Early Years

by Tim Shallice: Founding Director

It is a platitude to say that the human mind is one of the most complicated systems to be found in nature. But it is true. Nervous systems are at least 600 million years old.  Each newly evolving species with a nervous system will inherit its predecessor’s nervous system but adapted to face the new evolutionary niche it is in. The end product is often viewed as like a Swiss Army knife with lots of different systems each adapted to a different function. Thus the cerebellum at the back of the brain is concerned with the coordination, precision and timing of movements. 

But the situation is much more complex than that.  Take the hippocampus, a sea-horse like structure lying beneath the cerebral cortex.  In more primitive mammals, as discovered by UCL scientist, John O'Keefe, its primary function was to enable the animal to navigate round its environment. But in humans, and probably other high-level mammals like monkeys and apes it is also the structure critical for our knowledge of our own past. If it is damaged your memory of what has happened to you is limited to the last 30 seconds or so. The workings of our so-called "episodic memory" must reflect to a considerable extent their evolutionary origins as a data-base to allow you to find your way around your local environment.

Or take the cerebral cortex itself. Different regions of the cerebral cortex have different functions - from vision at the back to planning at the front.  But while inevitably there are many local differences, the basic structure of the cerebral cortex is the same everywhere, so here the  microstructure does not just reflect the overall function of the region; instead how the cerebral cortex works for a particular function must have major hardware constraints.

How do we discover how the mind, this enormously complex result of millions of years of evolution, actually works? Since 1879 in Germany and 1897 in Britain (at UCL) there have been psychological laboratories, in which purely behavioural means have been used of investigating how say perception, memory or thought works. There has been undoubted progress, particularly at the theoretical level. However, findings in experiments are typically the average of the behaviour of a set of participants, experimental subjects, usually university students. And people have different backgrounds, aptitudes and knowledge producing major individual differences. Thus one typically  does not get precise quantitative results as in physics, say, but just much cruder information such as  that one condition of the experiment produces better or worse performance than another.  So theories of functions like how you make a deduction when thinking or how working memory operates - the system  you use to hold information while you are  thinking - have proliferated and it remains very hard to tell which is on the right lines.  This is where cognitive neuroscience comes in. By seeing how the brain is operating while we carry out cognitive tasks, we get all sorts of much more specific information about what is happening. Progress can become much more rapid.

Cognitive neuroscience, then, is the investigation of how cognitive, perceptual, memory and emotional functions are organised and work, using experimental findings derived from the study of the brain as a physical object. The ICN – the Institute of Cognitive Neuroscience – was as far as I know the first centre of a major university in the world to be set up specifically on this topic. It was the brainchild of Derek Roberts, the provost of UCL from 1989 to 1999 and then again from 2002 to 2003. 

He came to the idea for a number of reasons. First, before becoming provost of UCL, he had as an electronic engineer, been head of research for the General Electric Company (GEC). He was personally very interested in how the brain computes cognitive functions. Second, at the time what came to be known as cognitive neuroscience was taking off. Its oldest component  - neuropsychology - the study of the effects of brain damage on perception, thinking and memory in humans and animals was already a century old. The study of the electrical activity of neurons during cognitive tasks was about 50 years ago. But a whole variety of new methods, such as imaging of the normal brain using positron emission tomography (PET) and later functional magnetic resonance imaging (fMRI) while subjects carried out  cognitive tasks, were coming on stream. Third, UCL was very strong in the area, but that strength was spread out not only in the Departments of Psychology and Anatomy in “old” UCL but also in the Institutes of Neurology, Child Health, and Opthalmology, which had  only recently become part of the College. Derek believed that only with some horizontal structure spanning across the old departments could the potential of that strength be properly realised. Finally, Derek could see one relative weakness in UCL's strength in area, that of computational modelling of cognitive and brain processes. However, he felt that if the Institute got off the ground, it  may be possible to obtain support to bring in a strong unit in that specific area.

