Exploring the Sensory–Microbiome Nexus Through a Hybrid Intelligence Framework (HIF): A Fibre-Enriched Food Intervention Case Study

This paper introduces a Hybrid Intelligence Framework (HIF) for human-AI collaboration in the analysis of complex biological and behavioural datasets. HIF combines generative-AI pattern surfacing with subject-matter-expert (SME) curation to support transparent, auditable interpretation in small, multivariate studies. We demonstrate the framework using a two-week fibre-enriched food intervention in autistic adults, in which twelve participants consumed a savoury nugget daily while providing oral and faecal microbiome samples, baseline diet information, and daily sensory enjoyment ratings. The dataset was analysed independently using two approaches: a conventional SME pipeline based on Healthy Diet Index (HDI) grouping and established microbiome statistics, and a HIF workflow that derived an alternative fibre-intake score and used iterative human-AI dialogue to explore patterns. Despite these different analytical routes, both approaches converged on the same core findings. Participants with lower baseline fibre intake showed larger microbiome compositional shifts over two weeks, alpha-diversity responses were heterogeneous, and beta-diversity separation was clearer among low-fibre or low-HDI participants. Both analyses identified reductions in opportunistic genera and an apparent link between higher product enjoyment, better adherence, and clearer microbial trajectories. This case study illustrates how HIF can complement expert statistical analysis by reframing small, noisy datasets, making analytical assumptions explicit, and linking sensory enjoyment with microbiome responsiveness. To our knowledge, this is the first application of a structured HIF in food and sensory science.