Autism Diagnosis and Therapy Using AI?

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition characterized by difficulties in social interaction, communication, and repetitive or restrictive behaviors. It is a spectrum condition because it affects individuals differently and to varying degrees. Globally, ASD is a significant public health issue with both social and economic impacts.? In the context of the developing world, the challenges associated with ASD are amplified due to factors such as:

• There is often a lack of understanding and awareness about ASD, leading to late diagnosis and intervention.
• Misconceptions and stigma surrounding ASD can lead to social exclusion and discrimination.
• Many developing countries lack the necessary resources, such as trained professionals and specialized educational facilities, to support individuals with ASD effectively.

Finding Clues for Early Diagnosis Using AI: Deep learning offers promise in identifying autistic patterns. It can analyze data and potentially provide evaluators with confirmation of a diagnosis or suggest the need for further evaluation. Sometimes is tricky to know if someone has autism early on. There's no single defining trait, and some autistic behaviors can occur in other conditions. Evaluators might hesitate to diagnose without absolute certainty, leading to delays in getting crucial therapies and support. Companies Pioneering AI-based Autism Diagnosis include two companies:

• Behavior Imaging uses an app called the Naturalistic Observation Diagnostic Assessment. Parents upload videos of their children, which clinicians then review for diagnosis. Additionally, Behavior Imaging is developing AI algorithms to analyze behaviors and potentially flag relevant patterns for human evaluation.
• Cognoa offers a mobile app for parents to screen children for autism. It uses multiple-choice questions and video analysis, but doesn't provide a definitive diagnosis.

Robots in Autism Therapy offer a safe and predictable environment for practicing social skills like facial expression recognition and appropriate responses. These robots are specially designed to help kids with autism understand emotions and social cues. They can change their faces to show as happy, sad, or surprised, just like people do. By practicing with robots in a safe space, kids can feel more confident talking to others.

• The SoftBank NAO robots are humanoid robots used in therapy. They can express emotions and adapt their behavior based on the child's responses. Many autistic children respond positively to NAO due to its non-judgmental and patient nature.
• MIT Researchers programmed robots to personalize interactions with autistic children based on data from video, audio, and physiological measures. This allows the robot to tailor its responses to the child's current state.
• A Denver startup, Manatee, is developing BiBli robots designed to guide children through challenging social interactions at their own pace. The robot can provide support without judgment and collect data to share with therapists for further insights.

Apps for Autism Therapy apps are less expensive (compared to robots) and can be easily integrated into various settings. For instance, the?Manatee app?is one of the first (no cost) AI apps offering a step-by-step list to move from easy to more advanced skills, and guidance with substantial focus on parental involvement. However, while these apps support children as they learn, they presently have several limitations:

• They often are limited in scope and focus on teaching specific skills like communication, facial recognition, and eye contact.
• Children using the apps need to be able to read instructions, be motivated to participate, and have a certain level of independent functioning.
• It's unclear if skills learned by interacting with apps can be transferred to real-world social interactions with humans.

Integration and Future of AI in Autism Care

GenAI can be the basis of potentially valuable tools for supporting early diagnosis and personalized therapy for autistic individuals, and thus a great addition to a child's therapy toolbox. However, it's important to remember that AI can't diagnose autism on its own, but it can offer valuable insights to doctors and a helping hand to parents and (limited) therapy for autistic children.? Furthermore, I feel that parents of an autistic child, often experiencing frustration, sadness helplessness, or anger, can also be helped with AI tools. Perhaps no one has paid much attention to this latter type of tool, yet!