25 August 2016

Identifying officers with extraordinary face recognition skills

Dr Sarah Bate is Principal Academic in Psychology and Lead Researcher within the Centre for Face Processing Disorders at Bournemouth University.  Here she explains about her work in super facial recognition.

What is super facial recognition?

Most people take their face recognition skills for granted, but this cognitive ability actually falls onto a wide continuum. At the bottom end, a minority of people have a condition called "prosopagnosia" or "face blindness", characterized by extreme difficulties in face recognition. This can affect recognition of the faces of family, friends, work colleagues, and even a person's own face. Prosopagnosia can rarely occur following brain injury, but many more people simply fail to develop normal face recognition skills.

At the other end of the spectrum are people with extraordinary face recognition skills. These people are often referred to by the media as "super recognisers". 

Why is this research taking place?

While initial work has examined these people as a means to inform our understanding of the human brain, a recent collaboration between Bournemouth University and Dorset Police is now investigating the underpinnings and identification of proficient face recognition. If aptitude tests can identify existing and new officers who perform extremely well on certain tasks, they may be an untapped resource that can be utilised in a variety of policing scenarios.

This work initially began as a series of neuropsychological investigations that aimed to explore expert face recognition as a means to further our theoretical understanding of the human face recognition system. However, as research into individual differences in face recognition has received increasing attention in recent years, the more applied value of the work is starting to be realised. In particular, recent terrorist attacks have brought issues of national security to the forefront, and the location of suspects has been plagued by difficulties in facial identification. The same issues are relevant to many areas of policing, and it has become increasingly clear that some people's identification skills are much more reliable than others.

Why are some people better at face recognition than others?

Some recent studies indicate that face recognition skills are more highly correlated in identical than non-identical twins, and there is also evidence to suggest that the developmental form of face blindness can run in families. Our face recognition skills therefore seem highly heritable, and some people may simply be born with a predisposition to be excellent at this task. This suggestion is supported by findings that it is very difficult to train people to be better at face recognition, and those with years of on-the-job experience in occupations such as passport control are no better at face matching than new recruits. Recent work from Bournemouth University provides some insights into the strategies that super recognisers use to recognise faces: In two separate experiments, people with extraordinary face recognition skills were found to look at the centre of the face (i.e. the nose) for a greater length of time than typical participants who focused more on the eyes.

How do you identify people with extraordinary face recognition skills?

Some people are aware that they are very good at face recognition, describing anecdotal evidence where they recognised someone encountered as an adult when they had only briefly met that person as a child, perhaps 20 or 30 years ago. They know their experience is unusual when the other person fails to recognise them. Some police officers are also renowned amongst their colleagues for making difficult identifications or spotting suspects in a crowd. However, much evidence suggests that we have little insight into our face recognition skills, so some people may very good at face recognition without realizing it. Bournemouth University's work with Dorset Police has therefore developed a battery of computerized face recognition tests that can instantly identify officers with excellent face recognition skills. This objective data-based approach allows new recruits to be screened without the need for analysis of years of on-the-job performance or self-report measures.

How many people have extraordinary face recognition skills?

Statistical approaches to the identification of super recognisers dictate the answer to this question. Because researchers do not currently know have a biological indicator of extraordinary face recognition skills, performance on screening tests are analysed using pre-defined cut-offs. These cut-offs are calculated using the standard psychological approach of taking the value that is two standard deviations from the control mean. This approach will identify the top 2.5% of the population. This does not mean that the top 2.5% of the population are super recognisers by any categorical definition of the phenomenon, but simply that the tests identify the best performing 2.5% of the population on a particular face recognition task.

How could the police use super recognisers?

There are a number of applications of super recognition in the police force. These officers might be used to compare images of perpetrators that have been captured on CCTV, to decide whether different indecent images depict the same child, to "spot" wanted people or known criminals in a crowd at a football match or festival, or may be mobilised in times of global or national security to make more difficult identifications, perhaps when target faces are of different ethnicities or have undergone substantial changes in appearance. Super recognisers may also be used in cases of missing children, where an identification needs to be made after many years have passed.

How else might this research be useful to the police?

Our primary aim is to develop a thorough battery of screening tests that can assess officers for superior face recognition skills. Our research to date has found that not all super recognisers excel at all face recognition tasks, and it is likely that there are subtypes of the phenomenon. By using eye-movement and neuroimaging technology to examine the underpinnings of different types of super recognition, we hope to expand our theoretical understanding of the human face recognition system. By examining this system at its most efficient, we may be able to develop more appropriate training techniques that can be used to improve face recognition skills in all officers, and perhaps even those with face blindness.

For more information please visit www.prosopagnosiaresearch.org or contact Sarah directly at sbate@bournemouth.ac.uk 



The College of Policing maintains this micro site to enhance public access to specific and background information about the What Works Centre.  Our goal is to keep this information timely and accurate, and any errors brought to our attention will be corrected as quickly as possible. The What Works Centre micro site may contain content produced by third parties, and the College accepts no responsibility or liability whatsoever regarding the accuracy, timeliness, or applicability of any third party content.