Can we use brain signaling to improve ABA?

This article from Wired describes a very interesting development in neuroscience that could be harnessed to improve teaching of autistic children (and, of course, NT children as well).

The essence of the discovery is that there are signals, prior to the application of a stimulus (in this case, something to be memorized), that seem to be correlated with the likelihood that the stimulus will in fact be remembered.

Of interest, and of course complication, is that this experiment used a magnetoencephalograph, or MEG, to measure the effect. MEGs differ from fMRI (functional magnetic resonance imagery) in the time horizon on which they operate. MEG measures magnetic fields created by neural electric impulses directly whereas fMRI uses the effect of the changes in iron content in blood flows as the basis for its measurement. Becaues of this difference, MEG can identify events that are approximately 20 times shorter in duration. That may explain one reason why this effect has only recently been discovered.

At first, the idea that something other than fMRI was used is interesting because fMRI requires a large machine to operate; it turns out after a bit of research that the MEG requires a magnetically shielded room! At least EEG can be done directly with external electrodes.

It seems that we must wait a while for either better magnetic shielding or a “translator” between MEG-discovered effects and EEG results (assuming it’s even possible) to envision the creation of a system, like a simple hat or non-superglued electrodes to track this information directly. Alternatively, the suggestion is made in the article that biofeedback techniques (which coincidentally require less complicated physical environments) could be used to increase or control the theta brainwave oscillations detected by the MEG in the underlying experiments.

Given the dramatic difficulties in teaching autistic children, from attention deficits to so-called co-morbidity issues such as other learning disabilities or mental retardation, any improvement in matching instruction to results could have a disproportionate impact on the amount learned by these children. Certainly, NT kids could benefit from the same techniques, but I imagine many parents will be unlikely to send their child to school with electrodes and a ponytail! Even intra-session teaching could be improved, if a therapist can be warned of the impending situation soon enough to present the optimal stimulus.

In my financial analysis work, we often use a functional analysis (brief example) to get an immediate understanding of a business’s real model for turning the opeational components of revenue into free cash flow. Here, the function that describes ABA learning may be briefly described as:

learning = learned/presented * presented/learning opportunity * learning opportunities/hour * hours/therapist * number of therapists.

(This equation should not be considered our final thinking about this; it’s just too important of a concept for its presentation to be delayed. The perfect is the enemy of the good.)

In this context, the ability to more closely match learning and presentation because of foreknowledge of improved learning conditions would improve the second factor, the ratio of presented items per learning opportunity. If we could teach children or otherwise increase the number of these improved conditions, we would improve the third factor, learning opportunities/hour. Why does breaking this process down help? It’s because we can more appropriately allocate our resources to improve the desired result. For example, at some point the length of a single ABA session becomes too long for a therapist (for the children, breaks seem to alleviate this problem more often than for the therapists!). At that point, it becomes necessary to find additional therapists to provide additional hours of therapy, rather than continuing to increase the session length.