APP56059 Connecting Brain Networks and Hormones together via Pseudospectra to Understand the Mechanisms of Depression

Depression is a multifaceted mental health condition affecting millions globally, arising from disruptions in the brain networks and hormonal systems that regulate mood, stress responses, and emotional well-being. These systems are deeply interconnected, yet studying their dynamics is challenging due to variability across individuals and limitations in obtaining comprehensive hormonal measurements. Mathematical modelling provides a powerful tool to address these complexities by capturing the interplay between brain networks and hormonal systems in a structured and quantifiable framework. Such models can integrate diverse biological data, simulate intricate feedback loops, and yield insights otherwise unattainable through direct observation. They enable researchers to test hypotheses, predict outcomes, and illuminate the mechanisms of depression, providing critical guidance for developing targeted interventions. This is especially vital given the limitations of current treatments, such as antidepressants and therapy, which often take weeks to act and fail for many patients. My research focuses on creating mathematical models that integrate real-world data to elucidate the dynamics of depression. In the model-building stage, I will incorporate data from advanced sensing technologies, such as fMRI and EEG, to capture how brain networks influence hormones. I will also embed inferred interactions for hormones thought to be implicated in depression, even when time series data is scarce, to examine their role in the downstream behaviour of measurable hormones like cortisol. These models will be tuned to fit human parameter values and account for complexities like time delays in hormonal signalling. By integrating diverse data sources and ensuring consistency across them, my work will unify brain networks and hormonal rhythms into a comprehensive framework. Upon deriving these models, I will use structured pseudospectra, which is a key innovation in my approach. Unlike standard pseudospectra, which analyse how systems respond to general perturbations, structured pseudospectra evaluate how specific perturbations—such as changes in hormone levels caused by medication or alterations in brain network structure from therapy—affect the system. This enables a detailed exploration of the bidirectional influences between brain networks and hormones, revealing the most sensitive components of these systems and guiding the development of more precise treatments. With advances in sensing technologies like real-time hormone monitoring and high-resolution brain imaging, I will derive data-driven models that I can also explore with structured pseudospectra. This will complement the mechanistic insights gained from my equations and in particular allow me to validate my equations and explore bio-variability across populations. The societal implications of this work are profound. By minimising trial-and-error prescribing, reducing unnecessary side effects, and enhancing the likelihood of successful outcomes, this research has the potential to transform mental health care. Simulating personalised responses to interventions offers tailored solutions for patients, while advancing our understanding of depression’s mechanisms could alleviate its societal and economic burdens, ultimately improving countless lives. Recognising the sensitivity of this topic, my research will include a component of Public and Patient Participation, ensuring that end-users of any potential technology are actively consulted and included throughout the research process. This engagement will help align scientific advancements with the needs and experiences of those they aim to benefit.