Autism Help: Robot For Robbie
NASHVILLE, Tenn. (Ivanhoe Newswire) - Parents cherish bedtime stories and games with their little ones, but children with autism can't always focus during those special times. Now, researchers are learning how robots could play a role in helping these kids learn how to pay attention.
With glowing eyes and a high-tech voice, a robot grabs the attention of Robbie Pruitt.
Robbie has autism. Researchers at Vanderbilt University are using a robot to teach him how to focus on people and objects in his environment.
"Try to get a kid to clean up when he can't pay attention to what's on the floor. It's difficult," Sabrina Pruitt, Robbie's mother, told Ivanhoe.
Dr. Zachary Warren said kids with autism seem to be drawn to technology. So robots could be key in teaching social interaction.
"Children with autism are spending much more time looking at the robot than they would be a human counterpart," Zachary Warren, PhD, Clinical Psychologist and Autism Researcher at Vanderbilt University, told Ivanhoe.
The robot directs Robbie to look at a wall monitor. Smart sensors detect that Robbie followed the prompt, so he is rewarded with a video.
The robot also directs Robbie's gaze by pointing.
"We've seen that the children are responding very accurately to the robot prompt," Dr. Warren said.
Tiring work, but Robbie's parents say his interaction skills are improving.
In a study published this year, researchers found that children with autism paid more attention to the robot than the human therapist and followed its instructions almost as well. Researchers say robots could ease the workload of autism therapists and they hope to study whether robots are a cost effective way to accelerate other learning skills as well.
BACKGROUND: Autism spectrum disorder, or ASD, is a range of complex neurodevelopment disorders. People can have various conditions along the autism spectrum, with classical ASD considered to be the most severe and conditions like Asperger syndrome being milder. The most common problems associated with ASD include social impairments, communication difficulties, and restricted and repetitive patterns of behavior. These impairments can make certain behaviors most people find natural, such as making eye contact or interacting with others, very difficult for someone with ASD. (Source: www.ninds.nih.gov)
SIGNS: The signs of ASD appear during early childhood, but some people with a very mild version of the disorder go undiagnosed for years. A few indicators of ASD are listed below:
- Not responding to their own name.
- Seeming to be unaware of others' feelings.
- Speaking with abnormal tone or rhythm.
- Doing repetitive motions, such as rocking or spinning.
- Developing specific routines or rituals. (Source: www.mayoclinic.com)
TREATMENT: Autism spectrum disorders cannot be cured, but some symptoms naturally improve as people age. The disorders are often treated using behavioral interventions where specially-trained therapists aid children in developing social and language skills. Medications for anxiety, depression, and OCD have also been used to treat people with severe behavioral problems due to an ASD. (Source: www.ninds.nih.gov)
NEW TECHNOLOGY: A specially-designed robot named NAO could help autistic children learn how to coordinate their attention with other people and objects around them, a basic social skill called joint attention. NAO was originally a commercial humanoid robot created in France, but engineers from Vanderbilt University changed the robot's control architecture by adding an adaptive structure they designed themselves called ARIA – Adaptive Robot-Mediated Intervention Architecture. The robot is programmed to give children either verbal prompts or gestures such as pointing to make them look at a video or image displayed on one of several mounted screens in a room. The room has multiple web cameras to track the child's head movement, and the child is given a hat with LED lights, so the computer can estimate what direction they are looking in. In a test, researchers found children with autism spent more time looking at the robot than a human therapist, proving NAO could be an effective tool when used in conjunction with a human therapist. NAO is also designed to be adaptive, so it behaves differently depending on the child's responses. The researchers hope NAO may also be programmed to help children with autism learn other important skills, like sharing and imitation. (Source: www.news.vanderbilt.edu)
Zachary Warren, PhD, Clinical Psychologist, Autism Researcher, and Associate Professor of Pediatrics at Vanderbilt University, talks about how robots are helping kids with autism.
How is a robot helping you guys when it comes to autism therapy?
Dr. Warren: We have known for a while that young children with autism often seem to display a preference for interactions with robots and technology. They also show profound challenges regarding social skills and communication. What we have tried to do is to use that advantage towards technology to help with skill deficiencies in terms of social communication.
What is it about the robots that they are so engaged with?
Dr. Warren: There is something about the movement, the sound, and the experience of the technology that is really hard to quantify. However, for whatever reason, across many different samples of children with autism, we see this preference for some young children to attend to robotic and technological movement rather than common social presses, like responding to their name when their parents call.
So, how did you come up with this final product?
Dr. Warren: For the past several years we have been collaborating with our team of engineers here at Vanderbilt University with thoughts of how could we apply their sophisticated adaptive technological platforms to the field of autism research. We were essentially thinking that if we have this ability to manipulate robots and create sophisticated systems and we have expertise in early development and autism spectrum disorders, we could potentially be able to put those together to see if we can make a difference.
