In the 1960s, the predominant view of autistic disorder was that its origins were environmental rather than biological. It is all too easy to scoff at such views in the current climate, when genetic and neurobiological accounts of autism are commonplace and widely accepted, but in the light of the evidence they seemed reasonable. For a start, a genetic basis to autism did not seem plausible because the condition did not appear to run in families; there was no evidence of parent-to-child transmission, and it was unusual for more than one child in a family to be affected. Second, there was no indication of any gross neurological damage, and the children looked remarkably normal in physical appearance; unlike many other children of low IQ, their demeanour often suggested high intelligence, giving an impression that there were true abilities locked away beneath the surface. Furthermore, experienced clinicians noted certain deficiencies of social interaction in some parents of children with autism. Taken together, a logical conclusion was that autism was the result of a failure of the child to bond adequately with a parent, leading to a severe disturbance in social interaction. It has to be said that we are still a long way from understanding the neurobiological basis of autism: even with modern imaging techniques it is difficult to detect consistent abnormalities in the brains of children with autism, and the most plausible accounts, in terms of abnormal connectivity and/or deficient neurotransmitters, still lack strong evidence. Nevertheless, what has become clear is that this is a disorder with a genetic basis: The tide started to turn towards acceptance of this idea with a twin study by Folstein and Rutter (1977)
and was strengthened by subsequent studies showing an increased incidence of milder “autistic-like” features in relatives of affected children (see Rutter, 2000
, for a historical review). Nowadays it is accepted that social oddities do characterize a subset of parents of children with autism, but this is seen as evidence for a shared genetically determined trait in parent and child rather than as an indication of a psychogenic origin to autism. In the 1960s, however, when Uta started to work on autism, only a minority of experts were prepared to countenance the idea that it might be an “organic” condition, and she freely admits that her own conviction that we should look for brain bases rather than family origins was based more on hunch than on evidence. Fortunately, she found herself in an environment where there was sympathy for this hunch, and she was encouraged to do studies of basic perceptual and cognitive processes that might give a clue as to what distinguished these children from others. It is important to realize that autism research in the 1960s and 1970s was a very different enterprise from how it is today. The concept of “high-functioning autism” was not recognized, and the children who were the topic of study had severe cognitive limitations. Some lived in institutions, and others attended the handful of schools that specialized in educating such children, in particular the Sybil Elgar School in West London, which had been opened by the National Autistic Society in 1965. Doing experiments with such children required both stamina and ingenuity, and Uta addressed the difficult problem of devising tasks that would exploit the special interests and cognitive peaks of children with autism. Initially she learned by implementing studies devised by Ati Hermelin: these experiments, included in the classic text Psychological experiments with austistic children
(Hermelin & O'Connor, 1970
) were concerned with the extent to which children's ability to remember verbal or nonverbal material was influenced by meaning. The comparison group were younger children who were matched on memory span: the result, seen for both verbal and nonverbal material, was that meaning improved recall for both groups, but its effect was far stronger in the typically developing controls than in the children with autism. This demonstration of a distinctive cognitive profile strengthened Uta's conviction that we were dealing with a neurological impairment rather than a social inhibition caused by poor parenting. She went on to devise her own experiment, which considered whether children with autism were able to take account of a stress pattern when remembering words. This could be investigated in two ways: first, did children remember more words when a sequence had prosodic structure; second, did they impose a natural prosodic structure on sequences that were presented with no stress? It was found that children with autism took less notice of prosody than did control children (Aurnhammer-Frith, 1969
). This led on to work on pattern processing in language and in vision, including an ingenious study of pattern production that may be regarded as the first ever study demonstrating that children with autism had a deficit in generativity (the ability to generate ideas—the inverse of rigidity) (Frith, 1971a
Throughout her doctoral studies, Uta benefited from the wholehearted support of Ati Hermelin and Neil O'Connor, and from them she learned the skill of designing elegantly simple experiments that illuminated the nature of underlying cognitive deficits. Once her PhD was completed, they offered her a post at their new Medical Research Council (MRC) Developmental Psychology Unit in central London; however, it was clear that she was now expected to develop a new line of research and so, rather than building on the work she had done on autism, she embarked on a series of studies of literacy development. It was only after Neil O'Connor's retirement, when John Morton took over as director of the re-named MRC Cognitive Development Unit in 1982, that Uta returned to studies of autism. Although by this time a more neuropsychological and biological approach to autism was gaining acceptance, in large part due to the influence of Uta's mentors at the IOP, Lorna Wing and Michael Rutter, there was still widespread resistance to the notion that there was a genetic component to the disorder. A remarkable development was the publication of a book entitled Autistic children: New hope for a cure
by the Nobel laureate Niko Tinbergen and his wife Elisabeth (Tinbergen & Tinbergen, 1983
). Tinbergen's reputation was founded on ethological work, and he had the highly original idea that the same observational methods that he had applied to the study of birds could throw light on the problems of children with autism. Unfortunately, his ingenious interpretations of autistic behaviour, and his consequent recommendation of “holding therapy” as a cure, were highly impressionistic and completely lacking in scientific rigour. This therapy involves parents in holding their child for prolonged periods, even if the child is resisting the embrace. The parent tries to establish eye contact and to share feeling with the child throughout the session. Uta wrote a scathing review of the book, but it continued to have a significant impact, with holding therapy continuing to be advocated as a treatment in Continental Europe as well as in the United Kingdom.
