Background
This PhD project aims to enhance our understanding of how treatments for anxiety and depression work by identifying key mechanisms and patient subgroups that respond differently to interventions. Traditional methods for mediation and moderation analysis are not well equipped to handle the high-dimensional data now available, such as genomic data, electronic health records, and digital phenotyping from smartphones and wearable devices. This project will integrate advanced machine learning techniques with traditional approaches, such as structural equation modelling (SEM), to more effectively analyse these datasets and uncover important mechanistic variables.
Novelty and Importance
By applying these cutting-edge methods, the research addresses the growing need to adapt mediation and moderation analyses to fully exploit large-scale, multimodal datasets. The resulting insights will be critical for personalising treatments and improving their effectiveness, particularly for specific subgroups of patients. This approach is novel, as machine learning has rarely been applied to this type of analysis, offering a new way to understand the mechanisms underlying mental health interventions and for whom they work best.
Aims and Objectives
This project will apply state-of-the-art methods to uncover mechanistic variables in high dimensional, multimodal datasets. The specific aims are to:
1. Identify mediators and moderators of treatment outcomes for anxiety and depression within large-scale datasets.
2. Apply and compare traditional and machine learning-based approaches to mediation and moderation.
3. Share the findings and provide training to make these advanced techniques more accessible for future research.
Ultimately, this PhD project will contribute to advancing personalised mental health treatments, improving their efficacy and targeting through a deeper understanding of underlying mechanisms.