How does it work?
Dr. Warren: Many studies have examined how children with autism generally respond with robots in the room. Do they look more frequently? Are they interested? Are we seeing them more engaged? This work tells us about the potential of robots, but we wanted to develop a system that was smart enough to start responding on its own. We wanted to develop a robotic technology that could really understand where that child was looking and how they are responding to the environment, so that we could improve those basic skills.
What kind of basic skills have you seen them help children with?
Dr. Warren: We focused on the area of joint attention. Joint attention refers to a concept of being able to share attention with others, such as understanding where another person is looking or directing others to look at a particular object. It is an early critical skill that relates to language and language development over time in typically developing kids, as well as kids with autism. So, we thought we could potentially use this technology to shift or accelerate joint attention skills.
What exactly does the robot do? Does it engage with the child?
Dr. Warren: The robot is essentially interacting with the child in a room that has many different monitors embedded within it. So, the robot starts to do things like call the child's name, shift its gaze to look across the room, or it points to a target within the room. The robot understands in real time where that child is looking. So, when the robot understands that the child is looking to the correct target, it actives that target and reinforces that look. And then tries to get the child to look to another aspect of that same environment.
So, does it talk to them as well?
Dr. Warren: It does. It provides a little bit of verbal praise as it goes, but really we are relying on some of these common reinforces that most preschoolers would respond to; such as, short segments of a funny cartoon clip or something like that to reinforce where they are looking in their environment.
So, is the robot looking first and trying to make the child look in that direction?
Dr. Warren: Correct. We designed the system to be able to scaffold or build skills together. Ultimately, we wanted to have an environment that would get the child's attention whether they were very interested or not. So, we wanted to get it to a system where the robot could just subtly sort of move its head to the target and the child would be able to follow. Children with autism have early difficulty regarding joint attention skills, and that real difficulty with translates into difficulties with language, with communication, social skills, and cognitive development. So, focusing on this area might be a core or pivotal skill that could impact many other skills over time.
What have you seen so far? Are there any results at this point?
Dr. Warren: It has been really interesting. We have seen that some children with autism are spending much more time looking at the robot then they would be at a human counterpart. Also, we have seen that the children are responding very accurately to the robotic prompts. So, they are looking when the robot prompts and we are seeing that this actually translates into being able to look at some of the subtler areas of prompting; maybe starting to look just with a name or a gaze shift rather than with a full point or activating the target.
So, what are the benefits?
Dr. Warren: If we can impact joint attention skills that are necessary for many human interactions outside of this sophisticated environment, we may see generalization into those other areas of impairment. We may see generalization towards early social communication development, language development, and later cognitive development. It might be something we could think of as an accelerant technology that boosts skills across many different domains.
Have you seen anything that encourages you towards that end?
Dr. Warren: Yes, we have seen that. We have seen within system improvements over time and what we would like to have is systems that can intelligently respond across very sophisticated and subtly adapted aspects of the environment to promote these skills in more powerful ways.
How long did it take to develop that, to get it to where you wanted it to be?
Dr. Warren: It took a lot of work, not necessarily to design the task or to design the robot itself, which is a totally different team, but to be able to design a closed loop technology. It's a technology that is smart enough to understand what you are doing in the environment.
Does it have a name?
Dr. Warren: The robot is called Nao.
Is this in the study phases or something that is available to everyone?
Dr. Warren: So, we have run a couple of studies with this right now and it is not necessarily widely available. What we would like to do is run a larger trial with this to really understand if it can truly accelerate these skills in a way that we want it to. If it can we would want to think about ways to make such an accelerant technology available to many children with autism.
Is this something you would use at home or is it used with a therapist?
Dr. Warren: That is a very good question. One of the very real questions with this is how can we really effectively harness robotic technology to improve the lives of children with autism? Robots are expensive, but many of the intensive behavior and educational interventions that we recommend are much more expensive when played out over periods of time. So, there may be ways where we could utilize robotics in a cost effective way to accelerate skill learning or prime learning for kids with autism.
How long down the road until you think you get the results that you are looking for or maybe see this more widespread?
Dr. Warren: One of the developments we would like to see take place is really moving towards noninvasive systems that are able to capture aspects of performance. In this study, what children did is they wore a small hat with lights on it that could basically triangulate wherever they were looking within that environment. The problem with getting young kids, including young kids with autism, to wear hats is that they do not like hats, right? So, what we really like to utilize is computer vision technology that can really sync up with where you are looking without having to wear a device and have that environment respond to that same technology.
FOR MORE INFORMATION, PLEASE CONTACT:
Vanderbilt University Medical Center