There followed a remarkably productive period of research on distinctive cognitive deficits in children with autism, fostered by the unique intellectual atmosphere of the Cognitive Development Unit, where John Morton encouraged his staff to question and debate both theory and experiments. During the early 1980s Uta's research students, Simon Baron-Cohen, Tony Attwood, and Amitta Shah, all made important new discoveries about autism.
Simon Baron-Cohen's influential doctoral studies (Baron-Cohen, Leslie, & Frith, 1985
) were stimulated in part by Premack and Woodruff's (1978)
work on chimpanzee cognition, in which the term “Theory of Mind” was first used to refer to the cognitive capacity to understand that others may have beliefs, desires, or intentions that differ from one's own. Uta recognized that this had potential relevance for studies of autism, and that Wimmer and Perner's (1983)
studies of false belief in children provided an ideal paradigm for investigating Theory of Mind in this population. The basic Sally-Anne task seemed so easy that Uta did not anticipate that children with autism would have any problem in responding correctly, and she was amazed when this task proved such a sensitive tool for revealing the underlying deficits in social cognition. Alan Leslie joined the collaboration working on this topic, providing a theoretical viewpoint that regarded Theory of Mind as a subset of more general metarepresentational knowledge. Throughout this period, the group benefited from John Morton's insistence that it was not enough just to describe what had been found: it was important to place it in a theoretical context that would allow for the development of new predictions that could be tested experimentally. It is not exaggerating to say that the Theory of Mind work transformed the conceptualization of autism by psychologists. Prior to this work, psychological studies on autism were not very coherent. The earlier studies on perceptual processes and pattern perception by Frith and others were clearly important, but it was difficult to relate these findings to the symptoms of the disorder, especially the core social impairments. In the United Kingdom, much theoretical debate had focused on the question of whether autism could be explained as a language disorder (e.g., Churchill, 1978
), and new insights had been gained by explicit comparisons of children with autism and those with receptive language disorders (Bartak, Rutter, & Cox, 1975
), but the core features of social impairment remained unexplained. In the United States, Marian Sigman, Peter Mundy and their collaborators were doing detailed observational studies of social interaction in children with autism, noting deficiencies in shared attention and nonverbal communication (Sigman, Mundy, Sherman, & Ungerer, 1986
). The Theory of Mind work provided a way of conceptualizing social impairment as stemming from a cognitive failure in a system that computes representations of other minds and their contents. It not only made theoretical sense; it also provided new avenues for thinking about intervention, and many people working in special education found it helped them to understand why a child with autism might react in an unusual way to social situations. The fact that Theory of Mind could be assessed using a relatively simple experimental task must also have played a part in the enormous volume of research stimulated by the 1985 paper. This included fruitful collaborative studies with Josef Perner, who had hitherto worked solely on typical development, but who continued to develop new ways of looking at Theory of Mind, including the now-famous Smarties task (Perner, Frith, Leslie, & Leekam, 1989
Tony Attwood was a part-time graduate student who also worked as a clinical psychologist. Previous studies had suggested that children with autism made deficient use of gestures: Attwood's studies showed that this was an oversimplification, and that the function of gestures was key. His findings related neatly to the Theory of Mind theorizing, showing that instrumental gestures, which do not require any interpersonal understanding, were intact, whereas expressive gestures, which serve a purely communicative function, were never observed in children with autism (Attwood, Frith, & Hermelin, 1988
Amitta Shah's work was distinctively different. She focused on the unusually good performance on embedded figures and block design tasks in children with autism (Shah & Frith, 1983
). Interest in strengths as well as weaknesses in the cognitive profile of autism was a distinctive aspect of work at the Cognitive Development Unit, and was key to the development of the theory of weak central coherence.
