Editorial: Capturing Developmental Trajectories of Change in Persons With Intellectual and Developmental Disability

Abstract

In her seminal 1998 synopsis on developmental disorders and the role of development in shaping outcomes, Annette Karmiloff-Smith made the following observation:We present this special issue of the journal with a great deal of excitement and expectation, as we see the fruits of a new generation of research discovery begin to fulfil the vision expounded by Karmiloff-Smith over a decade ago. Within the past few years alone, pioneering longitudinal studies have demonstrated that developmental changes are observable early in life and are disorder specific, such that different disorders have their own “signature” profile, and that changes continue across infancy, childhood, and even into adulthood. In essence, development is not static and rarely reaches an age plateau. These unique trajectories have now been identified across a range of different developmental disorders. In this special issue of the American Journal on Intellectual and Developmental Disabilities, we highlight some of these core new findings and their implications for the development of better targeted resources that will help reduce the clinical burden of disability for many families and increase academic potential of all affected individuals.By their very nature, longitudinal approaches to understanding developmental disorders are time consuming and costly, but the rich information accrued across different timelines and the translational impact of these data are substantial. Traditionally, researchers in this field have tended to focus on using cross-sectional methodology to explore age-related changes across developmental time. This approach does, of course, provide important insights into how different age groups function across different snapshots of time. However, these data can never fully capture the subtle changing profiles that are driven by individual differences as they occur across time in the same individuals. The new information presented in this special edition will contribute toward the essential next step in promoting progress toward developing interventions that recognize developmental change.One of the many complexities inherent in our field is how one measures change. There are several methods available, ranging from standardized measures that include rating scales to cognitive or behavioral measures that tap different aspects of behavior and cognition. These measures are sensitive to different and, in some cases, complementary aspects of change. For example, IQ indexes differences compared with chronological age–matched peers, and these can remain stable (e.g., Williams syndrome; see Mervis, Kistler, & John, this issue) or increase over time (as in the case of fragile X syndrome; see Cornish, Cole, Longhi, Karmiloff-Smith, & Scerif, this issue). At the same time, a measure like growth score, which indexes improvements in the ability to tackle progressively more difficult tasks–items (similar to, although more refined than, raw score) can reveal small but significant progressive improvements over time in children with developmental disorders, such as those shown by children with fragile X syndrome, despite their decreasing absolute IQ scores (Cornish et al., this issue). Other measures do not seem to be as sensitive to change in certain groups, as in the case of the Social Communication Questionnaire (Rutter, Bailey, & Lord, 2003) for children with fragile X syndrome (Cornish et al.), but they are for others (children with high-functioning idiopathic autism; see Pellicano, this issue).A second complexity is how, as researchers, we tease out the precise nature of any developmental changes that may be disorder specific. One of the most popular approaches is to compare the group being studied with a typically developing population of children matched on chronological age and/or developmental level. The latter approach is preferable because it is necessary to show that improvement does occur with increasing mental age level and that performance has not reached an age plateau. By only comparing performance with chronological age–matched, typically developing children, the information provided is limited, producing the obvious conclusion that children with developmental delay will have more behavioral issues and cognitively perform more slowly and less accurately than their age-equivalent typical peers. Moreover, there is a greater likelihood that this approach will produce what, at first glance, looks like “developmental freeze,” when, in fact, change is continuing but paced at a developmental age level not chronological age level. Small sample sizes and convenience sampling that includes more “capable” individuals as participants have also hindered a truly representative picture of disorder-specific developmental changes. Given such limitations, it is imperative that studies consider very carefully their choice of comparison group(s), bearing in mind such complexities as the target task or behavior itself. For example, some tasks need a chronological age–matched comparison group because of physiological changes that are highly sensitive to age-related changes even in typically developing children (as in the case for the heart activity measure in Roberts, Tonnsen, Robinson, & Shinkareva, this issue), whereas other, standardized measures need no explicit comparison group but require charting subtle trajectories of developmental change for individual, atypically developing children through multilevel hierarchical modeling (Mervis et al., this issue). Conversely, some research designs lend themselves to cross-syndrome comparisons to isolate phenotypic signatures of change (Fisch et al., this issue), whereas others focus on a single disorder to isolate developmental change (Gray et al., this issue) Clearly, choice of comparison group is an essential starting point in this field and, yet, is perhaps one of the most complex issues to address. There is no “one size fits all” model that is relevant or appropriate for all research studies. Matching will be study and task dependent. For the interested reader, Thomas et al. (2009) have outlined some key discussion points related to the selection and modeling of comparisons in Williams syndrome, Down syndrome, and autism, and similar considerations apply across developmental disorders.In this special issue, we highlight progress in this emerging field in a series of six articles. All authors use the longitudinal approach in slightly different ways, with a focus on three well-documented developmental disorders: autism, fragile X syndrome, and Williams syndrome. Some authors characterize IQ trajectories and