Autism in children can be dangerous as there are frequent portrayals of aggression. And it may occur in public places like parks and restaurants too. This is the reason that parents of these children restrict them in the four walls of their house but this can have an adverse effect on their development. Matthew Goodwin, a behavioural scientist in Northeastern University researched ways in which parents can predict such behaviours in advance.
According to Goodwin, it is important to understand the level of anxiety, stress, and arousal in the children. However, this can be difficult as people with autism cannot always express what they feel and therefore, Goodwin with his colleagues devised biosensors, mobile app, and machine that learns algorithm. This system predicts the aggressive behaviour in autistic children. He made this discovery with the help of 20 autistic children in a psychiatric hospital. The wearable biosensor had a clock in it which would record each aggressive episode and match it with the bodily function.
The data collected was then put in a machine-learning algorithm. He also observed that behaviour of a child in the past can actually predict his behaviour in the future. So the machine can predict whether or not the child will show aggression. This can prove to be a major success for parents and clinicians. Parents can clear the room in order to prevent the danger or engage their child in calming activities.
Goodwin calls these strategies similar to but an early version of Alexa or Siri. He received his funding from the Department of Defence and looks forward to working more to improve the predictions of the machine. He also thinks that there needs to be more research and programmes dedicated to children with autism. Therefore he says, “I do think it needs more attention and innovative approaches”.