Executive function in Williams and Down syndromes

Abstract

Williams (WS) and Down (DS) syndromes are characterised by roughly opposing ability profiles. Relative verbal strengths and visuospatial difficulties have been reported in those with WS, while expressive language difficulties have been observed in individuals with DS. Few investigations into the executive function (EF) skills of these groups have examined the effect of verbal/visuospatial task type on performance. Analogous verbal and visuospatial measures were administered to these populations within four EF domains: executive-loaded working memory (ELWM), inhibition, fluency and set-shifting. Performance in both groups was compared to that of typically developing (TD) children using regression techniques controlling for potentially influential cognitive/developmental factors. Individuals with WS showed the expected relative visuospatial difficulties, as indicated by poorer performance than TD individuals, on tests of ELWM and fluency. Individuals with DS displayed the expected relative verbal difficulty in the domain of set-shifting. In addition, each population showed pervasive deficits across modality in one domain; ELWM for individuals with DS, and inhibition for individuals with WS. Individuals with WS and DS showed EF difficulties in comparison to a TD group, but, their executive performance was affected by EF task type (verbal/visuospatial) and EF domain in different ways. While the findings indicated that EF in these populations is characterised by a range of specific strengths and weaknesses, it was also suggested that the relative verbal/visuospatial strengths associated with each population do not consistently manifest across EF domains. Lastly, syndrome specificity was indicated by the differences in groups’ performance patterns.

Introduction

The genetic conditions Williams syndrome (WS) and Down syndrome (DS) are characterised by roughly opposing ability profiles. Individuals with WS display relative verbal strengths alongside impairments on visuospatial tasks (Bellugi et al., 2001, Pani et al., 1999), particularly those involving a constructive element (Hoffman, Landau, & Pagani, 2003). Saccadic abnormalities (Brown et al., 2003, van der Geest et al., 2004), and problems with location encoding (Farran & Jarrold, 2005), as well as the perceptual grouping of elements (Farran, 2005), have all been suggested as contributory factors with regard to these spatial difficulties, while vulnerability of the dorsal stream – a brain region thought to mediate the processing of spatial location and movement (Milner & Goodale, 1995) – has been implicated at the neurological level (e.g. Atkinson et al., 2003, Galaburda and Bellugi, 2000). Individuals with DS show linguistic impairments, particularly in expressive language (e.g. Chapman, 2003), with verbal functioning reported to be below overall cognitive level (e.g. Vicari, Caselli, & Tonucci, 2000). These patterns have been reflected in short-term memory (STM) performance of both populations. Individuals with WS have shown impairments in comparison to matched individuals on visuospatial, but not verbal, measures (Jarrold et al., 1999, Robinson et al., 2003). Although this specific visuospatial STM weakness has not always been found in WS – Sampaio, Sousa, Férnandez, Henriques, & Gonçalves (2008) reported a generalised verbal and visuospatial STM deficit in their sample – other authors who have observed visuospatial STM difficulties have suggested that it may be the recall of spatial – rather than visual – material which may be compromised (e.g. Vicari, Bellucci, & Carlesimo, 2003). Contrastingly, individuals with DS have been reported to display specific difficulties recalling verbal, but not visuospatial, material (e.g. Brock and Jarrold, 2005, Visu-Petra et al., 2007).

Investigations into the performance of both groups on measures of executive function (EF) – a frontally mediated cognitive skill set conceptualised by Anderson (1998) as “those skills necessary for purposeful, goal-directed activity” (p. 319) – have indicated difficulties in this area. For example, individuals with WS and DS have shown poor performance across a number of EF domains in relation to both matched typically developing (TD) individuals and task norms (Lanfranchi et al., 2010, Porter et al., 2007, Rhodes et al., 2010). With regard to WS, Osório et al. (2012) have reported a higher level of association between performance in different EF domains, in comparison with a group of typically developing (TD) individuals, suggesting that any frontal impairment present in individuals with the condition may result in difficulties which manifest across several EF domains.

However, despite these undoubtedly interesting investigations into the nature of frontal task performance in WS and DS, few studies have formally acknowledged verbal/visuospatial task modality as a potential influence on the performance of either population. This is surprising when the respective ability profiles associated with each condition are considered. EF performance has been related to a range of outcomes in both populations, such as adaptive functioning in WS (Rhodes et al., 2010) and theory of mind performance in DS (Zelazo, Burack, Benedetto, & Frye, 1996), as well as suggested as a possible indicator of cognitive decline in the latter population (Adams and Oliver, 2010, Rowe et al., 2006). Factors such as task modality, which may influence performance in populations with uneven ability profiles, thus assume greater importance when the potential theoretical and prognostic significance of EF is considered.

The few studies to examine the effect of verbal/visuospatial modality on EF performance in either population have provided some evidence that verbal and visuospatial EF performance may be separable, to some extent, in both. With regard to WS, Menghini, Addona, Costanzo, & Vicari (2010) found a specific visuospatial set-shifting impairment, and Atkinson et al. (2003) observed a relative difficulty with visuospatial, but not verbal, inhibition. Correspondingly, Lanfranchi, Cornoldi, & Vianello (2004) suggested that ELWM skills in children and adolescents with DS may be marginally more compromised with regard to verbal stimuli.

Although these findings, in suggesting that verbal/non-verbal EF performance may be characterised by a level of separability in each population, are compatible with the literature on the more general respective difficulties associated with each condition, differences in the samples (Atkinson et al., 2003, Lanfranchi et al., 2004) and comparison methods (Atkinson et al., 2003) employed across experimental tasks in some of these studies mean that they should be interpreted with caution. Furthermore, it can be argued that many of the previously used tasks were not sufficiently ‘pure’ in terms of their verbal and visuospatial separability. For instance, the Trail Making Test used by Menghini and colleagues to assess non-verbal set-shifting employed alphabetic and numeric stimuli which could be verbally labelled.

In summary, further work addressing the role of verbal/visuospatial task modality in EF performance in WS and DS is required. This is due not only to the respective dissociations in these domains with which each population has been linked, but also the paucity of existing studies in the area.

The current study involved the administration of verbal and visuospatial EF task pairs which were analogous, in that processing requirements were as broadly similar as possible across modalities. Tasks within four established EF domains – ELWM, inhibition, fluency and set-shifting – were administered to individuals from both clinical populations (WS, DS), to assess potential differences in modality-specific EF performance. These EF domains were selected because they have been identified as key frontal/executive skill areas (Lehto et al., 2003, Miyake et al., 2000). Individual matching of children between the groups was not undertaken, due to the aim of including as many individuals as possible in the study. Instead, regression analyses using dummy-coded group variables, and controlling for the effects of chronological age (CA) and mental age (MA), were used to compare each clinical group with a group of typically developing (TD) children. The tentative hypotheses were based on the literature pertaining to ability profiles in each clinical group: (1) individuals with WS would display relative verbal strengths – i.e. they would not differ from TD children on verbal EF tasks, but they would show weaker performance than TD children on visuospatial EF tasks and (2) individuals with DS would show relative visuospatial strengths – i.e. they would not differ from TD children on visuospatial EF tasks, but they would show weaker performance than TD children on verbal EF tasks.