Foreword: development is not about studying children: the importance of longitudinal approaches.

03Original abstract

Paradoxically, numerous studies of infants and children are not developmental at all, because they take static snapshots, targeting a specific age group. A truly developmental perspective embraces a developmental way of thinking, regardless of the age of the population studied (Karmiloff-Smith, 1992, 1998). Even research on infants can be nondevelopmental, simply examining performance in, for example, 5-month-olds, whereas some studies of adults are developmental, because they focus on neural change over time in healthy aging even when behavioral output remains unchanged (e.g., Tyler et al., 2009) or on neurodevelopmental disorders as they continue to change in adulthood (Cornish et al., 2008). Indeed, it used to be thought that permutation carriers of the fragile X gene were unimpaired. However, important new research has shown that adult male premutation carriers who have a CGG repeat length of over 100 repeats are at risk of developing a late-onset neurodegenerative condition: fragile X tremor and ataxia syndrome (Cornish et al., 2008; Cornish, Hocking, Moss, & Kogan, 2011). In other words, neural and cognitive development is dynamic across the entire lifespan, and there is no static end state in either healthy adults as they age or in adults with neurodevelopmental disorders.To understand development in both typical and atypical individuals, it is vital to trace developmental trajectories across time (Cornish, Scerif, & Karmiloff-Smith, 2007); to assess progressive change from infancy onward at the neural, cognitive, and behavioral levels; and to pinpoint how parts of the developing system may interact with other parts differently at different times across ontogenesis (Steele, Brown, & Scerif, 2012). A process that is vital, say, at Time 2 may no longer play a role at Time 5. Yet its delay at Time 2 may have been crucial to a healthy developmental trajectory and outcome (Karmiloff-Smith, 1998). Indeed, developmental timing is among the most important of factors that must be taken into account when trying to understand human development, particularly in atypical cases. This is particularly true when examining comorbidities, which turn out to be very common; pure syndromes are actually relatively rare. The problem is that cross-sectional designs tend to assume a static relationship among comorbid conditions, whereas the relationship is likely to fluctuate and change over developmental time: hence, the importance of longitudinal designs.It is also worth noting that even when scores fall “in the normal range” in neurodevelopmental disorders, it cannot be assumed that this entails a normal developmental trajectory, without first examining the cognitive and neural processes underlying the behavior (Johnson, Halit, Grice, & Karmiloff-Smith, 2002; Karmiloff-Smith, 1998; Karmiloff-Smith et al., 2004). Yet researchers of a nativist persuasion will often label the domain in which scores fall in the normal range as intact next to one or more other domains considered to be impaired (Baron-Cohen, 1998; Duchaine, 2000; Landau, Hoffman, & Kurz, 2006; Temple, 1997; van der Lely, 2005). This approach, often inspired by adult neuropsychology, is captured well in the following quotation from Clahsen and Temple (2003, p. 347): “Williams syndrome can be explained in terms of selective deficits to an otherwise normal modular system.” However, neurodevelopmental disorders do not present with a neat juxtaposition of intact and impaired functions; development is far more complex. Indeed, the very notions of “intact–preserved–spared” have a static flavor and seem to imply genetic determinism, as if the brain were hardwired, unchanging and unaffected by ontogenesis and environmental influences. However, studies of humans and other species have clearly indicated that not only do the brain, behavior, and environment change dynamically over time, but gene expression also changes, influenced by other genes, by the environment, and, crucially, by the behavior of the animal in that environment (Bolhuis, Zijlstra, den Boer-Visser, & van der Zee, 2000; Kaffman & Meaney, 2007). As Gottlieb (2007) has repeatedly stressed, epigenesis is probabilistic and only under very broad genetic control, and this clearly holds for atypical development too (Karmiloff-Smith, 2007). Genetic mutations in neurodevelopmental disorders most often affect general processes like cell migration, synaptogenesis, dendritic growth, and pruning (Huttenlocher, 2002; Johnson, 2001; Kandel, Jessell, & Sanes, 2000; Thomas, Knowland, & Karmiloff-Smith, 2011). Mutations rarely, if ever, target a single cognitive domain, even if the outcome over developmental time looks like a single deficit (Karmiloff-Smith, 2006). Note, too, that there is often a slip from relative comparisons (e.g., the neurodevelopmental-disordered group performs worse in Domain A than in Domain B) to absolute conclusions (e.g., the group is impaired in Domain A and intact in Domain B).Many approaches to neurodevelopmental disorders ignore what Piaget (1966) deemed to be essential: the developmental history of the organism. Indeed, a crucial component of Piaget's epistemology stressed the growth of knowledge over ontogenesis, although at the time he did not generalize his theorizing to brain development. However, those of nativist persuasion tend to disregard everything we have since learned about the progressive development of the infant brain: in other words, the fact that several cortical regions are very immature at birth, particularly the temporal, parietal, and frontal regions, and that, in interaction with both endogenous and exogenous constraints, major structural, functional, and biochemical brain changes occur throughout development (Blakemore, 2010; Casey, Giedd, & Thomas, 2000; Crone et al., 2008; de Haan, Humphreys, & Johnson, 2002; Durston et al., 2006; Gou, Choudhury, & Benasich, 2011; Kinsbury & Finlay, 2001; Mills, Coffy-Corins, & Neville, 1997; Minagawa-Kawai et al., 2007). In fact, domain-specific outcomes may not even be possible without the gradual process of development over ontogenetic time (Karmiloff-Smith, 1992, 2009, 2011).Scientists need always to remember that nothing is static in biology or psychology. This is why it is crucial to examine longitudinal change in neurodevelopmental disorders rather than seeking to identify selective deficits in the phenotypic end state.

American journal on intellectual and developmental disabilities, 2012 · doi:10.1352/1944-7558-117.2.87