Gatsby Computational Neuroscience Unit



The Gatsby Computational Neuroscience Unit was established at UCL by the Gatsby Charitable Foundation in July 1998 to provide a unique opportunity for a critical mass of theoreticians to interact closely with one another and with other UCL research groups. The unit is one of the first centres in the world to bring together the fields of theoretical neuroscience and machine learning, and our researchers have pioneered research into the mathematical underpinnings of learning, perception and action in biological and artificial systems.

The unit was established with Geoffrey Hinton (founding director; Turing Prize 2019), Peter Dayan (director 2002-2017; Brain Prize 2017), Li Zhaoping and Zoubin Ghahramani (Royal Society Milner Award 2021). Since then, faculty members have included Peter Latham (current member), Maneesh Sahani (current director), Yee Whye Teh, Arthur Gretton (current member), Aapo Hyvaarinen, Peter Orbanz (current member) and Andrew Saxe (current member). 


In 2016 we moved from our home in Queen Square to Fitzrovia to create a collaborative partnership with the Sainsbury Wellcome Centre for Neural Circuits and Behaviour (SWC). The unit and SWC, both part of the Faculty of Life Sciences at UCL, work closely together, with parallel PhD programmes, common day-to-day activities, and joint appointments and research projects that bring together theoretical/computational and experimental neuroscience. We are also part of the cross-faculty Centre for Computational Statistics and Machine Learning (CSML), and collaborate with other UCL research groups in neuroscience and machine Learning. In 2020, the unit joined the ELLIS Unit at UCL, part of the wider European Lab for Learning and Intelligent Systems (ELLIS), Europe's leading network for AI research. We have strong links with two other Gatsby Charitable Foundation-funded research centres: the Center for Theoretical Neuroscience at Columbia University and the Gatsby Programme in Theoretical Neuroscience, Hebrew University.

Education and training

The unit runs an innovative, highly competitive four-year PhD programme in Theoretical Neuroscience and Machine Learning, and offers two postdoctoral training opportunities: the postdoctoral training fellowships and the joint postdoctoral fellowships with SWC. The latter was developed in partnership with SWC and aims to foster early career researchers who are interested to work at the interface of computational/theoretical and experiment neuroscience. In addition, we host an external seminar series with talks by experts in the fields, and organise scientific workshops and public events.


We are a unit of around 40 people, functioning as far as possible as a single large research group. This small size allows us to interact closely and frequently through regular research talks, journal clubs, lab meetings, external seminars and tea hours (with tea, snack and informal talks). On Friday afternoons we have joint tea hours with SWC. This strong cohesion is a defining characterising of the unit. We support flexible working and are committed to support the career development of our PhD students and postdoctoral fellows.


Around 90% of our alumni students and postdoctoral fellows work in a scientific setting, with over 60% holding an academic position and about 30% working in research development in companies such as DeepMind (itself co-founded by two former postdocs, Demis Hassabis and Shane Legg). Two, Yasser Roudi and Misha Ahrens, have won the Eric Kandel Young Neuroscientist Prize in 2015 and 2019, respectively. Please see The Greater Gatsby page for a list of former faculty members, postdoctoral fellows, PhD students and professional services staff.