plants can manage). The test conditions remain invariant for 11 minutes, at which time they change from a stimulus-infrequent to a stimulus-frequent condition. This monotony is a feature of the test, and serves as a challenge to sustained attention. The length of the test helps to assure reasonable statistics on errors of omission, which are taken as a measure of inattention. Errors of commission, which are typically more frequent, are taken as measures of impulsivity. Average response time is measured, as well as variability in response time. The latter is taken as the most revealing measure of ADHD. Results of TOVA testing for 342 subjects are shown for the four subtests in Figure 1.

Mean pre- and post-training results are shown in terms of standard scores for the four dependent measures of the TOVA. The data are segregated by severity of initial deficit for each measure. Standard scores of less than forty are not deemed to be meaningful, and are arbitrarily set at forty for this analysis (four standard deviations below the mean). For inattention and for variability, the data show that the most impaired group (starting score of 40) improved by two standard deviations. In the case of impulsivity, the most impaired group improved by three standard deviations. The effect size is seen to be quite significant. Data are not shown for those whose starting values were >100. Thus the actual number of subjects comprising each graph (as shown) is less than 342.

It is revealing to look at the individual data comprising the data of Figure 1. This is shown for impulsivity in Figure 2. The individual data reveal the consistency with which positive results are obtained. Some 84% of the data points are positive-going despite any test-retest variability, and even though the data include those subjects who test within the normal range. This Figure is proof that we are not dealing with a regression to the mean, if any doubt remained. The entire population moves upward, irrespective of starting point in terms of standard score. This observation demonstrates that essentially everyone is capable of responding to this training. This, combined with the fact that subjects can be readily moved to function above na‹ve norms, disposes of any residual placebo arguments. Some of the small number of cases in which scores declined significantly (beyond expected test-retest variability of perhaps half a standard deviation, or 7 points) may very well have done so in response to the training, as opposed to being "non-responders." The decline may be attributed to choice of training protocol, which may in these cases have been driven by issues other than impulsivity. A different choice of protocol might well have effected a recovery in those cases, but that opportunity does not always present itself in a clinical setting. Some declines in score of course have trivial explanations, such as illness on retest.

ADHD of mixed type requires a combination of approaches used for training the inattentive and impulsive subtypes, and these need to be properly titrated in order to achieve optimal results. For this reason, ADHD of mixed type is considered more of a challenge than the simpler subtypes, and is therefore listed in Table 2 as being of greater difficulty. It is appropriate, however, to incorporate it into this discussion. In addition to evaluations with the TOVA, IQ tests and other tests of cognitive function have been found useful in the past. In an early study of ADHD by our group that has not previously been published in a professional journal, Wechsler IQ scores were measured pre-post. The results are reproduced in Figure 3. The tests reveal the classic pattern for ADHD, namely depressed scores for Information, Arithmetic, Digit Span, and Coding. After the training, the same characteristic pattern is still recognizable, but at a much higher level. The increase in mean arithmetic scores, for example, is quite astounding. Since nothing in the training conferred arithmetic skills, one must attribute the gains to something like increased working memory. These children knew the rules of arithmetic all along. However, they failed in execution. The training allowed them to persist to completion, and to retain in working memory the task they were about. The increase in coding score is more modest. However, closer inspection reveals that a number of subjects did not change at all on the coding test; others made substantial changes. This study was performed early in our work (1990-1991), when a single protocol predominated. It is possible that the lack of progress in some scores is attributable to that paucity of approaches.

The general impression one has from these data is that the training improved level of function broadly. The average improvement in IQ score was 23 points. This change is much too large to be attributable to a test-retest effect, particularly since the retests were done typically nine months after the pre-test (with a six-month minimum interval). Verbal and Performance IQs changed comparably in most subjects. The largest improvements were seen in Picture Completion, which is not seen as a measure in ADD. The least change was seen in Block Design. Verbal and Performance IQs changed comparably in most subjects. This is noteworthy, because in most of them only left-side training was performed. The results imply that the training impinges on inter-hemispheric communication pathways as well.

The three categories of Arithmetic, Coding, and Digit Span together constitute a measure called 'Freedom from Distractibility.' All three also depend on sequential processing skills. The view commends itself that EEG biofeedback increases the 'continuity of mental states,' which manifests itself behaviorally in terms of reduced distractibility, and cognitively in terms of improved sequential processing ability and improved working memory. In support of the contention that

In support of the contention that the training influences function broadly, there is the additional evidence of the Benton Visual Retention Test. Whereas the IQ test showed the group to have been of above-average IQ (107) even before the training, they were in significant deficit with respect to visual retention, as shown in Figure 4.

After the training, some six subjects rated superior, having tested at average or less before the training. Everyone improved with the training, and one subject moved all the way from a defective to a superior rating. This subject unambiguously experienced an improvement in his level of functioning that cannot be explained by non-specific factors. The change is so startling that it does not require the weight of statistical evidence to prove the point.

Improvements were also noted in the tapping subtest of the Harris tests of lateral dominance. These results are shown in Figure 5. When these results are plotted up in terms of the ratio of right-to-left hand performance, an intriguing result obtains. We observe a depletion of mixed dominance and a loss of scatter in the data, as shown in Figure 6. Laterality normalizes. This test is unequivocal testimony to the fact that the training produces change in neurophysiological functioning. First of all, this test is unambiguously scoreable. There is typically 95% concordance between different testers. Secondly, the result was neither expected nor even wished for. (Inclusion of this test in the battery was almost an afterthought.) Thirdly, laterality presumably is not affected by non-specific aspects of the training, such as motivational factors. Fourth, the training itself does not involve any movement of the hands. Improvement in this regard must be ascribed to "central" effects of the training.

Some years ago, it was found that childbirth trauma significantly altered patterns of laterality. The study, published in Nature, examined fetal thumb-sucking and found that before birth, 95% of fetuses preferred their right thumb (Hepper, 1990). After birth, only 85% did so. The shift can be interpreted as an effect of birth trauma, which may bring about a compensatory shift to opposite hemisphere dominance or to mixed dominance when the natively dominant hemisphere has been injured. An appealing suggestion is that the EEG training remediates the functional injury. In this view, the tie-in to ADD becomes more apparent. The functions of vigilance and sustained attention have their own hemisphere-specific mechanisms. When birth injury disturbs hemispheric function to the degree that it impacts handedness, then perhaps it could also impact the management of vigilance and attention. Hence head injury in general, and birth injury in particular, is another confounding variable in the diagnosis of ADHD. This is not surprising. The original research in hyperactivity considered it to be grounded in minimal brain injury.

Click for Next Page