The Recognition of Infantile Autism

Leo Kanner was the first person to clearly delineate the syndrome of “early infantile autism” (Kanner, Reference Kanner1943), which he identified as an inborn disturbance of affective contact defined by two key features: (a) a marked lack of social engagement from very early in life and (b) insistence on sameness and avoidance of change. This description essentially defined one very specific, and indeed the classic, phenotype of autism. Although “new” to the developing field of developmental psychopathology there is little doubt that cases had been seen (without being recognized) much earlier (e.g., in some of the training schools in the USA in the 1800s (Donvan & Zuker, Reference Donvan and Zuker2016)) and likely even in the much earlier reports of “feral” or “wild” children – most likely children with autism or severe neglect who had either been abandoned by their parents or who had run away from them (see; Anderson et al., Reference Anderson, Law, Daniels, Rice, Mandell, Hagopian and Law2012; Boswell, Reference Boswell1989; Candland, Reference Candland1995; Favazza, Reference Favazza1977). Many aspects of Kanner's description remain quite applicable today, but it is clear that some early speculations about autism were fundamentally wrong. Among the false leads for early research were the initial confusion of autism with childhood schizophrenia, Kanner's observation of very high levels of occupation and academic success in parents (this leading to the assumptions that parents might be involved in syndrome pathogenesis and that autism was not seen in the “lower social class”), and that autism was not associated with other medical conditions. Given what Kanner saw as a lack of dysmorphic features of these children and their ability to engage in some nonverbal tasks at very high levels, most notable was his early impression that infantile autism was not associated with intellectual disability (or, as it was then termed, mental retardation). This misimpression was corrected as follow-up studies became available and the first wave of identified children functioned, as adults, mostly in the severely impaired range with only a relative few achieving adult independence and self-sufficiency (Eisenberg & Kanner, Reference Eisenberg and Kanner1956; Howlin, Volkmar, Paul, Rogers, & Pelphrey, Reference Howlin, Volkmar, Paul, Rogers, Pelphrey, Volkmar, Rogers, Paul and Pelphrey2014; Kanner, Reference Kanner1971). Other early assumptions – that autism was a disorder associated with higher social status, an early form of childhood schizophrenia, and not associated with other medical conditions – also proved false over time (Jackson & Volkmar, Reference Jackson, Volkmar and Volkmar2019). For many years, research was limited, particularly by the confusion of autism with childhood-onset schizophrenia. However, it very quickly became apparent that despite very low levels of overall adaptive and cognitive functioning, some individuals – perhaps 10% – had some special skills or ability, usually involving some aspects of memory such as drawing, musical ability, calendar calculations, and so on (Hermelin, Reference Hermelin2001; Selfe, Reference Selfe1979; Wiltshire, Reference Wiltshire1988). Shortly after Kanner's initial publication, Kurt Goldstein and colleagues (Scheerer, Rothmann, & Goldstein, Reference Scheerer, Rothmann and Goldstein1945) reported the case of an 11-year-old boy with an IQ of 50 who had remarkable abilities in number manipulation, verbal memory, music, and calendar calculation. He was studied over a period of 5 years by means of standard tests, experiments, and a reliable record of his previous development. Despite his special skills, he was otherwise severely impaired in his daily life and cognitive ability. Their term for this condition (“idiot savant”) (Rimland & Fein, Reference Rimland and Fein1988; Treffert, Reference Treffert1988) has now effectively been changed to autistic savant as such cases are almost invariably associated with autism and intellectual disability. The observations of these islets of amazing preserved ability (in the face of otherwise severe cognitive dysfunction) immediately raised important questions about continuities and discontinuities with other types of intellectual deficiency and the question of the course of cognitive development in these cases.

Research on autism gradually began to accumulate as it became clear that autism was not associated with schizophrenia (Rutter, Reference Rutter1972), was seen in all social classes (Wing, Reference Wing1980), was highly genetic (Folstein & Rutter, Reference Folstein and Rutter1977), and was related to evidence of central nervous system dysfunction such as seizures (Volkmar & Nelson, Reference Volkmar and Nelson1990). It also became clear that the early psychodynamic treatments for autism were not effective, but structured teaching could be (Bartak & Rutter, Reference Bartak and Rutter1973), and evidence-based treatment approaches began to be developed at academic centers in the USA (National Research Council, 2001). These treatment models began to increase in breadth and sophistication, essentially with a focus on helping the developing child compensate for his/her lack of social connectedness (National Research Council, 2001).

