Friday, March 29, 2019

Two Observations Impacting Robot-mediated Behavior Intervention Therapy in autism

Two Discoveries Are Impacting Robot-mediated Behavior Intervention in autism Therapy
Like any new advance into an emerging discipline, there are unexpected discoveries. I have been studying the elements of robot-based, mediated intervention therapy to try and understand and discover the why? What works and what doesn’t.
Let’s set the stage. What is robot-mediated behavior intervention? It is the use of robots as an assistive technology in delivering autism intervention therapies. This further means, that robots are used to engage, present and deliver robot executed behavior therapy routines presented by the robot to those diagnosed on the autism spectrum. These routines address defined and targeted aspects of social interactions and learning skills development. Robots are particularly well suited to the task. In the general case, robots are engaging, they don’t tire, they don’t get off track, they don’t emit frustrations and they are not prone to giving off facial or verbal stimuli that can be perceived and thus received as upsetting. In summary, they are typically found to be non-threatening. Multiple University based research programs attests to these attributes. And, they can be programmed to embrace ABA and EBP guidelines. But seemingly more is going on.

There are two elements that are coming into focus.

First, in a typical therapy session usually conducted at the outset in an environment somewhat uncomfortable to a child, there is the distraction of an unfamiliar environment, new faces, ‘clipboards’, clocks, image cards, and therapist interactions, etc. used to facilitate the therapy. What I am seeing is that when a robot is introduced and utilized there seems to be an unlocking, an establishment of ‘engagement’ facilitated by the robot with the therapist now viewed as a friend and partner in the exercise. Kind of like “we are doing this together”. I would postulate that the therapeutic environment shifts from one of being ‘I am a target’ to that of a social exercise engaged with a friend.

Second, and I know this will seem as heresy, but most robots attempt to present a ‘face’ through which emotional expressions are conveyed in an attempt to mimic human social interactions. In and of itself, not a bad idea. However, recent research has pointed to a factor that I always wondered about. Why are some robots seemingly more engaging than others. We see attempts at using robot models that project faces that are cartoon-like, animal-like, computer screen like, puppet-like, used to crack the child engagement barrier. But as I mentioned University research has postulated that in many children diagnosed on the spectrum there is a delay between what is seen and what is being heard. The experience of watching these robots is like watching a T.V. show where the voice is not in sync with the image being projected. (See: The Journal of Neuroscience/ Vanderbilt Study). Animal based robots help alleviate some of this ‘unsycronization’. I now believe this is why I have always been biased to the NAO robot’s interactions in robot mediated behavior interventions. The NAO has no moving ‘lips’ or ‘facial’ expressions. The child only hears and does not have to mitigate the synchronization.

Put these two concepts together and I sense a combination that has impact and traction. I know this will be challenged by some and maybe even many but based upon my practical, field-based experience I am comfortable that both concepts have merit. I will let formally trained researchers solve for a ’proof’.

With Stanford Researchers recently noting that there is an average wait time for an autism specialist of 18 months and the cost of therapy per child per year is estimated at $40-60 thousand dollars, I just want to help these children, young adults, and their families. Robots can help and be positively utilized to be assistive in this challenge.

Michael D. Radice is Chairman of the Technology Advisory Board for ChartaCloud Technologies, ChartaCloud Robotics, and