|
It is often difficult for people with autism or Asperger syndrome to decode, understand and manage social situations. Since social behavior is frequently a well-integrated and almost unconscious part of our own behavior, we who are neurotypical* commonly have difficulties understanding and interacting socially with people who have these disabilities. A person with high functioning autism/Asperger syndrome can, of course, learn social rules by observing our social interaction (notice that it is usually they who learn our rules and not we who learn theirs). But because the automaticity with which we learn social behavior does not work for a person with autism, this learning takes much longer. In addition, there are so many implicit messages in social situations that make it difficult to correctly interpret what is happening and thus learn the right associations and rules.
The opposite is also true. Our neurotypical social thinking often functions on an unconscious level, which means that we may find it difficult to interpret what a person with high functioning autism/Asperger syndrome does. Such interpretations often require mental effort and awareness; we make mistakes, in spite of it all, when our built-in interpretation mechanism automatically kicks in.
The easiest solution may appear to be that everyone involved asks as soon as there is something they do not understand. It is often difficult, though, to ask about social rules/social thinking in the actual situation because by questioning, a person reveals that he or she has not mastered social interaction. And if the question is asked of the wrong person, he or she can exploit the weakness that has been exposed.
With this in mind, we believe that it would be of great benefit, both for people with high functioning autism/Asperger syndrome and for those who are neurotypical, to have access to a tool that enables them to study social rules and relations on their own.
Since motivation is an essential and key concept in all types of learning, such a tool should also be fun to use. This places demands on the presentation and design of the tool. An element of excitement should be included, providing quite drastic responses at times depending on what you do. At the same time, it is important that you should always be able to go in and find out what the underlying reasons are for the way different people react. And, of course, it should be possible to choose characteristics and backgrounds of the different people who are participating in a given situation. For example, it should be possible to go in and study how person A's behavior is influenced if she is happy, angry (at someone) or sad, or if she has had too little sleep.
In each scenario it should be possible to interact with different actors, to choose to be different actors and to try and see what will happen depending on what a participant says and does. To increase motivation, the program should probably be designed a bit like a game, but without any external judgments of what is "right" or "wrong". It should also be possible at any time to get an in-depth explanation as to why different actors in the scenario behave the way they do. You should have the choice of going in as a neurotypical person (in which case the opponent is a person with autism) or as a person with autism (in which case the opponent is neurotypical). To add variety, you should be able to choose traits/distinctive attributes and background events for the people involved.
* The word "neurotypical" was coined by people with autism spectrum
disorders as a designation for those with average, normal functioning central
nervous systems, that is, the opposite to those who are neurologically atypical,
such as people with autism.
You can read more about The Social Simulator Project in the final report.
The program as it is presently constructed includes text, sound and images. The user communicates with the program by clicking or using the keyboard. This, of course, limits the simulation, but on the other hand, there are still very many unspoken "rules" that can be brought to light in this manner.
The project started out with more expert system-like, fixed scenarios. In those, the user has 1-3 alternatives to choose from, making more detailed interaction possible. The alternatives, though, are often both limiting and controlling, and it can be quite easy to guess which of the alternatives the program prefers.
Thus, we decided to change approaches and try to find a computer program structure that provided opportunities for the user to "do what he or she wanted" and then let the program respond to these actions. This is a major problem in general, of course. Attempts have been made in the area of AI (Artificial Intelligence) to create intelligent computer systems since the 1950s and it would have been a bit over optimistic to think that we now could suddenly solve the problem in a year. However, we succeeded in finding a solution that works satisfactorily by introducing a large number of limitations. The user, for example, only communicates with the computer through typewritten text or by clicking. The computer only responds with still images and prerecorded sound along with text. Efforts have been made to place the user in situations in which there are not too many actions to choose from.
It turned out that these types of scenarios were felt to be more flexible and in that way could be exciting, but it was extremely demanding to develop them. In addition, it became apparent that the editing software was not adequate for supporting the way one usually describes a situation ("He did this, and I did that, and then this happened. . ." etc.) As a result, new software is in the process of being developed that will better support usage in these kinds of interview situations. We have also reverted to fixed choices, simply because it is too demanding to create more flexible types of scenes from real events. New versions of the program have now been developed and are accessible as beta versions for testing.
You are welcome to beta test our editing program, SimulatorEdit and playback program, SimulatorPlay. The beta version has only been tested using Windows 2000.
SimulatorSetupFiles.zip (bugfix 2, 2002-11-29, 1.18 MB)
Example.zip (O.45 MB) (In Swedish but illustrates how the program works.)
If you have installed an earlier version of SimulatorEdit and SimulatorPlay, it is best to uninstall these. Sometimes the new files are not copied over older files with the same name.
Copy the file SimulatorSetupFiles.zip to a folder on your computer. Unzip the files (a good program for doing this is WinZip. You can download it from http://www.winzip.com) and then click on the Setup.exe. This initiates the installation program and you just need to follow the instructions. Then copy the Example.zip file and unzip that in an appropriate folder (the one you installed the software in, for example). Then you should be able to start the program and test it.
All viewpoints are welcomed!
Questions & comments: Charlotte Magnusson, Tel: +46 46-222 46 96, e-mail:
Charlotte.magnusson@certec.lth.se
Updated: 05-02-17