Mapping biological networks of the human brain remains one of the complex endevaours in human history. Connectomics is a field of study that aims to map all neural connections in the brain (estimated to be 100 trillion) using MRI, fMRI and EM images obtained from various projects under the Human Connectome Project. Unfortunately, these microscopic neuroimages come in wide variety of sizes, rotations and colours. Spliced 2D images of various segments of the brain with non-standardised fluorescent labelling have to be assembled into a three dimensional representation, in order to understand the neural connectivity (comprising of billions of neurons). To ensure high labelling accuracy, manual tracing (by neuroscientists) remains the gold standard in deciphering these neural connections. This is a time-consuming, trivial and labor-intensive task especially when we consider the number of plausible connections at the microscale resolution.
A citizen scientist initiative to gamify this annotation process was launched in 2012 (Eyewire) with the aim of mapping the retinal neurons. The game has attracted more than 100,000 participants from all over the world. Gamers are tasked with tracing a neuron and its connections based on images of brain slices obtained from a brain tissue. The reconstruction of this branch is validated by experts for its accuracy. Concurrently, the human tracing behaviour is adapted to an AI engine in an effort to improve the automatic annotation capability. Although the activity is simplistic and does not require prior scientific knowledge, this is not your average colouring exercise (i.e., where keeping it inside the line would suffice). There is a minimal training required before your work can be accepted.
On the one hand, the task seems trivial and is highly suited for machine learning adoption, however, the intricacies of tracing the path interconnecting these neurons require innate human traits unknown to conventional supervised learning algorithms. The amount of unannotated data further limits the capability of supervised machine learning approaches for this task. On the other hand, reinforcement learning (RL) might provide a better alternative. Rapid progress has been achieved in developing computer gamers through utilisation of deep neural networks trained using RL paradigms. Advanced RL approaches exemplified by AlphaGo Zero (and its variants) and OpenAI have demonstrated that these systems outperform the best human players in their specialised domains (e.g., Go, Chess, Shogi, 3D games, and other RPG computer games).
RL learns by competing against another agent (could be human or another machine) and in principle does not rely on the availability of labelled data for training. Is it possible to extend the capability of RL to map the human brain? The ability to accurately map our brain carries a significant impact for the future of mankind. Notwithstanding this grand contribution, the proposed research will contribute to further understanding in the area of neuroscience and may result in the development of newer approaches in AI and data analytics towards artificial general intelligence.
Applicants should hold, or expect to obtain, a First or Upper Second Class Honours Degree in a subject relevant to the proposed area of study.
We may also consider applications from those who hold equivalent qualifications, for example, a Lower Second Class Honours Degree plus a Master’s Degree with Distinction.
In exceptional circumstances, the University may consider a portfolio of evidence from applicants who have appropriate professional experience which is equivalent to the learning outcomes of an Honours degree in lieu of academic qualifications.
If the University receives a large number of applicants for the project, the following desirable criteria may be applied to shortlist applicants for interview.
The University offers the following levels of support:
The following scholarship options are available to applicants worldwide:
These scholarships will cover full-time PhD tuition fees for three years (subject to satisfactory academic performance) and will provide a £900 per annum research training support grant (RTSG) to help support the PhD researcher.
Applicants who already hold a doctoral degree or who have been registered on a programme of research leading to the award of a doctoral degree on a full-time basis for more than one year (or part-time equivalent) are NOT eligible to apply for an award.
Please note: you will automatically be entered into the competition for the Full Award, unless you state otherwise in your application.
The scholarship will cover tuition fees at the Home rate and a maintenance allowance of £19,000 (tbc) per annum for three years (subject to satisfactory academic performance).
This scholarship also comes with £900 per annum for three years as a research training support grant (RTSG) allocation to help support the PhD researcher.
Due consideration should be given to financing your studies. Further information on cost of living
Submission deadline
Friday 7 February 2020
12:00AM
Interview Date
23 to 24 March 2020
Preferred student start date
Mid September 2020