Attempts were made to define the syndrome of infantile autism more carefully (National Society for Autistic Children, 1978; Rimland, Reference Rimland1964; Rutter, Reference Rutter1978) and, by 1980, infantile autism was first recognized as an official diagnostic category in the third edition of the Diagnostic and statistical manual of mental disorders (DSM-III) (American Psychiatric Association, 1980). This definition was, in many ways, very much in line with Kanner's original description, emphasizing the earliest form of the disorder both in its title “infantile autism” and the criteria it provided. The issue of potential developmental change was addressed with another category of “residual infantile autism” for cases that had once met the criteria for the infantile form of the condition but no longer did so. Given the need for a new “class” or category for the disorder, autism and associated conditions were placed in a new class of conditions – pervasive developmental disorders (PDD). Criteria were also given for childhood-onset forms of the disorder (i.e., after some years of normal early development) and for a subthreshold category “atypical PDD” for cases not meeting full criteria but with problems suggestive of autism (Cohen, Paul, & Volkmar, Reference Cohen, Paul and Volkmar1986).

DSM-III Infantile Autism – What is in and What was out?

At the point in 1980 when autism was first officially recognized, several points are of note. In the first place, it appeared that autism was one of the best examples of a “disorder” in child psychiatry (i.e., it did not seem to shade off into normalcy) (Garmezy & Rutter, Reference Garmezy and Rutter1983). Furthermore, autism, as then strictly defined, was relatively rare, impacting perhaps one in several thousand children and was much more common (three to five times) in boys than in girls (Fombonne, Reference Fombonne and Volkmar1997). In addition, at that point in time, most individuals (over 50%) were mute as adults and functioned in the intellectually impaired range although unusual patterns of cognitive abilities were frequently observed (Volkmar & Cohen, Reference Volkmar and Cohen1986) Many of these notions began to change as research on the condition increased following its official recognition in DSM-III (American Psychiatric Association, 1980), including a marked growth in research on genetics and neurobiology (Rutter, Reference Rutter2013). Greater awareness also led to more diagnoses and the potential for the implementation of evidence-based intervention (Reichow & Volkmar, Reference Reichow, Volkmar, Reichow, Doehring, Cicchetti and Volkmar2011). In turn, this appeared to lead to better outcomes in many cases, with some children making remarkable gains either in the preschool years or sometimes in adolescence (Fein et al., Reference Fein, Barton, Eigsti, Kelley, Naigles, Schultz and Tyson2013). Diagnostic tests (interviews and in-person assessments) began to be developed (Lord et al., Reference Lord, Corsello, Grzadzinski, Volkmar, Paul, Rogers, Pelphrey, Volkmar, Paul, Rodgers and Pelphrey2014), partly in the hope that consensus on the diagnostic concept could stimulate genetic research that identified underlying genes (Rutter & Thapar, Reference Rutter, Thapar, Volkmar, Paul, Rogers and Pelphrey2014). All these factors set the stage for a developing and now ongoing debate on the boundaries of autism and the autistic phenotype.

The Expanding Diagnostic Concept and the Broader Autism Spectrum

Perhaps somewhat paradoxically, agreement on the classic “Kannerian” view of autism in DSM-III (American Psychiatric Association, 1980) sparked a much larger debate about the boundaries of autism. These issues included a debate about the potential validity of other conditions (notably Asperger syndrome) and “atypical” cases – the beginning of the broader autism phenotype (BAP) debate – as well as related discussions of the need for a more developmentally appropriate definition of autism (as compared to the one in DSM-III) and the growing and expanding awareness of the complexities of genetic factors.

