Jaswal & Akhtar (J&A) consider four behaviors typical of autistic people, challenging the interpretation that is often attributed to these behaviors. While there is general agreement as to these autistic atypicalities, J&A claim that misinterpreting the motivation behind then has a negative impact on autism intervention and research. However, the way the authors use the term autism, offering no further qualifications, renders their claims uninterpretable. In fact, generalized references to “autistic people,” such as appear in the current article, have had negative bearings on research related to the causes of the disorder, impeding translational impact (Müller & Amaral Reference Müller and Amaral2017).

J&A use “autistic” to refer to people who have received a diagnosis of autism (footnote 1). But this is a highly heterogeneous group that eludes phenotypic as well as neurobiological description. The Diagnostic and Statistical Manual of Mental Disorders (5th ed., known as DSM-5; American Psychiatric Association 2013) has gone a long way in acknowledging this heterogeneity, taking note of levels of severity of autism, along with language status, IQ, and personal history (e.g., age of diagnosis, additional deficits). People with syndromic autism, that is, people who are diagnosed with a neurodevelopmental disorder and autism, whose autism typically varies according to the comorbid diagnosis, are also part of the spectrum.

Most notable is the fact that autism phenotype can vary even among monozygotic twins, not only in comorbid features but also in core autism spectrum disorder (ASD) parameters and severity measures. Furthermore, variation is seen in time of onset of ASD as about 25–40% of children regress after 2 years of seemingly normal development, about 60% have intellectual impairments (for a detailed review and meta-analysis of the prevalence of regression in ASD, see Pearson et al. Reference Pearson, Charman, Happé, Bolton and McEwen2018) and 25–40% have minimal or no language at all. Attempts to account behaviorally for autism subtypes did not demonstrate course of development or treatment validity and showed low interrater agreement in assessment (Lord et al. Reference Lord, Petkova, Hus, Gan, Lu, Martin, Ousley, Guy, Bernier, Gerdts, Algermissen, Whitaker, Sutcliffe, Warren, Klin, Saulnier, Hanson, Hundley, Piggot, Fombonne, Steiman, Miles, Kanne, Goin-Kochel, Peters, Cook, Guter, Tiernagel, Green-Snyder, Bishop, Ester, Gotham, Luyster, Miller, Olsonk, Richler and Risi2012).

Importantly, along with the diagnostic features of ASD, the non-diagnostic deficits accompanying ASD occur in one combination or another in all, or nearly all, people diagnosed with ASD, while social-communicative difficulties and especially routine-repetitive behaviors are not infrequent in young typically developing children. It is generally accepted that the extensive behavioral heterogeneity that is seen in autism exceeds that which characterizes other psychiatric disorders (Waterhouse Reference Waterhouse2013).

Virtually every aspect of autism that has been studied reflects an unexplained diversity. To date, there are reports of structural and mutational variations of between 200 and 1,000 genes involved in ASD susceptibility (Berg & Geschwind Reference Berg and Geschwind2012). The contribution of common genetic variants to ASD is likely mediated by a large, heterogeneous number of mutations, each contributing a minute risk of the disorder. Still, although it is estimated that while 15–40% of the risk of ASD is accounted for by common variants, only two genome-wide significant loci have been found. Furthermore, even these loci have not been straightforwardly replicated (Anney et al. Reference Anney, Klei, Pinto, Almeida, Bacchelli, Baird, Bolshakova, Bölte, Bolton, Bourgeron, Brennan, Brian, Casey, Conroy, Correia, Corsello, Crawford, de Jonge, Delorme, Duketis, Duque, Estes, Farrar, Fernandez, Folstein, Fombonne, Gilbert, Gillberg, Glessner, Green, Green, Guter, Heron, Holt, Howe, Hughes, Hus, Igliozzi, Jacob, Kenny, Kim, Kolevzon, Kustanovich, Lajonchere, Lamb, Law-Smith, Leboyer, Le Couteur, Leventhal, Liu, Lombard, Lord, Lotspeich, Lund, Magalhaes, Mantoulan, McDougle, Melhem, Merikangas, Minshew, Mirza, Munson, Noakes, Nygren, Papanikolaou, Pagnamenta, Parrini, Paton, Pickles, Posey, Poustka, Ragoussis, Regan, Roberts, Roeder, Roge, Rutter, Schlitt, Shah, Sheffield, Soorya, Sousa, Stoppioni, Sykes, Tancredi, Thompson, Thomson, Tryfon, Tsiantis, Van Engeland, Vincent, Volkmar, Vorstman, Wallace, Wing, Wittemeyer, Wood, Zurawiecki, Zwaigenbaum, Bailey, Battaglia, Cantor, Coon, Cuccaro, Dawson, Ennis, Freitag, Geschwind, Haines, Klauck, McMahon, Maestrini, Miller, Monaco, Nelson, Nurnberger, Oliveira, Parr, Pericak-Vance, Piven, Schellenberg, Scherer, Vicente, Wassink, Wijsman, Betancur, Buxbaum, Cook, Gallagher, Gill, Hallmayer, Paterson, Sutcliffe, Szatmari, Vieland, Hakonarson and Devlin2012).