In early 1996, Derek organised a meeting which showed the widespread support in College for the setting up of such a cross-departmental Institute.  He offered me the directorship of the nascent Institute and recruited an outstanding executive committee to facilitate it coming into operation. The strength and breadth of cognitive neuroscience in the College can be seen from the membership of the first executive committee.  It consisted of  Brian Butterworth FBA of the Psychology Department, cognitive neuropsychologist and pioneer of the study of  developmental dyscalculia, Richard Frackowiak, who later received the IPSEN and Wilhelm Feldberg prizes, then director of   the Institute of Neurology  Functional Imaging Lab, the leading such unit in the world, Uta Frith FRS FBA , later  co-holder of the European Latsis and Jean Nicod prizes, pioneer of the study of autism, then Deputy Director of the MRC Cognitive Development Unit, Michael Morgan FRS, a noted vision scientist then of the Institute of Ophthalmology and John O’Keefe FRS of the Anatomy Department, Nobel prize-winner to be and discoverer of place cells in the hippocampus.

The first two years were mainly devoted to developing a cross-departmental community of cognitive neuroscience researchers. It was done successfully through the organisation  of three series of monthly seminars on Computational Modelling in Cognitive Neuroscience, the Neurobiology of Cognitive Development and the Cognitive Neuroscience of Consciousness.  By the end of the first year the Institute had 81 members at the postdoctoral level or above.

A key development in the second year was the fulfilment of one of Derek Roberts’ original aims, namely the recruitment into UCL of a world-leading computational neuroscience group. Neuroscience is an area where there are very large number of experimentalists using a huge variety of different types of technique. However, by comparison with a subject like physics, there is a real lack of theoreticians.  UCL was well represented at a more junior level by Karl Friston  of the Institute of Neurology and Neil Burgess, then of Anatomy. However, its strength in depth of theoreticians was not comparable to that of some of the more experimental areas. In a series of negotiations between the Gatsby Charitable Foundation, UCL and Geoffrey Hinton, then at the Department of Computer Science, University of Toronto, it was agreed that the Gatsby Charitable Foundation would set up a unit in computational neuroscience to be linked administratively and physically to the ICN, which would then have a bricks-and-mortar home. Hinton, at the time, was probably the leading neural network modeller in the world, and has since become famous as the Godfather of Deep Learning, the theoretical basis of many artificial intelligence recognition programs. As I was a former collaborators of his, it was straightforward for the ICN to  be the link between the new Gatsby unit and the College. Hinton brought with him an extremely strong group of theoreticians - Peter Dayan, later director of the Gatsby unit, Zhaoping Li and Zoubin Ghahramani. 

One major issue that remained to be resolved was where the core of the ICN and the Gatsby unit could be located.  This was dealt with in a  dramatically effective fashion by Derek Roberts.  A six-story building,  Alexandra House in Queen Square had been the offices of a firm of solicitors. In July 1997 the building became available to rent for a 15 year period.  UCL had one week in which to decide whether or not to rent it. Fortunately, the provost made the very courageous decision to rent it with  two top floors being allocated to the Gatsby unit and the lower four floors to the ICN.

This enabled the ICN to house a core group of scientists and their groups of researchers in an ideal location, directly opposite the main National Hospital for Neurology and Neurosurgery building and the main Institute of Neurology building and just a few doors down from the Functional Imaging Lab. From Psychology came Brian Butterworth (see above), Patrick Haggard, later to found  a new field, the brain basis of agency, what makes an action intentional and Paul Burgess, expert on the frontopolar cortex, the part of the cortex most distant from the senses. From Anatomy came the human behavioural side of the work of John O'Keefe (see above) and his colleague Neil Burgess, currently the director of the ICN. The Institute of Ophthalmology was represented by Michael Morgan (see above). In addition, the College provided new positions so that we could recruit Jon Driver, a leading world expert on attention, from Birkbeck College, who was to become the second director, Michael Rugg, Europe's leading expert on the study of the electrical rhythms of the brain, from St Andrews, and most critically Uta Frith (see above) to become Deputy Director.  Soon to join us, from the Institute of Neurology were Masud Husain and Geraint Rees, later to become the third director, bringing invaluable clinical research, and from Oxford  Vincent Walsh pioneer of Transcranial Magnetic Stimulation, a method of producing very short-term completely reversible malfunctioning of the human brain.

To have so many leading researchers with their groups in the same building, and in a building also holding Gatsby Computational Neuroscience Unit and so close to the imagers of the Functional Imaging Lab and the clinical researchers of the National Hospital produced a remarkable synergy leading to a mass of outstanding research, some of which you can see on the website.