The debate about the BAP is, in many ways, a long-standing one. Both before and after Kanner, cases similar to those he described were reported. For example, Rank (Reference Rank1949, Reference Rank and Caplan1955) and Putnam and Rank (Reference Putnam and Rank1953) proposed a concept termed “atypical personality development” – a term that became remarkably similar to the term “atypical PDD” in DSM-III. More importantly, the year after Kanner's initial report (Kanner, Reference Kanner1943), Hans Asperger described a group of boys who had marked social troubles but preserved language abilities along with notable special skills and often a positive history of social difficulties in family members (Asperger, Reference Asperger1944). His same use of the word autism immediately suggested an important point of similarity although, throughout his life, even after becoming familiar with Kanner's description, he regarded the two conditions as distinctive (Asperger, Reference Asperger1979). The disorder then attracted little attention until Wing's review of the topic (Wing, Reference Wing1981), a year after DSM-III appeared. Subsequently, research interest gradually increased, although so did reports on best approaches to diagnosis of the condition with some advocating strict adherence to Asperger's original reports and others attempting to modify Wing's description so that several different conceptualizations can occur, although all viewed it as having some similarity to autism (see Klin, Volkmar, & Sparrow, Reference Klin, Volkmar and Sparrow2000; Szatmari, Reference Szatmari1991). These somewhat conflicting views further contributed to the lack of consensus on diagnosis, even after the publication of DSM-IV (American Psychiatric Association, 1994).

In addition, work on cases with many features of autism but not meeting full DSM-III criteria began to increase (Towbin, Reference Towbin, Cohen and Volkmar1997; Towbin, Reference Towbin, Volkmar, Rogers, Paul and Pelphrey2014). This work began, in important respects, to converge with a growing body of work in genetics (Rutter & Thapar, Reference Rutter, Thapar, Volkmar, Paul, Rogers and Pelphrey2014; Woodbury-Smith & Scherer, Reference Woodbury-Smith and Scherer2019; Yuenn et al., Reference Yuenn, Szatmari, Vorstman and Volkmar2019) showing the complexities of genetic contributions to autism and the potential for many forms of autism (Gillberg et al., Reference Gillberg, Allely, Bougeron, Coleman, Fernell, Hadjikhani, Sarovic and Volkmar2019).

As the term ASD came into common use and began to replace the term PDD (Lord, Reference Lord, Cohen and Volkmar1997), there began to be recognition of a BAP, suggesting that – like many other conditions – autism was associated with a range of problems in family members (Ingersoll et al., Reference Ingersoll, Wainer, Volkmar, Rogers, Paul and Pelphrey2014; Piven, Palmer, Jacobi, Childress, & Arndt, Reference Piven, Palmer, Jacobi, Childress and Arndt1997), thus questioning the early notion that autism did not shade off into normalcy (Ingersoll et al., Reference Ingersoll, Wainer, Volkmar, Rogers, Paul and Pelphrey2014; Piven et al., Reference Piven, Palmer, Jacobi, Childress and Arndt1997) but was associated with a range of problems both developmental and behavioral in nature.

From DSM-III to DSM-5

Although DSM-III helped stimulate an explosion of research on autism, problems with it were apparent. These included a lack of developmental awareness, an overfocus on “infantile” autism, limitations of the provided monothetic criteria, and the rationale for other disorders included in the PDD category (e.g., residual infantile autism and childhood-onset PDD) (Volkmar, Reference Volkmar1987; Wing, Reference Wing and Gillberg1989a, Reference Wing and Gillberg1989b; Wing & Gould, Reference Wing and Gould1979). Attempts were made to address some of these problems in DSM-III-R (American Psychiatric Association, 1987), although this version was quickly replaced by the fourth edition of the Diagnostic and statistical manual of mental disorders, DSM-IV (American Psychiatric Association, 1994). The definition for autistic disorder (as it was now termed) was developed in collaboration with that for the International Classification of Diseases (ICD-10) (World Health Organization, 1994) and based on a large multi-site, international field trial (Volkmar et al., Reference Volkmar, Klin, Siegel, Szatmari, Lord, Campbell and Kline1994). This definition included 12 criteria grouped into the traditional three areas of problems in (a) social interaction, (b) communication and play, and (c) restricted and repetitive behaviors and interests apparent from early in life. In this new approach, over 2,000 combinations of criteria could lead to an autism diagnosis. It appeared to cover the range of ages and intellectual functioning well, and converged well with expert clinician diagnoses. In addition, both DSM-IV and ICD-10 included a number of other “new” conditions, including Asperger disorder, a category for the very rare cases of autism that developed after a period of normal development, and a subthreshold category of atypical autism/pervasive developmental disorder – not otherwise specified (PDD-NOS).