others, distinct cognitive trajectories. Other authors focus on the trajectory of autism symptomatology, whereas another author studies more general behavior. All authors ask to what extent early behaviors can predict later outcomes.In her foreword to this issue, Karmiloff-Smith presents findings that were inspired by her approach to developmental disorders (Karmiloff-Smith, 1998) and reiterates the importance of tracking change over developmental time, rather than assuming a priori the stability of behavioral and cognitive profiles in any developmental disorder.With a focus on fragile X syndrome, Roberts et al. investigate the early emergence and predictive relationship of heart activity as a biomarker of autistic behavior in infants and toddlers. They found that heart activity measured at an average of 20 months predicted severity of autistic behavior at an average of 60 months. Of interest, they report a nonlinear effect indicating that low arousal predicts severity of autism during the first year of life with a shift to high arousal predicting severity of autism by preschool age. These findings would not have been possible without the adoption of a developmental perspective and demonstrate the feasibility of testing very young children with developmental disorders. Too often, research in atypical populations tends to exclude infants and toddlers instead focusing on the childhood years; however, the richness of data accrued from this younger age group has a vital role to play in understanding the start states of developmental change.Cornish et al. investigate trajectories of outcomes related to IQ, autistic spectrum symptomology, and behavioral and cognitive inattention over a 3-year period in children (3–10 years old) with fragile X syndrome. Trajectories across measures did not show an even profile. Behavioral ratings of autistic symptomology and attention deficit/hyperactivity disorder symptoms showed no change over time, and others, namely nonverbal growth scores and cognitive inattention markers, showed small but nonetheless significant improvements over time. These findings demonstrate the importance of looking beyond traditional outcome measures that may focus on overt behaviors or standardized scores that in some populations may mask important yet subtle developmental changes. Of particular intrigue is the issue of IQ versus growth scores, in which the latter measure showed an improvement over time and the former a decline over time. Growth scores, in contrast to gross standardized IQ scores, appear to be a much more sensitive measure of subtle changes in performance that reflect changes from a child's earlier performance rather than rely on comparisons with ever-increasing older, typically developing peers matched on chronological age. The study by Cornish et al. includes one of the largest samples of boys with fragile X syndrome to date and is the first to compare and contrast the rate of developmental changes in both behavioral and cognitive outcomes of attention. Fisch et al. add a further dimension by including cross-syndrome comparisons of intellectual functioning and adaptive behavior in fragile X syndrome, Williams-Beuren syndrome, and Wolf-Hirschhorn syndrome. Their findings highlight how critical it is to compare different disorders on the same measures and then to follow the participants longitudinally to tease apart disorder-specific “signatures.”With a focus on Williams syndrome, Mervis et al. examine intellectual abilities and the stability of IQ standard scores. This is one of largest longitudinal studies of individuals in Williams syndrome and one of the few to incorporate multilevel modeling. Findings suggest stable onset differences and trajectories when assessing IQ between the ages of 4 and 16 years. The data also highlight large individual differences within this group, both in terms of individual functioning (intercepts) and its change over time (slopes). When investigating verbal IQ, in particular, the current data provide some intriguing insights: The early literature on Williams syndrome emphasized, in relative (or even absolute) terms, the surprising strengths in verbal abilities for individuals with Williams syndrome, whereas the current data on both intercepts and slopes are consistent with subtle and developmentally sustained verbal impairments, as reported by Mervis et al. (e.g., see Mervis & John, 2012) as well as Karmiloff-Smith et al. (e.g., see Karmiloff-Smith, Ansari, Cambell, Scerif, & Thomas, 2006) and others since those initial suggestions, because they point to significant and long-lasting verbal delay behind chronological age.With a focus on autism, Pellicano presents an important area of inquiry regarding the change and stability of autistic features in cognitively able young children with autism. The longitudinal design and well-characterized samples are particular strengths of this work. The results clearly demonstrate the importance of looking beyond common assumptions of developmental change (or lack thereof, as is commonly assumed) in children with autism. The study raises an important yet unaddressed issue of those children (albeit a minority) whose functioning changes dynamically with age. Who are the children belonging to this subsample? Are there early cognitive–behavioral precursors that could identify these children even earlier in development? The study by Pellicano also highlights the role of age of onset of intervention as a predictor of outcome, and this, in turn, points to future studies needed in investigating this variable further. In a complementary study, Gray et al. examine outcomes in older individuals with autism over a 15-year period to detect how early emerging symptoms affect outcomes. This is an important area of research, and the study has a number of strengths, including the large sample size, excellent retention rate, range of outcome measures that have both scientific and clinical meaning, as well as its length. The findings have the potential to make an impact on improving diagnostic and treatment efforts. There are very few longitudinal cohort studies of this kind, particularly on such a large group of individuals with autism, and, so, the results should be of value to the autism research community.The studies presented in this special issue embrace a new era of developmental research in persons with developmental and intellectual disabilities. Research of this nature is not without its challenges, but the benefits accrued from mapping subtle yet specific pathways across developmental time and across different disorders provide a previously unrealized opportunity to target resources that recognize unique signatures of change beginning as early as possible in development.