 |
International Journal of Scientific Research and Engineering Development( International Peer Reviewed Open Access Journal ) ISSN [ Online ] : 2581 - 7175 |
📑 Paper Information
| 📑 Paper Title | AI-Based Adaptive Learning for Autistic Children: Effects on Task Engagement and Motivation in Cameroon's Inclusive Primary Schools |
| 👤 Authors | Che Myra Ndum, Nyock Ilouga Samuel |
| 📘 Published Issue | Volume 9 Issue 4 |
| 📅 Year of Publication | 2026 |
| 🆔 Unique Identification Number | IJSRED-V9I4P23 |
📝 Abstract
Background: Behavioural engagement strongly predicts academic achievement, yet it is systematically depressed among autistic children, whose heterogeneous learning profiles are poorly served by uniform whole-class instruction. In sub-Saharan Africa, where less than 1% of global autism research is conducted, and teachers receive little preparation in differentiated methods, no experimental study has tested whether AI-based adaptive learning can raise the engagement of autistic learners.
Objectives: This study evaluated the effects of AI-based adaptive learning on the task engagement and behaviouralmotivation of children with ASD Level 1 in inclusive primary schools in Yaoundé, Cameroon, compared with traditional whole-class instruction over twelve weeks.
Methods: A quasi-experimental study embedded in a convergent parallel mixed-methods design engaged 48 children with ASD Level 1 (24 experimental, 24 control) across one public and one private inclusive school. Grounded in the Zone of Proximal Development, Cognitive Load Theory, and Self-Determination Theory, the intervention used two adaptive applications: Khan Academy Kids and Autism ACE, an offline-capable application co-developed for local infrastructure conditions. Data were collected through the Structured Behavioural Observation Grid, automated usage logs, and teacher and parent interviews analysed using reflexive thematic analysis.
Results: Children in the adaptive condition completed over three times more task items per session (3.37 vs. 1.04; t= 6.38, p < .001, d = 1.21), displayed more positive affect (d = 0.60) without increased distress, and disengaged less often. Engagement followed a U-shaped trajectory inconsistent with a novelty effect. Teachers reported reduced mathematics avoidance and strengthened capacity to differentiate instruction; the principal challenge was managing overexcitement and session transitions.
Discussion: AI-based adaptive learning is feasible in Central African inclusive classrooms and substantially raises the behavioural engagement of autistic children, providing the first controlled evidence for this population and context. The findings position engagement as the gateway through which adaptive instruction operates, and identify infrastructure reliability and teacher professional development in inclusive AI use as the conditions under which these gains can translate into broader learning outcomes.
Objectives: This study evaluated the effects of AI-based adaptive learning on the task engagement and behaviouralmotivation of children with ASD Level 1 in inclusive primary schools in Yaoundé, Cameroon, compared with traditional whole-class instruction over twelve weeks.
Methods: A quasi-experimental study embedded in a convergent parallel mixed-methods design engaged 48 children with ASD Level 1 (24 experimental, 24 control) across one public and one private inclusive school. Grounded in the Zone of Proximal Development, Cognitive Load Theory, and Self-Determination Theory, the intervention used two adaptive applications: Khan Academy Kids and Autism ACE, an offline-capable application co-developed for local infrastructure conditions. Data were collected through the Structured Behavioural Observation Grid, automated usage logs, and teacher and parent interviews analysed using reflexive thematic analysis.
Results: Children in the adaptive condition completed over three times more task items per session (3.37 vs. 1.04; t= 6.38, p < .001, d = 1.21), displayed more positive affect (d = 0.60) without increased distress, and disengaged less often. Engagement followed a U-shaped trajectory inconsistent with a novelty effect. Teachers reported reduced mathematics avoidance and strengthened capacity to differentiate instruction; the principal challenge was managing overexcitement and session transitions.
Discussion: AI-based adaptive learning is feasible in Central African inclusive classrooms and substantially raises the behavioural engagement of autistic children, providing the first controlled evidence for this population and context. The findings position engagement as the gateway through which adaptive instruction operates, and identify infrastructure reliability and teacher professional development in inclusive AI use as the conditions under which these gains can translate into broader learning outcomes.
📝 How to Cite
Che Myra Ndum, Nyock Ilouga Samuel, "AI-Based Adaptive Learning for Autistic Children: Effects on Task Engagement and Motivation in Cameroon's Inclusive Primary Schools" International Journal of Scientific Research and Engineering Development, V9(4): Page(193-205) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.
📘 Other Details