The recognition and definition of Asperger disorder was the focus of considerable debate (Ghaziuddin, Reference Ghaziuddin2011; Mayes, Calhoun, & Crites, Reference Mayes, Calhoun and Crites2001; Woodbury-Smith, Klin, & Volkmar, Reference Woodbury-Smith, Klin and Volkmar2005). Depending on the diagnostic approach used, potentially important differences were, however, found from autism relative to family history and neuropsychological profiles (Klin, Volkmar, Sparrow, Cicchetti, & Rourke, Reference Klin, Volkmar, Sparrow, Cicchetti and Rourke1995; Woodbury-Smith et al., Reference Woodbury-Smith, Klin and Volkmar2005). The DSM-IV system worked well for nearly two decades and stimulated a massive body of work on autism as well as on conditions such as Asperger syndrome and the BAP.

The entire DSM was revised in a fifth edition, published in 2013. DSM-5 (American Psychiatric Association, 2013) differed from its predecessor in many ways. Some important, positive changes were made relative to modifications to categories and their definitions. Both of these, however, were not without controversy (Greenberg, Reference Greenberg2013). The term ASD was adopted both for the condition itself and the category to which it belonged. It was observed – probably not surprisingly – that approaches to the diagnosis of Asperger syndrome and PDD-NOS varied considerably from site to site, so these terms were dropped in favor of the overall diagnostic term (Happe, Reference Happe2011; Lord et al., Reference Lord, Petkova, Hus, Gan, Lu, Martin and Risi2012). The new approach to diagnosis for ASD grouped social communication into a single criterion set, all of which had to be met, and a second set of four restricted interests/repetitive behavior criteria, two of which had to be met. The latter also included, for the first time, an item on sensory hypo/hypersensitivity as a defining feature of the condition (which had not worked well in the DSM-IV field trial (Volkmar et al., Reference Volkmar, Klin, Siegel, Szatmari, Lord, Campbell and Kline1994)). The initial report of the DSM-5 approach (Huerta, Bishop, Duncan, Hus, & Lord, Reference Huerta, Bishop, Duncan, Hus and Lord2012) suggested that it had a good balance of sensitivity and specificity although in fact specificity relative to the PDD-NOS was fairly poor. A new communication disorder category (social communication disorder or SCD) was also included in DSM-5 for individuals with problems in pragmatic communication who did not have the difficulties with restricted interests observed in ASD.

Unfortunately, a number of issues were immediately apparent. As a practical matter, this new approach vastly reduced the number of criterion combinations that could lead to an ASD diagnosis (over 2,000 in DSM-IV and 12 in DSM-5). In addition, the actual term for the class of conditions – ASD – while welcome in many ways, was in the singular (i.e., it failed to recognize the breadth of the autism spectrum) (Ingersoll et al., Reference Ingersoll, Wainer, Volkmar, Rogers, Paul and Pelphrey2014). The first study of the draft DSM-5 criteria set provided worrisome results (Matson, Kozlowski, Hattier, Horovitz, & Sipes, Reference Matson, Kozlowski, Hattier, Horovitz and Sipes2012). Subsequently, a reanalysis of cases from the DSM-IV field trial, but using the final new DSM-5 criteria set, confirmed and expanded these concerns (McPartland, Reichow, & Volkmar, Reference McPartland, Reichow and Volkmar2012). In the McPartland et al. study, many individuals with clinician-assigned diagnoses of autism who functioned within the average intellectual range lost their diagnosis, as did the vast majority of those with Asperger syndrome or PDD-NOS. Concerns were also quickly raised about the exclusion of toddlers from early intervention because of the stringency of the new approach (Barton, Robins, Jashar, Brennan, & Fein, Reference Barton, Robins, Jashar, Brennan and Fein2013; Matson et al., Reference Matson, Kozlowski, Hattier, Horovitz and Sipes2012). A number of studies have now confirmed these findings (for a meta-analysis, see Smith, Reichow, and Volkmar, Reference Smith, Reichow and Volkmar2015). These concerns resulted in a last-minute change to DSM-5 that “grandfathered in” cases that previously held “well established” diagnoses of conditions such as Asperger syndrome, PDD-NOS, and autistic disorder. This solution did not, of course, help those individuals just attaining a diagnosis and had a particular impact in the USA where the diagnostic label is often closely associated with intervention services. In summary, although presenting some advances, the newer DSM-5 definition, despite the name change, has reverted to a narrower view of the diagnostic concept. This narrower DSM-5 approach has ramifications for the classification of major social disabilities in special groups such as women, very young children, culturally diverse groups, and the broader autism spectrum in general. This is particularly unfortunate since a growing body of research indicates that, with identification and treatment, many children can make important gains and no longer suffer the associated problems of intellectual disabilities once so common in autism (Magiati & Howlin, Reference Magiati, Howlin and Volkmar2019)