De novo mutations are also involved in ASD. Genetic susceptibility to ID, ASD, attention-deficit/hyperactivity disorder (ADHD), and schizophrenia often arises from de novo mutations in the same genes, suggesting that these disorders share common mechanisms (Fromer et al. Reference Fromer, Pocklington, Kavanagh, Williams, Dwyer, Gormley, Georgieva, Rees, Palta, Ruderfer, Carrera, Humphreys, Johnson, Roussos, Barker, Banks, Milanova, Grant, Hannon, Rose, Chambert, Mahajan, Scolnick, Moran, Kirov, Palotie, McCarroll, Holmans, Sklar, Owen, Purcell and O'Donovan2014) as well as common affected pathways (Ben-David & Shifman Reference Ben-David and Shifman2012). De novo mutations in ASD as well as in ID are seen in genes that are relatively immune to mutations in the general population (Samocha et al. Reference Samocha, Robinson, Sanders, Stevens, Sabo, McGrath, Kosmicki, Rehnström, Mallick, Kirby, Wall, MacArthur, Gabriel, DePristo, Purcell, Palotie, Boerwinkle, Buxbaum, Cook, Gibbs, Schellenberg, Sutcliffe, Devlin, Roeder, Neale and Daly2014). Importantly, while reporting common mutations in ASD, ID, and schizophrenia, Shohat et al. (Reference Shohat, Ben-David and Shifman2017) found gene expression patterns that were specific to each disorder. They suggest that convergence among disorders is due to pathways that are affected by mutated genes, while diverse loci of gene expression contribute to specific phenotypes.

Diagnostic as well as non-diagnostic symptoms of autism such as ID, language disorders, ADHD, and seizures often occur with a variety of brain impairments, likely reflecting the complexity of the behavioral manifestations of social-communicative behavior and the varied manifestations of repetitive behaviors. Furthermore, ASD diagnosis or elements thereof have been found with many developmental syndromes, including single-gene Mendelian syndromes that have a variety of genetic and brain bases (Arnett et al. Reference Arnett, Trinh and Bernier2019).

Studies of the neuropathology of ASD have identified abnormal brain growth trajectories and disordered cortical organization and subcortical connectivity. In a recent review article, Waterhouse et al. (Reference Waterhouse, London and Gillberg2016) provide a representative sample of the heterogeneous and often contradictory findings that characterize attempts to uncover the brain abnormalities that characterize people with autism. A summary of this work is beyond the scope of this commentary. Suffice it to say that the general conclusion is that ASD lacks biological and construct validity. Waterhouse et al. (Reference Waterhouse, London and Gillberg2017, p. 1182) boldly suggest that the diagnosis of ASD, which they refer to as an arbitrary, unscientific “convenient fiction,” should be abandoned in research. Müller and Amaral (Reference Müller and Amaral2017) offer a more conservative view. They agree that the inadequacy of behaviorally defined ASD mandates a change in research, without which there is little hope for translational progress. Nevertheless, Müller and Amaral argue that for the time being, there is much to be gained by continuing to work along the lines of the clinical diagnosis of autism.

All of the above and more stress the need to qualify reference to “autistic people” in research, in intervention and in daily conversations, so as to avoid the impression that autism refers to a unitary group of affected people. The claims made by J&A would be meaningful if we knew more about the subgroups of people with a diagnosis of autism with respect to whom these claims are made.