Currently accepted at: Journal of Medical Internet Research
Date Submitted: Mar 23, 2026
Date Accepted: Jul 14, 2026
This paper has been accepted and is currently in production.
It will appear shortly on 10.2196/95887
The final accepted version (not copyedited yet) is in this tab.
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Multimodal Analysis of Child–Horse Interactions Using Vision–Language Models in an Australian Aboriginal Equine-Assisted Learning Program: A Development and Evaluation Study
ABSTRACT
Background:
The human-horse bond has played a role in human development through history and has more recently been applied to mental health promotion through Equine-Assisted Learning (EAL) programs. In EAL settings, therapeutic assessment commonly relies on practitioner observations documented through behavioural checklists, structured interviews, and progress notes, which can pose challenges due to time demands, subjectivity, and inconsistent standardisation across practitioners and contexts.
Objective:
To address these challenges and support more quantitative analysis of therapeutic interactions, this study developed an AI-powered framework based on Visual Language Models (VLMs) to analyse child-horse interactions in EAL settings. The framework was developed in collaboration with the Yawardani Jan-ga EAL program, which supports the social and emotional well-being of Aboriginal youth in the Kimberly region of Western Australia.
Methods:
The proposed framework comprises three core components: (1) an activity classification module employing a VLM with few-shot learning based on SigLIP2 to identify nine therapeutic activities; (2) a proxemics analysis module using a qwen2.5-VL-7B VLM to classify physical contact between the child and horse; and (3) a body posture analysis module applying multimodal Chain-of-Thought (MCoT) prompting and self-consistency techniques to classify children’s body posture into confidence-related states with interpretable reasoning.
Results:
The framework was trained and evaluated using real-world images collected during therapeutic sessions with Aboriginal children participating in the Yawardani Jan-ga program. The activity classification module achieved F1, accuracy, precision, and recall scores of 91.6%, 90.7%, 92.2%, and 91.7%, respectively. The proxemics analysis module achieved an accuracy of 88.6%, precision of 88.3%, and an F1 score of 88.1%. For the body posture analysis module, system reliability was assessed using stability metrics, achieving an average Consistency of 0.916, an Agreement Rate of 0.705, a Rationale Coherence of 0.857, and a Vote Entropy of 0.174.
Conclusions:
The findings suggest that AI- and VLM-based approaches may provide complementary support for therapeutic assessment in EAL and warrant further investigation.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.