Several contributors have pointed out that there is evidence about a relationship between brain structures and functions, as well as about the effect of the environment on the developing brain. Of course this is perfectly true.
My concern is not about this established evidence, but about the very loose logic that jumps from evidence about learning to the assumption that specific aspects of the brain have been changed by experience. Such changes MAY have occurred, but they need to be measured independently before we decide this. One of the reasons I am concerned about this faulty logic is that I believe it is often deployed for commercial purposes, suggesting to the audience that interventions are much more empirically-supported than is the case. I am not alone in this concern, by the way. Here are two interesting discussions of the matter as it applies to education:
Hirsh-Pasek, K., & Bruer, J.T. (2007). The brain/education barrier. Science, Vol. 317, p. 1293.
Willingham, D.T. (2006). Brain-based learning: More fiction than fact. American Educator, Vol. 30(3), pp. 30-37.
It is simply not enough to show that an existing brain mechanism may be at work when a proposed intervention is used. The intervention needs to be shown to be effective-- not necessarily easy to do, as I have pointed out before. Then, if there is a reason to consider other applications of the same brain mechanism, it is necessary to show that the intervention was actually a use of that mechanism. "Confident speculation" is not evidence for either of these possible conclusions. Regrettably, many customers for commercially-available intervention packages are easily persuaded by statements about brain development that may in fact be completely irrelevant to the issue.
An example of these problems is the FastForWord program for treatment of dyslexia. This program is thoroughly grounded in excellent science. The basic idea of the program is that reading involves making a connection between a heard sound and a seen symbol. Children who have difficulty processing sounds will obviously have more trouble with this. To process sounds accurately, certain brain processes must occur extremely quickly, because there are very rapid changes in sound pattern during the course of what seems to be a single speech sound. It is possible to train people to run brain processes more quickly than they naturally do. A video game makes it possible for children to practice sound processing and speed up their abilities, therefore making their discrimination of sounds more accurate and improving their ability to connect sound and symbol, and therefore to read better.
All of these pieces of the puzzle are based on good evidence. So, is FastForWord an effective program? http://www.fcrr.org/FCRRReports/PDF/FastForWordLanguage.pdf provides a number of statements about this, but on examination we see that most of them involve descriptions of the postulated mechanism or of related but not identical interventions. What is the problem? Once an intervention has become proprietary, the owners do not wish to release information about it for the use of business rivals. This means, of course, that the information is not available to anyone who might do an independent replication of the intervention's effects, nor can there be a report of the owner's assessment in a professional journal, which would require transparent reporting of the nature of the treatment. (Some writers on this subject have even suggested that no proprietary treatment can ever be properly referred to as "evidence-based" .)
I would not dream of saying that nothing affects the brain. Obviously we use medications that affect mood, behavior, and neurotransmitters. Studies of deep brain stimulation (G. Miller, "Rewiring faulty circuits in the brain", Science, 2009, Vol. 323, pp. 1554-1556) are currently exploring ways to create major changes. One question I raise is whether it is appropriate to claim that sensory or cognitive events can cause brain changes of a type that it would now be possible for us to measure. I believe it is inappropriate to make this claim, and perhaps even more inappropriate to claim that an intervention must be effective because a known brain mechanism can be postulated to be relevant. Making these claims in a commercial context is an especially worrisome practice that is becoming ever more common.
The problem of establishing an evidence basis for treatment is a broad one, not limited to claims about brain plasticity. There is further discussion of this issue in an article I have written with Randal Pennington, Monica Pignotti, and Linda Rosa ("Dyadic Developmental Psychotherapy is not ‘evidence-based' ", Child and Family Social Work, 2009), which is now on line at http://www3.interscience.wiley.com/journal/120735670/issue.