Current Controversies and Complexities

Since DSM-5 appeared in 2013, a number of controversies have continued or emerged. One is that the problem of diagnosing Asperger syndrome remains unresolved. There is general agreement that the DSM-IV definition could have been substantially improved and that authors have essentially adopted their own forms of it, making it unsurprising that different centers use the term in different ways (Woodbury-Smith et al., Reference Woodbury-Smith, Klin and Volkmar2005, #28287). In the meantime, the number of studies on Asperger syndrome dramatically increased, and new data were emerging just as the decision was made to drop the category. For example, Chiang and colleagues conducted a large meta-analysis of autism and Asperger syndrome and found different neuropsychological profiles associated with the two conditions (Chiang, Tsai, Cheung, Brown, & Li, Reference Chiang, Tsai, Cheung, Brown and Li2014). Similarly, Sharma et al. reviewed nearly 70 research studies using various approaches to diagnosis of Asperger syndrome and noted areas of overlap with autism as well as some important potential differences (Sharma, Woolfson, & Hunter, Reference Sharma, Woolfson and Hunter2012). Other studies have also supported the recognition of Asperger syndrome in some form as an official category (de Giambattista et al., Reference de Giambattista, Ventura, Trerotoli, Margari, Palumbi and Margari2019; Montgomery et al., Reference Montgomery, Allison, Lai, Cassidy, Langdon and Baron-Cohen2016), and the term has attracted such widespread attention that it has now entered into everyday discourse. As Greenberg (Reference Greenberg2013) notes substantial numbers of individuals had identified with this label effectively brining it into common use regardless of DSM terminology. At the moment, the category is in somewhat of a state of limbo, with cases diagnosed before DSM-5 retaining their diagnosis and new cases no longer identified (but continuing to exist). The best approaches to including and defining Asperger syndrome clearly is an important topic for the next iteration of the DSM.

Other issues arose with the dropping of the subthreshold PDD-NOS concept and the inclusion of SCD partly in its place. The large majority of cases of PDD-NOS as diagnosed according to DSM-IV no longer meet the criteria for autistic disorder in DSM-5, but continue to merit service provision and support. It is not simply the case that these cases automatically “make” the new SCD label, given that cases with restricted, repetitive behaviors are excluded from that diagnosis (for contrasting views, see Ozonoff & Miller, Reference Ozonoff and Miller2014; Tufan, Reference Tufan2014). The entire rationale for this concept has been questioned (Flusberg, Reference Tager-Flusberg2018). In effect, this very large group of cases no longer has a diagnostic “home” in DSM-5, even as their frequency is increasingly realized and likely accounts for at least some of the apparent dramatic increase in cases of autism in epidemiological surveys (Meyers, Chavez, Hill, Zuckerman, & Fombonne, Reference Meyers, Chavez, Hill, Zuckerman, Fombonne and Volkmar2019). Follow-up studies of young children with significant pragmatic deficits into later life are clearly needed.

Developmental Issues

The first reports of autism were of school-age children, with this group – particularly boys of borderline and mild intellectual disability – traditionally the general target group of diagnostic instruments (Lord et al., Reference Lord, Corsello, Grzadzinski, Volkmar, Paul, Rogers, Pelphrey, Volkmar, Paul, Rodgers and Pelphrey2014). The challenge for diagnostic criteria is, of course, to address the entire age range of the condition. For autism, this has been a challenge at times, both for adults and infants and young children. Given that autism is clearly an early onset disorder, the latter problem might seem somewhat paradoxical.

