Mining the Plate: A Conceptual Framework for a Precision Neuro-Nutritional Behavioural Forecaster for Neurodivergent Children

Neurodivergent conditions, specifically Autism Spectrum Disorder (ASD) and Attention-
Deficit/Hyperactivity Disorder (ADHD), are increasingly recognised as complex systemic
disorders shaped by the microbiota-gut-brain (MGB) axis and an individual’s metabolic
milieu. Despite mounting evidence that dietary patterns act as critical modulators of be-
havioural phenotypes, current nutritional interventions remain reactive, relying on broad,
stressful trial-and-error elimination protocols that lack predictive power for individual chil-
dren.
To bridge this precision gap, we introduce Mining the Plate, an AI-driven computational
framework that transitions paediatric neuro-nutrition from reactive dietary restriction to
proactive, individualised behavioural forecasting. Our system integrates three complemen-
tary data streams: (1) a child’s lifetime genetic vulnerabilities derived from Genome-Wide
Association Study (GWAS) data and bidirectional Mendelian Randomisation (MR) anal-
ysis; (2) continuous real-time metabolic monitoring centred on the protein-to-sugar ratio
to anticipate insulin-mediated neurochemical disruption; and (3) behavioural biometric sig-
nals, wherein stereotyped motor acts such as eye-poking or toe-walking are algorithmically
reframed as quantitative predictive biomarkers of internal physiological distress.
Applying machine-learning techniques including pattern mining, temporal outlier anal-
ysis, and supervised classification, the framework synthesises these modalities into a Person-
alised Behavioural Risk Score (PBRS): a single, dynamic metric that quantifies the adverse
behavioural impact of a candidate food item for a specific child. Proof-of-concept eval-
uation across four dietary items demonstrates that the PBRS correctly stratifies high-risk
processed foods (cheese spread, PBRS = 14.53) from protective whole foods (vegetable soup,
PBRS = 0.20), establishing face validity for the scoring mechanism. The proposed frame-
work empowers clinicians and caregivers with actionable, data-driven mealtime strategies,
transforming nutrition into a targeted mathematical intervention for stabilising neurodiver-
gent children’s neurochemical environments.