In parallel with these exciting developments on the cognitive front, there was a gradual broadening of the concept of autism. An epidemiological study by Lorna Wing and Judith Gould had already drawn attention to the existence of a large number of children with social abnormalities who did not have classic Kanner syndrome but nevertheless shared many features with these cases (Wing & Gould, 1979
). A subsequent report noted that although most children with these characteristics were mentally retarded, this was not true for all (Wing, 1981a
). At the same time, Asperger's report of a syndrome akin to autism but accompanied by relatively good language skills was summarized in English (Wing, 1981b
), starting a lively debate as to whether this should be regarded as a subtype of autism or a distinct disorder. Meanwhile, as described above, genetic and family studies by Michael Rutter and associates were forcing researchers to the conclusion that there was a spectrum of autism, ranging from classic Kanner syndrome at one extreme to a broader phenotype at the other extreme, where the affected individual might function normally in society but have mild abnormalities of social interaction or communication. This was the start of a change in the nature of research on autism. Hitherto, cognitive studies had focused on children with intellectual retardation; gradually, as the concept of autism broadened, researchers turned more and more to the so-called high-functioning cases, who were far easier to work with and who, it was hoped, might throw autism-specific deficits into much sharper focus. Uta translated Asperger's text in 1991, and subsequently she became one of the world's experts in this syndrome.
One graduate student who took advantage of the broadened concept of autism to study cases of high-functioning autism and Asperger syndrome was Francesca Happé, who arrived at the Cognitive Development Unit in the late 1980s. Language in autism had always interested Uta, and she could see that investigation of the relationship between language and Theory of Mind would be a fruitful topic for study. Happé (1991)
used Sperber and Wilson's (1986)
Relevance Theory to generate predictions about how children with autism would interpret non-literal language and found an impressive fit between theory and data. For Uta, the puzzle of language in autism made sense if one recognized that the core deficits were pragmatic failures in the appreciation of relevance, with other aspects of language impairment (e.g. in structural aspects such as grammar and phonology) being correlates of the condition rather than key features (Frith, 1989b
In 1989 Autism: Explaining the enigma
, Uta's synthesis of more than two decades of work on cognitive bases of this disorder (Frith, 1989a
), appeared. She made a strong case for autism as a neurobiological disorder and presented compelling evidence for Theory of Mind as a core area of deficit, but she also drew attention to the a new idea concerning Weak Central Coherence as another aspect of the autistic mind. Unusually for an academic text, this was written in a clear, informal, and engaging style and accompanied by charming illustrations, and it immediately became a best-seller, with translations into numerous other languages.
Francesca Happé stayed on at the Cognitive Development Unit after completing her doctorate, and she and Uta worked together to consider how Weak Central Coherence related to other aspects of autistic cognition, including strengths in certain aspects of memory and perception as well as deficits in the ability to use context. This culminated in a detailed exposition of the Weak Central Coherence account by Frith and Happé in Cognition
in 1994 (see also Happé & Frith, 2006
, for an update of the theory).
The mid 1990s saw another major development that was to fundamentally change the course of Uta's research: the development of functional brain imaging. The earliest studies were done using the positon-emission tomography (PET) system at the MRC Cyclotron Unit at the Hammersmith Hospital, where Chris Frith was involved in developing activation paradigms using this new technology. Initially PET was used to probe the neural basis of Theory of Mind in normal adults (Fletcher et al., 1995
); subsequently the same paradigm was applied to adults with Asperger syndrome, who showed a specific decrease of activation in the medial prefrontal cortex, which was the region activated by Theory of Mind tasks in control participants (Happé et al., 1996
). Work on brain imaging gathered pace with the development of imaging facilities in central London at the Wellcome Department of Cognitive Neurology, where Chris Frith and his colleagues pioneered methods for analyzing brain activation in both PET and functional magnetic resonance imaging (fMRI) paradigms. Studies using fMRI provided yet more evidence for involvement of the medial frontal lobes in Theory of Mind (Gallagher, Frith, & Snowling, 2000
Chris Frith has always been a major influence in Uta's academic life, right from the earliest days at the IOP, when he helped her through her struggles with English language and culture. During 1970s and 1980s Uta and Chris would always discuss ideas and read one another's papers, but it was in the 1990s that their research collaboration really took off, as the new imaging techniques became increasingly important. In more recent years, their work has moved in a new and fruitful direction, with the integration of neuropsychological work on autism and schizophrenia. Both disorders involve abnormalities in social cognition as well as executive function impairments, but whereas people with autism under-utilize Theory of Mind, it appears that Theory of Mind is overactive in those with para-noid–schizophrenic symptoms, who see meaning and communication in situations where it is not intended. The cross-fertilization between research on these two disorders has led to innovative theoretical approaches to both autism and schizophrenia, engendering a line of work that promises to be enormously productive over the next few years.