Autism in Infants and Young Children

As noted previously, DSM-5 presents some challenges for diagnosis in younger children, particularly those who are more cognitively able. For all individuals, both the social–communication problems of autism and the restricted behavior features must be met before an ASD diagnosis can be made. Just as in older children and adults, there is a wide range of heterogeneity, but even when antedated by parental concern, the more obvious behavioral manifestations of ASD emerge for most children in the second year of life and are often initially expressed in the realm of social–communicative features rather than restricted behaviors (i.e., lowered levels of social communication, limited social attention, and lack of social engagement and eye contact) (Chawarska, Klin, Paul, Macari, & Volkmar, Reference Chawarska, Klin, Paul, Macari and Volkmar2009; Chawarska et al., Reference Chawarska, Shic, Macari, Campbell, Brian, Landa and Young2014; Miller et al., Reference Miller, Iosif, Hill, Young, Schwichtenberg and Ozonoff2017). Delays in these areas can also be noted in other conditions such as intellectual disability and language disorders, meaning that making use of appropriate comparison groups is critical in evaluating best approaches to diagnosis (Macari et al., Reference Macari, Powell, Lyons, Saulnier, Vernetti, Chawarska, Chawarska and Volkmar2020; Manwaring, Stevens, Mowdood, & Lackey, Reference Manwaring, Stevens, Mowdood and Lackey2018).

Studies using high-risk groups (e.g., infant siblings of those diagnosed) also show that some siblings will have early difficulties suggestive of ASD but may then go on to develop more typically or to become part of the BAP (Georgiades et al., Reference Georgiades, Szatmari, Zwaigenbaum, Bryson, Brian, Roberts and Garon2013; Macari et al., Reference Macari, Campbell, Gengoux, Saulnier, Klin and Chawarska2012; Ozonoff et al., Reference Ozonoff, Young, Belding, Hill, Hill, Hutman and Iosif2014).

Studies of this group of high-risk siblings have helped to clarify the earliest features of ASD. By 12 months of age, those who later go on to have a diagnosis of autism often show atypical social–communicative responsivity and expression (Chawarska, Macari, Vernetti, & Brunissen, Reference Chawarska, Macari, Vernetti, Brunissen, Chawarska and Volkmar2020). The restricted interests/repetitive behavior domain behaviors begin to appear at the same time early in life, but also are seen in typically developing infants where they do not, however, consistently engage in attention with an adult social partner (Chawarska et al., Reference Chawarska, Macari, Vernetti, Brunissen, Chawarska and Volkmar2020; Chawarska et al., Reference Chawarska, Shic, Macari, Campbell, Brian, Landa and Young2014). Unfortunately, these generalizations do not always apply in individual cases (Elison et al., Reference Elison, Wolff, Reznick, Botteron, Estes, Gu and Piven2014;

As noted previously, the approach adopted in DSM-5 is rather less flexible than in DSM-IV (Mandy, Charman, & Skuse, Reference Mandy, Charman and Skuse2012). As a result, the youngest children who go on to have ASD may not yet have met the criteria for the condition. In one sample of 1- to 3-year-olds, 72% of those diagnosed who would have had PDD diagnosed in DSM-IV did not meet ASD criteria using DSM-5 (Mayes, Black, & Tierney, Reference Mayes, Black and Tierney2013). In general, the loss of the more flexible polythetic and three-domain approach in DSM-IV seems to have had a major impact on this group of cases (Barton et al., Reference Barton, Robins, Jashar, Brennan and Fein2013).

Autism in Adolescents and Adults

Another set of issues pertain to adolescents and adults for whom the question of an ASD diagnosis arises for the first time. This issue has become more frequent with awareness of the BAP and an awareness that individuals with milder autism traits may not be diagnosed until adulthood (Lai & Baron-Cohen, Reference Lai and Baron-Cohen2015). Often, these are very cognitively able individuals who may present with complex mental health histories. As we have noted elsewhere (Volkmar, Woodbury-Smith, Macari, & Oien, Reference Volkmar, Woodbury-Smith, Macari, Oien, McPartland and Stavropoulosin press), there are several issues arising. How well do current criteria apply to this group? Do unusual diagnostic assessment instruments assist with diagnosis? What are the needs of this group of cases?

Adults who present with a question of a possible ASD diagnosis are a diverse group and present several challenges (Huang, Arnold, Foley, & Trollor, Reference Huang, Arnold, Foley and Trollorin press; Lai & Baron-Chen, Reference Lai and Baron-Cohen2015). Sometimes, symptoms are relatively mild and only come to attention in new situations, such as starting college. This can sometimes be because learning demands change and students need support: in one study by White and colleagues (White, Ollendick, & Bray, Reference White, Ollendick and Bray2011), between 0.7% and 1.9% of a large freshman sample met the criteria for ASD but had not been previously diagnosed. In other cases, an adult who is functioning well enough will be able to manage with supports, but when these disappear or radically change, the extent and severity of their needs become apparent. As we discuss shortly, a number of these cases are women for whom it is possible that a milder phenotype delays diagnosis until later in life. In addition, adults are usually seen in mental health services, and some will already have other diagnoses.

