My name is Dr Sarah Morgan. I work primarily at the Cambridge Department of Computer Science and Technology, as an Accelerate Science Research Fellow. I am also affiliated to the Cambridge Psychiatry Department, and to The Alan Turing Institute in London, so I can sometimes be found at one of those places!
My research applies machine learning, network science and Natural Language Processing to better understand and predict mental health conditions.
One of my main interests is using brain Magnetic Resonance Imaging (MRI) to study schizophrenia and other mental health conditions. In particular, MRI brain images can be used to investigate brain connectivity, by calculating MRI brain networks where nodes represent large scale brain regions and edges represent connectivity between brain regions. MRI brain networks from patients with schizophrenia often show altered connectivity patterns compared to healthy volunteers. My research explores both whether we can use these connectivity patterns to predict individual patients' disease trajectories, and what they can teach us about the biological mechanisms underlying schizophrenia.
I am also interested in using data science to investigate other aspects of mental health, for example using network science and natural language processing methods to study patients' speech.
There are broadly two types ways my research could be applied in future. The first is that developing a better understanding of the biological mechanisms underlying schizophrenia (for example from brain MRI) might help lead to new therapeutics. About 20-30% of patients with schizophrenia don’t respond well to current treatments, so new therapeutics could potentially be life changing for those individuals.
The other way in which my research could have real world applications is by identifying signals that can help predict or monitor disease outcome for patients with psychotic disorders. For example, we are currently exploring whether there are signals in speech data that can predict outcome for people who have some early stage symptoms of psychosis. If so, that could help clinicians target treatments better at patients who are likely to have poor disease outcomes.
I saw the Henslow Fellowship advertised by Lucy Cavendish College.
I had been working as a postdoc at the Cambridge Brain Mapping Unit in the Psychiatry Department for about a year when I applied for the Henslow Fellowship. I had lots of ideas about how the research I was doing could be extended (for example by applying a new method that had just been developed in Cambridge to construct structural brain networks for patients with schizophrenia) and I thought a Henslow Fellowship would give me time to do that. I also knew the research community at Lucy Cavendish College is friendly and supportive and I thought my work would fit in well there.
My Henslow fellowship was instrumental in giving me time to develop my expertise in brain imaging and network neuroscience. It also gave me the chance to meet researchers from other disciplines, both at Lucy Cavendish College and at the Philosophical Society. A lot of my research is highly interdisciplinary, pulling together ideas from Computer Science, Physics and Psychiatry, so having those interactions has been extremely helpful in forming new collaborations.
For my work with brain MRI, one of the biggest challenges is that magnetic resonance images only give us approximately millimetre resolution, whereas the biological mechanisms we’re interested in happen at a much smaller scale. We have recently started linking brain MRI data to genetic and genomic data to try to traverse these different scales. This sort of approach is quite new though, and we’re still learning the best ways to go about combining these different data modalities.
For my work with speech data, a key challenge is that there is relatively little data available at the moment. We are collaborating with clinicians who are hoping to start collecting more data soon, which is very exciting!
Show All
Dementia is a topic of considerable public interest. How empirical evidence has contributed to this societal awareness and indeed fear will be covered in this talk. It will span research from the 1980s when not much was understood about dementia up to contemporary perspectives. The focus will be on the epidemiological and public health evidence base, and how this relates to the results published from clinical and lab based research. The findings from UK and other high income countries of reduced age specific prevalence (%) will be explored, and the implications of results from brain based studies that dementia is not inevitable in the presence of ‘alzheimer’ type changes. The role of inequalities, risk varying across countries and time and our knowledge about protective factors have strengthened during recent years, and the balance of high risk with whole population approaches to reducing risk for society will be considered.
Registered address:17 Mill LaneCambridgeCB2 1RXUnited Kingdom
Business address:6A King's ParadeCambridgeCB2 1SJUnited Kingdom
Office hours at the business address:Monday and Thursday: 10am-12pm and 2pm-4pm.
Please contact philosoc@group.cam.ac.uk to agree a timing before visiting the office.