The Tyranny of Scientific Fundamentalism

There is a curious democracy of facts operative in the field of psychology which I find puzzling as a physicist. When a solid state scientist announces that he was able to put a gate oxide on indium antimonide, and then another one announces that he failed to do so, these two pronouncements do not counterbalance each other. The latter does not serve as an eraser of the former and cancel it out. Now in psychology, a researcher does not even have to obtain a negative result in order to erase the positive results achieved by others; he simply has to be influential and skeptical to erase them. Hence, in some doubting minds, Russell Barkley's skepticism alone, absent any data, can simply erase concrete findings by hundreds of clinicians on thousands of persons with ADHD. This is truly remarkable behavior on the part of scientists. Among academic authority figures, there is a propensity to confuse rigidity with rectitude.

If the results are not obtained via controlled studies, they are deemed to be invalid. Barkley says, with the presumed support of his professional colleagues, that such results are not even worth looking at. This is epistemologically unsupportable. Outcome studies do have scientific validity, and in many cases of interest to us, controlled studies are neither practical nor appropriate. The technique has been helpful to cases of epilepsy, traumatic brain injury, cerebral palsy, and near-drowning victims. It is not possible to establish a good control group for such cases. They are their own best controls. And no one expects them to get better by themselves after long baselines are established. So why are the remarkable recoveries achieved with EEG training not creditable? Psychologists like Barkley are so in thrall to the current, primitive understanding of brain function that they cannot allow themselves to acknowledge new information. Confronted with more and more clinical data on improvements achieved in ADHD children with EEG biofeedback, he is now reported to claim that the effects of training do not last! (Observe the asymmetry: when it comes to discrediting an unorthodox scientific claim, anecdotal data serve quite nicely.) Of course we only have twenty years of history on the method, but that's still a lot longer than Ritalin lasts. Schaef (1992) refers to attitudes such as this as "scientific fundamentalism", the primacy of dogma over good sense. It should be treated with the disdain that it deserves.

Even if control groups can be found, as in the case of ADHD, for example, there are further methodological problems with controlled studies of EEG biofeedback. One popular proposal is to use "sham" training. Members of the control group get somebody else's EEG to train on, rather than their own. This approach has an essential flaw. If the biofeedback being provided is good, then the person involved in it becomes an active part of the process. Moreover, he is very well aware of being an active participant. Such a person would not sit still for very long in front of a display of information relating to someone else. He would soon become aware of the subterfuge and become bored, or irritated. We have actually observed this in a clinical setting, where a child had been training successfully until the software was updated, at which time an error in the software caused the feedback to be gradually more and more delayed. The child was aware of the difference within a matter of minutes.

Hence, the only kind of biofeedback which can be successfully tested in a sham protocol is bad biofeedback. And the only clients you can do it with are the compliant ones who will do anything you ask. That leaves out young, hyperactive kids. The only way these children are able to participate at all is that they quickly grasp the connection between what is going on on the screen and what is happening in their brain. Pacman may grab their interest for the first few minutes, but that novelty wears off. What holds them is the process itself. You cannot fake that.

The second approach recommended for controlled studies is A-B reversal designs. This poses a different methodological problem. Biofeedback cannot be "withdrawn" in the B-phase in the same manner as medications. It is learning, after all. And you cannot ask the brain to unlearn any more than you an unring a bell. One of the persistent and recurring lessons of this field is that once the brain has achieved a higher state of stability through EEG biofeedback, it tends to be self-sustaining until disrupted by further insult to the brain. The brain quite naturally rein forces those strategies which it has found to be successful. Concern about back-sliding, once the brain has consolidated its gains, is largely overdone.

There is an even more fundamental methodological problem with controlled studies of EEG biofeedback. The key driver for con trolled studies of drug efficacy is not merely to control for tester biases, but rather to distinguish the real effect of the drug from the ever-present "real" placebo effect, namely the subject's physiological response to the proverbial "sugar pill". Ullman and Sleator (1986), in a double-blind placebo-controlled crossover study of methylphenidate found 18 subjects (out of 118) who made 50% improvement on teacher rating scales with placebo, which matched their improvement on methylphenidate. On the basis of this finding, they recommend that all children considered for medication be tried on placebo first. The "real" placebo effect is so strong that the best medical studies incorporate a placebo washout period to identify the placebo responders before the subjects are assigned to treatment and control groups. Evidence of the body's own healing response is marbled throughout medical research. It is the skeleton in their closet. It is the noise in their system.