In adults with diagnoses later in life, there may be different patterns of syndrome expression (e.g., where special interests and routines have some advantages for occupational success). However, many adult mental health providers are much less likely to recognize the underlying social disability so that misdiagnosis and inappropriate treatment may be more common (see Volkmar et al., Reference Volkmar, Woodbury-Smith, Macari, Oien, McPartland and Stavropoulosin press). Application of the more child-centered DSM-5 criteria may be problematic (Magiati & Howlin, Reference Magiati, Howlin and Volkmar2019). It does appear that some of the standard assessment instruments may not work as well in adults as in children, while others may work reasonably well (Fusar-Poli et al., Reference Fusar-Poli, Brondino, Rocchetti, Panisi, Provenzani, Damiani and Politi2017). Several adult-focused assessments are now available to address this problem (Volkmar et al., Reference Volkmar, Woodbury-Smith, Macari, Oien, McPartland and Stavropoulosin press). Sadly, however, appropriate supports and services are often lacking (Huang et al., Reference Huang, Arnold, Foley and Trollorin press).

Gender

As already noted, issues of gender, specifically the nature of the ASD phenotype in females, have been the focus of increasing debate. For many years the apparent preponderance of boys with autism meant that much less attention was focused on girls, with some studies excluding girls from their sample to try and avoid technical complexities like gender matching. In addition, both the early diagnostic criteria and assessments were developed with a focus on school-age boys with mild to borderline levels of intellectual disability. As a result, important aspects of the female phenotype have been overlooked. As we discuss elsewhere (Volkmar et al., Reference Volkmar, Woodbury-Smith, Macari, Oien, McPartland and Stavropoulosin press), there has been an implicit assumption that diagnostic criteria are gender neutral. The available literature on this topic suggests that females with ASD must have greater degrees of impairment and symptomatology to meet the cutoff for diagnosis (Dworzynski, Ronald, Bolton, & Happé, Reference Dworzynski, Ronald, Bolton and Happé2012). This is of interest as, in general, females, tend to present with more internalizing symptoms profiles than males (Mandy, Charman, & Skuse, Reference Mandy, Charman and Skuse2012). It has also been suggested that, compared with males, females are more likely to be able to engage in “camouflaging” or “pretending to be normal,” which could also account for later presentation and diagnosis (Bargiela, Steward, & Mandy, Reference Bargiela, Steward and Mandy2016). Several researchers have questioned the long-standing assumption of male predominance in ASD (Becker, Reference Becker2012; Constantino & Charman, Reference Constantino and Charman2012; Haney, Reference Haney2016).

Over the last decades, several theories have been proposed to account for what appears to be this male predominance of ASD. One theory is that females must present with greater symptom severity and lower cognitive abilities to raise concern and ultimately receive a diagnosis (Volkmar, Szatmari, & Sparrow, Reference Volkmar, Szatmari and Sparrow1993). Others have speculated that there may be a genetic protective effect of being female (i.e., females have a genetic protection from ASD) (Robinson, Lichtenstein, Anckarsäter, Happé, & Ronald, Reference Robinson, Lichtenstein, Anckarsäter, Happé and Ronald2013). However, it must then also be questioned whether this is the fact for all neurodevelopmental disorders where the male to female ratio is predominant towards males. It does generally appear to be the case that females as a group exhibit some noteworthy phenotypic differences from males. For example, they exhibit lower levels of restrictive, repetitive behaviors (Charman et al., Reference Charman, Loth, Tillmann, Crawley, Wooldridge, Goyard and Baron-Cohen2017; Frazier, Georgiades, Bishop, & Hardan, Reference Frazier, Georgiades, Bishop and Hardan2013), less severe sensory issues and better joint attention (Oien et al., Reference Oien, Hart, Schjolberg, Wall, Kim, Nordahl-Hansen and Shic2017), and a lesser degree of social avoidance (Oien et al., Reference Oien, Vambheim, Hart, Nordahl-Hansen, Erickson, Wink and Grodberg2018). These differences in phenotypic expression could result in less pronounced ASD symptoms (which can affect parental concern), failure to meet usual diagnostic criteria, or being below symptom thresholds on the usual assessment instruments. It is noteworthy that in large-scale studies of screening instruments, sex differences are observed very early in life (Oien et al., Reference Oien, Vambheim, Hart, Nordahl-Hansen, Erickson, Wink and Grodberg2018).

