Mind Hacks points to a good article on the current state of research into the effectiveness of biofeedback on ADHD. Its conclusion that biofeedback is a promising but unproven treatment is fair one.
What is interesting about the article is that it talks about what makes a good research study and delves into the problems of having a control group. In medication trials the control group takes a placebo pill, an inert sugar pill, whilst the study group takes the real medication. Both pills look the same and none of the medical staff handing out the pills know who is getting what. This minimises the confounding variables in the experiment.
In biofeedback studies, the problem is what to do you do with the control group. The treatment typically consists of 30 minutes a day using a computer and a biofeedback device. The control group needs to have a dummy treatment that is identical so that no one knows who is getting the real treatment but yet is guaranteed to have no effect. By its nature, biofeedback is an interactive process so the control group must have some sort of interactive experience. A device that just randomly responses to the the biofeedback would be quickly spotted.
This problem, when applied to a treatment such as the Dore program, becomes even more significant. Dore is series of exercises forming a 12 month program of twenty minutes a day. What activity could be used as a placebo that isn’t immediately recognised as the placebo? The only possibility would be to give one group the proper Dore exercises whilst the control group gets a random selection of exercises. But, the random exercises are too much like the real thing and there is a chance they help develop the cerebellum. Though they certainly would not be as effective it would introduce an unknown variable into an experimental set-up designed to remove unknowns.
A secondary problem is that children are assessed every six weeks under Dore using sensitive balance and eye-tracking devices. The child gets regular empirical evidence that the treatment is having an effect long before any improvements are seen in academic work. Obviously the users of the placebo should not see any improvement and this could have a significant effect on the child’s general confidence.
There is a way to conduct trials in these situation. You have two treatment groups, A and B plus a control, Group C. Group A gets the treatment whilst Group B gets a different activity, for example one-to-one help for 20 minutes a day. Group C gets no treatment. After Group A has received the treatment, the groups switch so Group B get the treatment and Group A gets the other activity. Once Group B has completed the treatment the experiment ends.
Both Groups A and B should of made more academic progress than the control Group C because of the one-to-one help they received and confounding variables such as the Placebo and Hawthorne effects. If the treatment worked better than the one-to-one help, then Group A would be expected to show academic improvements during the first period greater than Group B. This progress would slow down in the second period after the two groups switched roles and Group B would catch-up. If the treatment did not work, Group B would be ahead after the first period thanks to the one-to-one help and Group A would catch-up in during the second period.
There are some obviously difficulties in using this experimental design with Dore. Firstly it is a year long treatment. That means the whole experiment will last two years. To allow for people to move schools or drop out without having a major impact on the statistics, each group needs to be quiet large. Given that for the two treatment groups, there is a lot of investment of time in doing the exercises or taking children to the one-to-one sessions, the drop-out rate is likely to very high. Each group would probably need to start with about 50 people.
The cost of all this is significant. One-to-one teaching everyday for 50 people for year, twice, won’t be cheap. Nor will provision of the Dore treatment. Overheads in managing the experiment, tracking the academic performance of the children all add up. A gold standard experiment like this costs tens of thousands of pounds. Of course if Wynford Dore pays for the experiment then it won’t be an independent study but no one else will fund the experiment. The only dyslexia research body with that sort of funding is the department at York University. It is run by Professor Snowling who is rabidly opposed to Dore so its unlikely that any funding will come from there.
Another problem with an experiment on Dore is an ethical one. Asking a child to take part in a drug trial for four weeks is OK because if the treatment has no benefit then the child isn’t effected in anyway. For a Dore trial, the child has to spend twenty minutes a day for a year. If the treatment doesn’t work then the child has lost a huge amount of time and effort that could of been spent on more established therapies and the child would of fallen even further behind academically. This is major problem as the basis of any ethical experiment is that in no way, regardless of the experiment’s outcome, should the subjects experience any detrimental effects.
There are good reasons to criticise Dore’s scientific research and similar research by other alternative treatment but it has to be seen in context. The practicalities of an effective study that proves in one go the treatment works are both difficult and expensive. Dore and others are stuck in a catch-22 situation where if they pay for research there will be immediate accusation of bias but if they don’t pay for it, no one else will. Finally the ethical issues make the whole feasibility of an experiment doubtful.