Brain changes in adolescence-it is about time to get serious in autism spectrum disorder research.

03Original abstract

In a study published in this issue, Lawrence and colleagues examine age-related trajectories of brain network organization in teens with autism spectrum disorders (ASDs) and typically developing (TD) peers, using resting state functional connectivity MRI (rs-fcMRI). Between approximately ages of 13 and 16 years, TD adolescents show changes in three selected networks (central executive, salience, and default mode) that are consistent with expected maturational changes. Such changes are not seen in teens with ASDs. More remarkably, the developmental increase in anti-correlations between central executive and default mode networks in TD adolescents, presumably reflecting enhanced network segregation and specialization, is reversed in adolescents with ASDs. While the reported findings surely have specific relevance to those who believe that brain network connectivity (rather than any localizing feature) is well suited to account for ASD symptomatology, the study by Lawrence et al. has broader significance. Numerous functional neuroimaging studies have included teens in the age range examined by Lawrence et al., and some of these have tested for age-related changes (e.g., Nomi & Uddin, 2015). However, these studies have been exclusively cross-sectional. Lawrence and colleagues, on the other hand, acquired longitudinal data to test for developmental changes. What may appear to be a purely methodological advance has in fact important conceptual implications. There is a critical need for such data that goes beyond the general strengths of longitudinal designs in reducing uncontrolled variance of no interest. This is related to the known heterogeneity of ASDs and the possibility that cross-sectional findings may be confounded by cohort effects. While it is possible to control for some basic demographic and cognitive factors (e.g., gender, handedness, IQ), others can usually not be modeled with sufficient precision (e.g., history of interventions, medications), although they likely affect functional connectivity measures (Linke, Olson, Gao, Fishman, & Müller, 2017) and may therefore confound cross-sectional comparisons. Yet another set of factors is entirely beyond our grasp at the current state of knowledge: It appears certain that ASDs are the outcome of many different etiological pathways (Geschwind & State, 2015), but as these differential etiologies are not well understood, they cannot be matched in cross-sectional comparisons of, say, a cohort of 13-year olds and another cohort of 16-year olds. The need for and importance of longitudinal neuroimaging in ASDs is of course generally accepted. Indeed, a growing number of recent studies have presented longitudinal data capturing brain development in the first years of life in ASDs (Wolff, Jacob, & Elison, 2018). Remarkably, longitudinal data for adolescence are almost completely missing in the neuroimaging literature of ASDs. This is unacceptable, both from a scientific perspective and in regard to public health. The common definition of ASDs as a set of neurodevelopmental disorders is adequate when the focus is on causation: In order to understand why a 2-year old may receive the diagnosis of an ASD, we need to better understand in what ways disruptions of early brain development may cause the behavioral problems that result in a diagnosis. However, this focus on early development cannot be exclusive because ASDs are commonly lifespan disorders. Despite progress in earlier diagnosis and intervention in the past decades, most children with ASDs continue to meet diagnostic criteria for the disorder in adulthood (Volkmar & Wolf, 2013). There has been some recently growing awareness that “adult ASD” and aging in people with ASD require more study, given that we have almost no knowledge of what happens in the brains of adults with ASDs as they grow old (e.g., Piven, Rabins et al., 2011). Equally important, however, is the transition from late childhood to young adulthood. This is an important juncture that may affect the development of individuals with ASDs across the later lifespan. It is known that adolescents with ASDs are at greatly increased risk of depression and anxiety disorders (Hofvander et al., 2009; Lugnegard, Hallerback, & Gillberg, 2011). The discrepancy between prevalence rates of depression and anxiety in ASD compared to TD populations increases with age during adolescence (van Steensel & Heeman, 2017), indicating specific vulnerability in adolescents with ASDs. Rates of suicidality are higher in the ASD than in the TD population—a difference that also emerges during late adolescence and the transition to young adulthood (Veenstra-Vander Weele, 2018). Another area of specific concern in adolescents with ASDs is adaptive functioning, which may be a strong predictor of adult outcome (Farley et al., 2009). Deficits in adaptive functioning are linked to executive impairments, which have been in turn found to be associated with depression and anxiety in ASDs (Troyb et al., 2014). There is thus a whole array of aggravating challenges that adolescents with ASDs face as they approach adulthood. The importance of this life stage stands in stark contrast to the very limited understanding of how brain changes that have been observed in TD adolescents (Casey, Jones, & Hare, 2008) may differ in ASDs and may affect the risk of depression, suicidality, and low adaptive functioning. The study by Lawrence et al. is a first step toward filling this large evidential gap. One may consider it a small step with a rather specific set of findings in a small sample of adolescents with ASDs. However, the study may serve to raise awareness for a life stage in ASDs that has been appreciated for its critical importance in clinical and public health spheres, but has been woefully neglected in neuroscientific research.

Autism research : official journal of the International Society for Autism Research, 2018 · doi:10.1002/aur.2042