About Me

I am Aravind, trained as an engineer and applied mathematician, and now working as a neuroscientist. At the core of my work is a long-standing curiosity about how systems evolve over time and why they behave the way they do. This naturally led me to dynamical systems theory, which forms the backbone of my research. Real-world systems are rarely linear, predictable, or noise-free — and that is precisely what makes them interesting. My work therefore focuses on nonlinear and stochastic dynamical systems, both as mathematical objects and as tools for understanding complex biological phenomena.

My career trajectory mirrors my interest in nonlinearity. I trained initially as a mechanical engineer and completed a PhD in Applied Mechanics at IIT Madras, working on nonlinear dynamical systems in engineering contexts. I then moved to the UK, first applying these ideas to rotorcraft modelling at the University of Bristol, before transitioning into computational neuroscience at the University of Birmingham. Rather than a change in intellectual direction, these moves reflect a gradual shift in application domain. In May 2024, I joined the University of Surrey to work in a clinical research environment, with the explicit aim of bringing mechanistic modelling closer to patient-relevant questions.

I am currently the Principal Investigator on a UK Dementia Research Institute Pilot Award at the Surrey Sleep Research Centre. My work focuses on the collection and analysis of human scalp EEG to characterise differences between healthy ageing and Alzheimer’s disease. I have developed and validated a comprehensive EEG analysis pipeline in MATLAB, covering artefact handling, spectral and connectivity analyses, event detection, and statistical summarisation. This framework has been applied consistently across multiple cohorts, including young adults, healthy older adults, and individuals with Alzheimer’s disease.

On personal initiative, I have continued to work on mechanistic modelling in epilepsy since leaving Birmingham. The way I approach this problem reflects my development as a researcher. During my PhD, I often pursued increasing complexity as a route to better models: for example, in modelling magnetic interactions, I moved from an inverse-square law to dipole approximations based on Taylor expansions, and eventually to formulations involving elliptic integrals. Over time, however, I have come to appreciate that all models are approximations to an underlying ground truth, and I now favour models that are simple enough to reason about for the question at hand, yet rich enough to capture essential dynamics. In this context, I have worked with Matthew Szuromi (Boston University) to develop a parsimonious neural mass model that explains why first-line treatments fail to terminate seizures in a substantial fraction of status epilepticus cases, and how second-line treatments succeed when initial interventions do not. A detailed account of this work is available here.

Looking ahead, my aim is to study the brain explicitly as a living dynamical system — one that adapts, fluctuates, and occasionally fails. By combining mechanistic modelling with human neurophysiological data, I am interested in understanding how normal brain function emerges from interacting processes across scales, and how co-morbid diseases reflect disruptions to these dynamics rather than isolated effects. I see this systems-level perspective as essential for developing interpretable biomarkers and interventions that meaningfully connect theory, data, and clinical practice.