|Intelligent Digital Technology to Advance Treatment, Procedural Fidelity, and Employment for Neurodiverse Individuals and Caregivers|
|Monday, May 30, 2022|
|10:00 AM–11:50 AM |
|Meeting Level 2; Room 254A|
|Area: AUT/OBM; Domain: Service Delivery|
|Chair: Donald A. Hantula (Temple University)|
|Discussant: Donald A. Hantula (Temple University)|
|CE Instructor: Donald A. Hantula, Ph.D.|
Rapid advances in intelligent agent technology and artificial intelligence technology present new challenges and new opportunities for neurodiverse individuals, their caregivers, and the professionals who work with them. The COVID-19 pandemic became a strong motivating operation for integrating these technologies into work with neurodiverse individuals and their caregivers. This symposium shares groundbreaking new work on applications of intelligent agent technology and artificial intelligence technology with neurodiverse individuals from pre-K to adulthood. Drawn from research conducted by a NIH and NSF funded network of academic and private sector researchers across several states, four illustrative examples show how these digital technologies have been incorporated into work with neurodiverse individuals. One presentation shows how an intelligent agent platform can enable families to access expert guidance in implementing home-based behavioral treatment for children with ASD. A second demonstrates how a similar platform can be used to instruct school age children with ASD. The third presentation evaluates how an intelligent agent based tool increases caregiver treatment fidelity when teaching children basic living skills. The final presentation discusses design considerations for developing a platform that will enable neurodiverse adults to work as data annotators in the IT industry. Although all presentations will review data, there is an equal focus on issues of usability, acceptability and reaction from caregivers and employees using the intelligent agent technology. The overarching theme for this symposium is that forward-thinking applied behavior analysts can help create and leverage innovative technologies to assist the neurodiverse individuals, their families, and the professionals who work with them succeed.
|Instruction Level: Advanced|
|Keyword(s): Cargiver support, Intelligent agent, Neurodiverse emoployees, Treatment fidelity|
|Target Audience: |
Advanced: prerequisite skill/competency would include experience in supervising work with children with ASD and/or the work of neurodiverse employees; some basic educated lay person familiarity with AI, experience with efforts to improve program fidelity
|Learning Objectives: Learning objectives: At the conclusion of the presentation, participants will be able to: (1) discuss how an intelligent agent can improve caregiver provided ABA therapy; (2) identify how an intelligent agent can be used to improve instruction with school age children and its supervision; (3) describe the concerns of neurodiverse employees working in data annotation jobs.|
Supporting Caregiver-Delivered Behavioral Intervention for Children With Autism With an Intelligent Agent Platform
|ALIYA YAGAFAROVA (Auburn U), Corina Jimenez-Gomez (Auburn University), Cecelia Drummond (Auburn U), Emily A Phaup (Auburn U), Donald A. Hantula (Temple University), John T Nosek (Guiding Technologies)|
Caregiver involvement is an integral component of behavioral interventions for children with autism spectrum disorder (ASD) to ensure generalization and maintenance of skills. Coaching and supporting caregivers in the implementation of behavioral interventions often requires closely working with behavior analysts, which can be resource intensive and may not be feasible in some settings (e.g., rural communities). Further, ensuring sustained treatment fidelity may require additional monitoring and re-coaching. An intelligent agent platform that coaches and guides caregivers in the implementation of behavioral interventions may be useful for minimizing resources required to support caregivers and may aid in maintaining high, long-term treatment fidelity. Caregivers of children with ASD receiving services at a university-based clinic were recruited to deliver behavioral interventions in the home under the direction of written instructions or an intelligent agent platform. Treatment implementation fidelity, percentage of correct responses by the child, and acceptability of each support system served as the main dependent measures. Caregiver acceptability of the intelligent agent technology is discussed.
Evaluating Intelligent Agent Technology for Acquisition and Instruction of Hand-Washing in Children With Autism
|ELIZABETH R. LORAH (University of Arkansas), Brenna R Griffen (University of Arkansas), Donald A. Hantula (Temple University), John T Nosek (Guiding Technologies)|
Intelligent agent technology can help improve procedural fidelity and maintain high levels of performance by ABA therapists and the clients that they serve. However, this nascent technology has only very recently been introduced to ABA therapy. Much remains to be learned about intelligent agent technology and its effects on ABA therapists, including social validity and acceptability This study evaluated the use of an intelligent agent and data collection system for ABA therapist use while teaching handwashing to three school aged children with a diagnosis of autism. Using a multiple baseline design data were collected on therapists’ fidelity of implementation, as well as child acquisition of handwashing. Data were collected until the child participants mastered the ability to independently demonstrate washing hands. Following this, therapists were given the option to either continue with the app or use a traditional paper based protocol and data collection method for the purposes of handwashing instruction as a test of acceptability. The results of therapist reactions, as well as implications for digital technology and intelligent agent use will be discussed.
Intelligent Agent Technology for Caregiver Treatment Fidelity and Life Skills of Children With Autism Spectrum Disorder
|KAORI G. NEPO (NeurAbilities), Donald A. Hantula (Temple University), John T Nosek (Guiding Technologies)|
Behavior Analysis provides effective interventions for individuals with ASD and other developmental disabilities. However, the shortage of qualified and trained professionals to implement such interventions has been an ongoing problem nationally and internationally over the past decades. This study evaluated the effect of an innovative intelligent agent technology (GAINS) as a new tool for the caregivers of children with ASD to teach them important life skills such as dressing, feeding, hand washing, or packing a snack. Children with ASD (ages between 3-8) who receive services from a regional behavioral health organization and their caregivers participated. A single subjects design across participant dyads was used to evaluate independent completion of the task analyzed target life skill of the child and treatment fidelity data of caregivers following the cues from the intelligent agent technology. Both behavioral data and participant reaction and satisfaction with the intelligent agent technology were assessed. Recommendations for designing intelligent agent technology for use with this population are discussed.
|Design Considerations for Building a Platform to Enable Neurodiverse Employees to Work in Data Annotation|
|ELIZABETH GARRISON (Temple University), Slobodan Vucetic (Temple University), Eduard Dragut (Temple University), Matthew Tincani (Temple University), Donald A. Hantula (Temple University), Ray Hong (George Mason University)|
|Abstract: Neurodiverse individuals often struggle to obtain employment. With the surge of large-scale data-driven innovation in Artificial Intelligence, data annotation tasks found in Amazon’s MTurk and similar platforms have presented significant employment opportunities for neurodiverse individuals. We recruited nine neurodiverse and ten neurotypical participants between the ages of 18-30 and built an interactive web-based training platform to determine when, how and why the annotation performance and their perception of images and text tasks vary between the two groups of participants. After we collected data using our platform, we conducted additional semi-structured interviews with neurodiverse participants to gain a deeper understanding of the reasoning for their particular responses. Our approach (1) highlights differences between neurodiverse and neurotypical workers in data annotation, (2) identifies which characteristics result in differences between neurodiverse and neurotypical data annotation responses, and (3) explains possible reasons for those responses. We suggest design considerations for building future neurodiverse-centered data annotation user interfaces.|