Autism spectrum disorder (ASD) is a complex neurodevelopmental condition characterised by persistent challenges in social communication and interaction, alongside restricted and repetitive behaviours. Despite almost a century of research, precise aetiologies and biomarkers are yet to be elucidated. While metabolic and circadian dysfunction has been linked to ASD, the exact connection remains elusive.

Additionally, ASD diagnosis remains reliant on qualitative assessments, which are time-consuming, resource-intensive, and inherently subjective. ASD rates have increased eightfold in the last two decades, underscoring the need for faster, objective assessments. The absence of an accurate, quantitative diagnostic tool delays early identification, intervention, and support for children with ASD.

In this work we use time-localised, phase-based, multiscale analysis approaches
to gain diagnostic and mechanistic insights into systems implicated in ASD. Electroencephalographic measurements are inherently susceptible to movement artefacts. We demonstrate that phase-based connectivity measures, such as wavelet phase coherence (WPC) and dynamical Bayesian inference (DBI), can detect interactions between brain regions despite these disturbances. By applying WPC and DBI, we reveal distinct connectivity patterns in the frontal cortex of young males with ASD, suggesting a potential biomarker. Circadian dysregulation is also prevalent in ASD, yet remains poorly understood; time-localised analysis such as wavelet transforms, ridge extraction and harmonic analysis allows us to establish the presence of behavioral modes and trace their changing frequency content over time. We reveal an irregular circadian rhythm that may contribute to disrupted sleep patterns in ASD. Given the established links between cellular energy metabolism and ASD, we also propose a simple phase oscillator-based model that simulates altered metabolic
pathways with significantly fewer parameters than mass-based approaches.
Applying physics-based approaches to understand cellular dynamics, electrophysiology, and circadian regulation, contributes towards a cohesive framework to understand the multifactorial nature of ASD. Collectively, these findings provide mechanistic insights while enhancing diagnostic capabilities through practical guidance on measurement and analysis. By explicitly considering key physical principles of biological systems, the framework presented can significantly advance the assessment of ASD and other neurological conditions, such as ADHD, depression and dementia.