Level 2 Screening With the PDD Behavior Inventory: Subgroup Profiles and Implications for Differential Diagnosis

Abstract

The PDD Behavior Inventory (PDDBI) has recently been shown, in a large multisite study, to discriminate well between autism spectrum disorder (ASD) and other groups when its scores were examined using a machine learning tool, Classification and Regression Trees (CART). Discrimination was good for toddlers, preschoolers, and school-age children; generalized across clinical diagnostic sites; and agreed well with Autism Diagnostic Observation Schedule (ADOS) classifications. Results also revealed three subtypes of ASD: minimally verbal, verbal, and atypical that differed in developmental history, behavior profiles, and biomedical findings. Seven subtypes of Not-ASD children were identified, two of which were relatively common. Three of the remaining five relatively rare Not-ASD subgroups had highly atypical profiles marked either by extreme aggressiveness or by extreme ritualistic behaviors. PDDBI profiles of these rare subgroups were not previously characterized. In this study, profiles of all CART subgroups based on parent and teacher PDDBIs are described, along with their implications for diagnosis and assessment.

Artificial Intelligence (AI) in academic research. A multi-group analysis of students’ awareness and perceptions using gender and programme type

“…The benefits of AI in educational circles and research have been well documented in previous studies. For example, it is stated that AI technology like chatbots is used for review of literature (Clark, 2020), intelligent tutoring and automated data collection (Heffernan & Heffernan, 2014), student collaboration and personalisation of learning experiences (Luckin et al, 2016;Chiu et al, 2022;Mertala et al, 2022), monitoring progress of a work (Swiecki et al, 2019), automated data collection and analysis (Okada et al, 2019;Vij et al, 2020;Yuan et al, 2020), profiling respondents' background (Cohen et al, 2017), as well as analysis of data using different statistical packages . Other areas are AI tools like ChatGPT that have the internal capacity to assist students and researchers, in areas which include writing tasks, text generation, language translations, and responding to academic queries (Dwivedi et al, 2023;Kasneci et al, 2023;Lund et al, 2023).…”