Cambridge

Varun Warrier & Ed Bullmore
Academic leads
Samina Begum
Co-production lead

Brain network biomarkers

Objectives

1

Primary: We will use magnetic resonance imaging (MRI) to measure individual differences in human brain network formation (network phenotypes) that are caused by variation in the genetics of the immune or metabolic systems.

2

Exploratory: We will use positron emission tomography (PET) to measure inflammatory changes in parts of the immune system, located close to the brain, that have been linked with brain and mind disorders.

What we are doing

We are working on several big, openly accessible datasets, generously released by the UK Biobank and the Psychiatric Genomics Consortium, to analyse genetic effects on brain MRI scans from hundreds of thousands of people.

We are using advanced statistical and computational methods to analyse these big (and growing) datasets. This is technically challenging work. However, the unprecedented scale and richness of the data now available, and the power of recent developments in AI, can be combined to give us new insights into some fundamental questions.

How does genetic inheritance influence the normal developmental process of human brain network formation, from mid-pregnancy through childhood and adolescence to adult life?

How do commonly occurring, minor DNA differences between people in their immune or metabolic genes cause changes in their brain networks that make them individually more resilient or more vulnerable to severe mental illness?

Answers to these questions will tell us which of the many millions of phenotypes we can measure in a single MRI scan are most strongly caused by DNA differences in genes for the immune or metabolic systems.

We can then use these particular MRI phenotypes as validated biomarkers in future clinical trials of drugs or other interventions targeting the immune or metabolic pathways to severe mental illness.

What we are planning to do

We will access and re-analyse PET data on hundreds of people with major depression, previously collected for research studies using a radio-active tracer that binds to inflammatory cells in the brain and other parts of the body. These data have previously been analysed to show that depression is associated with increased inflammation in parts of the brain. We will re-analyse these data, focusing this time not on the brain itself but on compartments of the immune system in the bone marrow of the skull and the meningeal membranes that are folded between the brain and the skull.

It has been recently shown that immune cells concentrated in the skull bone marrow and brain membranes can probably move into the brain and increase levels of brain inflammation, contributing to neurodegenerative disorders and depression (see published research here, here, and here).

We will use AI to make new tools for measurement (segmentation) of inflammation in the skull and meninges and we will test the prior hypothesis that these “non-brain” sources of inflammation are associated with brain and mental health disorders.

Once we have done this work, we will be able to decide if PET images of non-brain inflammation could be useful biomarkers in future clinical trials of anti-inflammatory drugs or other interventions for brain and mind disorders.

Find out more

Since about 2015, a growing international research community has made several new discoveries about the human brain’s “genetic architecture”. There are several recent reviews of the exciting range of work ongoing globally at the interface between genetics and brain imaging.

We have shown that genetic effects are typically widespread throughout the brain. Most DNA variations that make any difference to the brain phenotypes that we can measure using MRI tend to make a difference across networks or gradients or hierarchies of multiple inter-connected brain areas (rather than each specific DNA variant making a difference to only one or two localised areas). You can see research on this here and here.

We have also shown that genes which normally influence the structure of highly connected hubs in brain networks are also genes which increase an individual’s risk for schizophrenia. This result highlights the growing consensus that the “genes for mental illness” are commonly in fact genes for human brain development. You can see research on this here and here.

Get involved

There are no opportunities at the moment. Check again soon!