Decoding neural mechanisms underpinning human cognition in health and disease using machine learning

Summary This project aims to understand the neural mechanisms underpinning higher-order cognitive function in humans. This student will be guided in using state-of-the-art neuroimaging techniques and machine learning to decipher the complex network of brain circuits that give rise to individual variation in cognition in healthy adults as well as abnormal brain changes associated with cognitive decline in neurodegenerative diseases such as Alzheimer’s disease and mild cognitive impairment.

Project Details

Complex higher-order cognitive function resulting from the greatly expanded forebrain is one of the most unique features that distinguish humans (i.e., homo sapiens) from other species such as non-human primates. However, the evolutionary leap in the size of the most frontal part of the human brain could not explain all the cognitive advances and the highly variable cognitive capacities among humans. Indeed, the performance of higher-order cognitive function varies substantially across individuals and is strongly related to various domains of individual life outcome, such as educational achievement, social mobility, occupational attainment and job performance as well as the quality of life while ageing. Yet, the precise neural mechanisms giving rise to such inter-individual variation remain unclear.

We are seeking an enthusiastic and motivated master/honours candidate to investigate the relationship between brain network circuits and higher-order cognitive function in health and neurodegenerative diseases using large neuroimaging datasets including the Human Connectome Project (n=1200) and the UK biobank (n=500,000).

We will focus on brain phenotypes derived from functional brain imaging modalities such as resting-state and task-evoked functional magnetic resonance imaging (fMRI) acquired at ultrahigh field strengths (e.g., 3T and 7T). Other imaging modalities such as structural and diffusion MRI could also be analysed depending on specific study hypotheses.  Novel machine learning techniques will be developed and utilised to facilitate the identification of reproducible and generalizable neural markers underpinning human cognition as well as brain changes associated with cognitive dysfunction.

Accurate mapping of neural markers associated with various domains of cognition will be an important step towards the understanding of intellectual ability unique to humans. Moreover, it will provide the fundamental premise for developing effective interventions to ameliorate the negative effect of cognitive decline resulting from neurodegenerative diseases while ageing.

The successful applicant will work on rich imaging-derived brain phenotypes (or raw imaging data), medical records, behavioural and cognitive measurements. The student will be responsible for conducting a literature review, the development of a more specific proposal and hypotheses, computer programming as well as statistical analyses. The student participating in this project will learn advanced neuroimaging and machine learning techniques. A student who is interested in or has a good understanding of neuroscience, biology and medicine as well as some experience in computer programming would be well suited to this project.