Cultural Issues

Increased interest in ASD research and treatment has increased dramatically in developing countries in recent years (Freeth, Milne, Sheppard, & Ramachandran, Reference Freeth, Milne, Sheppard, Ramachandran, Volkmar, Paul, Rogers and Pelphrey2014) As interest has increased, so have issues of cultural similarities and differences in the expression, assessment, and treatment of autism and related conditions. While the body of work on this topic is relatively small, there clearly are some areas of potential difference, for example regulation of eye contact across cultures (Freeth et al., Reference Freeth, Milne, Sheppard, Ramachandran, Volkmar, Paul, Rogers and Pelphrey2014) that might impact the usual Western-oriented assessments. As Freeth and colleagues rightly note, the bulk of research has been conducted within well-developed Western cultures and Japan. Often, screening and diagnostic instruments have been translated but sometimes with mixed results in non-Western countries (Freeth et al., Reference Freeth, Milne, Sheppard, Ramachandran, Volkmar, Paul, Rogers and Pelphrey2014).

Within the USA, studies of ethnic minority groups have suggested some potential complexities in the application of usual models/instruments in the diagnosis of autism. In one study of Hispanic children, issues arose with the application of usual research diagnostic instruments, particularly around the area of communication resulting in lower sensitivity and specificity (Vanegas, Magaña, Morales, & McNamara, Reference Vanegas, Magaña, Morales and McNamara2016), and similar issues have been noted by others studying minority groups in the USA (Mandell, Ittenbach, Levy, & Pinto-Martin, Reference Mandell, Ittenbach, Levy and Pinto-Martin2007; Palmer, Walker, Mandell, Bayles, & Miller, Reference Palmer, Walker, Mandell, Bayles and Miller2010; Tek and Landa, Reference Tek and Landa2012). Screening instrument results may also vary in different minority groups (e.g., Dai et al., Reference Dai, Porto, Skapek, Barton, Dumont-Mathieu, Fein and Robins2020; Khowaja, Hazzard, & Robins, Reference Khowaja, Hazzard and Robins2015; Kimple, Bartelt, Wysocki, & Steiner, Reference Kimple, Bartelt, Wysocki and Steiner2014; Rea, Armstrong-Brine, Ramirez, & Stancin, Reference Rea, Armstrong-Brine, Ramirez and Stancin2019; Windham et al., Reference Windham, Smith, Rosen, Anderson, Grether, Coolman and Harris2014).

Differences have been noted more generally in the application of DSM-5 in different developed Western countries; for example, the system seemed to work well in the UK but not so in Finland (Mandy, Charman, & Skuse, Reference Mandy, Charman and Skuse2014). One of the great complexities of research in this area is the broader meaning and context of studies of how autism may be viewed differently across cultures (Grinker et al., Reference Grinker, Kang-Yi, Ahmann, Beidas, Lagman and Mandell2015).

Summary

Professor Ed Zigler was one of the first to investigate the potential implications of experience and endowment for all the various forms of intellectual deficiency. His work on the role of biological factors and psychopathology, best approaches to treatment, and the nature of cognitive processes (both similar to and different from those in typical development) has stimulated an entire body of research. This work is now moving into the area of autism and related conditions where it has become increasingly clear that, in a very similar way, many different forms of autism can be identified – all pulled together by the central of a significant social learning challenge that impacts growth in other areas of a child's life. In this paper we focused on two important issues. First, despite the move to a better name for autism (ASD), the definition in DSM-5 has actually narrowed the diagnostic concept. Second, there is an emerging awareness of important diagnostic difficulties and tensions relative to issues of age and developmental level, gender, and culture. All these have yet to be adequately addressed. Future approach to the diagnosis may be guided by specific connections to known genetic causes and associated neuropsychological profiles. In sum DSM-5 was ideally placed to drive both the research agenda and clinical service but its attempt to cover both has fallen short and fails to resolve important diagnostic controversies.