Brain imaging yields activation maps that reproduce well on group level. However, the maps also show marked individual variation. Such variation needs to be understood in order to enhance the interpretability of brain imaging and to unleash its great potential in science, clinic and society. Our project takes brain imaging to the next level, from group-level description to individual-level prediction. We determine individually unique features and reproducibility within individuals, variability between individuals, and relationship of brain-level variability to variation in behaviour, utilising machine learning methods. Our goal is to use brain imaging to predict individual disease trajectories in neurological patients (e.g., stroke, mild traumatic brain injury), and to enhance the sensitivity of neuroimaging by grouping individuals based on brain patterns instead of coarse categorisations like gender or age. Our work will enable more informative, reproducible and accurate neuroscience.

2023

2027