Data Science

Platform Leadership

Rudolf Uher, MD, PhD
Frank Rudzicz, PhD

Platform Operations

The Data Science Platform operates at the following research site:

• Dalhousie University (Halifax, Nova Scotia)

Platform Overview

The Data Science Platform is guided by the central question: “Which individual will benefit most from which treatment?” To address this, the platform uses advanced computational approaches to analyze datasets collected across all CAN-BIND platforms. These data are applied to develop, fine-tune, and test models that help us understand patterns in treatment response and identify meaningful subtypes of depression.

The Data Science Platform analysts use diverse data sources, including speech recordings, actigraphy, electroencephalography (EEG) data, and other clinical and behavioural measures to fine-tune large models that estimate symptoms and predict treatment responses. In addition to prediction, the platform focuses on identifying patterns across symptoms, behaviours, and treatment outcomes that may not be easily visible through traditional analyses. By applying computational approaches to complex, multi-source data, the platform aims to capture relationships and trends that can inform a more precise understanding of depression and its treatment.

This work supports the development of practical, data-informed tools that can help guide treatment decisions in clinical settings and improve how new treatments are discovered and evaluated. To achieve these goals, CAN-BIND’s multidisciplinary Data Science Platform brings together expertise in computational approaches, machine learning, deep learning, biostatistics, and data integration, ensuring that diverse data sources can be effectively combined and interpreted.