By contrast, self-healing is what biofeedback is all about. For us, self-healing is signal, not noise. Why would we want to control for it? Are there good and bad kinds of self-healing? Biofeedback is not something we do to a person, it is something the person does to support and strengthen his own brain's intrinsic competences for self-regulation. The only burden on us is to determine whether learning is in fact occurring, and we do that by testing. If one wishes to guard against tester bias, one can have the testing done independently. This we have done in all of our studies (with the exception of the T.O.V.A. study, where the computer does the testing and the scoring).

Whereas testing is the way in which we should measure progress and prove ourselves to the rest of the world, the biofeedback therapist does not have to wait for post-training test results to confirm that something has been accomplished. Again, if good feedback is being given, such confirmation comes on a session-by-session basis, as the client (or the parent) reports the changes he observes. Nothing we have learned to date about EEG biofeedback was initially established by controlled studies. It was and remains simply a matter of good, skilled observation.

We have been criticized for making broad claims for the EEG training technique. Our objective was to compel the professional biofeedback community to break out of compartmentalized thinking, to shift from a fixation on specific disorders to a focus on process. Nevertheless, it is the number and breadth of claims with respect to various disorders that elicit skepticism, and the message about process may be getting lost. I daresay that every thing we have claimed to date will eventually be confirmed. However, we recognize that each particular "claim" might not be able to survive scrutiny on its own at this stage. Nevertheless, all the individual "claims" are mutually supportive of the under lying message of brain "responsivity", and of biofeedback as an effective tool for eliciting it. The essential message is that the biofeedback training, seen as a tool with general applicability, should have a favorable impact on a variety of mental disorders which have the common elements recited previously: lack of control of arousal level; lack of flexibility of brain state; and diminished cortical stability. (We make no claims for all the others.) Moreover, with respect to the hope we hold out to people in these regards, the watchword is progress, not perfection; remediation, not cure.

Sterman did not discover some unique feature of epilepsy which allows it to respond to his protocol. Lubar did not discover some idiosyncratic feature of ADHD which renders it susceptible to remediation by training. And Peniston did not discover a unique characteristic of alcoholism that causes it to yield to our ministrations. They all discovered aspects of the intrinsic plasticity of the brain, its ability to learn about itself, which is one of its most basic capacities. The first person to appreciate the generality of the method was perhaps Les Fehmi, who found that the entire enterprise of EEG biofeedback could be understood in terms of how the brain pays attention. Significantly, he is able to elicit similar transformational experiences reported for biofeedback through verbal channels, by employing only the language of attention. By teaching the brain the skill of paying attention, either verbally or by biofeedback, it is impossible not to simultaneously influence autonomic function toward home ostasis.

Without a doubt, the research of Joe Kamiya and Les Fehmi, and the controlled studies of Barry Sterman and of Joel and Judith Lubar were necessary to establish the field. Without them, none of us would have the professional opportunities we now enjoy. However, once a new paradigm is established, inductive methods may be more fruitful. After twenty-five years, Sterman has done research only on epilepsy; and after twenty years Lubars have worked only with ADHD. Building on what has been done, clinicians are now expanding the field with a necessarily more comprehensive vision.

This larger view of EEG biofeedback was not "revealed truth", and it was not innate wisdom on anybody's part. It compelled itself on clinicians by virtue of results that were being obtained with clients. If one is doing EEG biofeedback well, these results happen. They are not always favorable. However, if undesirable results are being obtained, they simply call for redirection of the training strategy. We educate the client to anticipate a range of results. When these results are experienced, the client reports them, and we make mid-course corrections where required. Gradually we proceed from dealing with the most egregious symptoms to the more benign. The client progressively learns about himself, and is thus empowered.

With reference to ADHD, for example, this means that there is no single protocol for ADHD, or even two; there are many. The condition is not monolithic, and there are many etiologies. Answering the question of whether the person is ADHD is just the beginning of our inquiry. With regard to all the subsidiary issues, we are guided both by testing and by thorough interview. The EEG data obtained then adds a final bit of information that may confirm what we have already determined. This the neurologist already knows. In his realm, an EEG can only confirm a diagnosis of seizure disorder; it can never overrule it.

The EEG training is a matter of increasing the person's competence sequentially and incrementally. An analogy I find amusing is to increasing the "flight envelope" of an aircraft: higher ceiling (peak performance); higher g-turns (stress tolerance); and lower landing speed (ability to relax from a vigilant state). This does not entail a single approach, but usually several. As key issues are resolved with a particular protocol, subsidiary issues come to the fore. They will usually require a different approach. When the training is done well, the brain wants this new competence