The Predictive Clinical Neuroscience lab at the Donders Institute and RadboudUMC aims to: (i) shift perspectives on mental health through methodological innovation; (ii) advance precision medicine by integrating across multifaceted data measuring biology, behaviour and environment, in daily life and across the lifespan; (iii) make the innovations we develop accessible to both the scientific community and clinical practice. We further aim to lead in good research practice, set standards for transparent data analysis, open science and scientific software development.
In more detail, we develop statistical and machine learning techniques to predict and stratify brain disorders and to understand their underlying neurobiological mechanisms on the basis of diverse data sources including brain imaging, digital phenotyping and remote sensing satellite data. We focus both on supervised techniques for predicting clinical variables as well as unsupervised approaches for stratifying clinical groups on the basis of the underlying biology.
Some specific methodological techniques of interest include:
- Normative modelling approaches for modelling variation in clinical cohorts
- Machine learning methods for densely sampled longitudinal timeseries
- Neural networks for learning latent representations
- Stochastic processes models
- Bayesian non-parametric methods for pattern classification and regression
- Doubly multivariate statistical methods for finding associations between datasets
- Markov chain Monte Carlo methods for inference in probabilistic models
- Multi-output, multi-task and structured output learning methods
Current Projects
MENTALPRECISION: Linking brain and behaviour for precision stratification of mental disorders
environMENTAL: Reducing the major environmental challenges on mental health
PRECOGNITION: Learning latent cognitive profiles to predict psychosis
Raynor Cerebellum Charts: Growth charting of the human cerebellum
Predictive Clinical Neuroscience Lab 