Towards Designing An Interactive and Intelligent Tool for Social Skill Development of Individuals with HFA

Background:

Research and anecdotal evidence suggests that individuals with high functioning autism (HFA) respond well to social skills instruction and related applications, such as Social Stories. These techniques can be used to introduce socially appropriate behaviors in specific circumstances. However, individuals with ASD have a hard time adapting to unexpected events (e.g. when things “go wrong”). They often respond to these events with behavior that is not socially acceptable or age-appropriate. These inappropriate behaviors limit their social opportunities, which negatively impact their quality of life.

Objectives:

We see an opportunity for technology to provide custom generated scenarios for the individual to experience social situations and appropriate responses to unexpected events. Technology can also assist and support reflecting on those experiences. Finally, it can gather data to assess the individual’s skill level and adjust the scenarios appropriately. A long term goal is to have the assessment and scenario generation automated via artificial intelligence.

Methods:

A pilot study was conducted in the (8 boys/2 girls, age 7-12, IQ 67-99). Children were presented with a bridge crossing scenario through an interactive computer animation (ICA, completed twice) and were then presented with a paper prototype (PP, completed once) to retell the experience. All sessions were video-taped.

Results:

In the ICA study we found that six out of nine children were aided by the rehearsal feature. In the PP study the children used different perspectives to tell the story, and used a rich set of interaction modes with the given pieces. Two different types of prompts were required: One to continue the story and one to elaborate on the details.

Conclusions:

After a more in depth literature review and input from experts in the clinical field, we are now starting the second iteration of the project. The first step is to prototype a set of scenarios that will connect to the personal experience of our defined user group, which will run in a software demo. For the reflection method we will develop a set of puzzle pieces which contain the same pictures as seen in the scenario. These puzzle pieces can then be sequenced in the right order, providing visual feedback via lights, as a reward and as a prompt to continue the activity. In addition, sound modules will contain the dialog of the scenario or allow for the user to record short messages with each picture.

We see an opportunity for technology that may greatly benefit important skill development for high-functioning children with autism. We concluded that it is possible to create an ICA with the flexibility of the PP in an intelligent system that will provide the necessary prompting. With our system there is the freedom to practice different outcomes, learn proper behaviors by increasing in levels and generalize the learnt skills to other contexts. We will be conducting a pilot study of the new system in early January.