Abstract
Executive function (EF) is critical to children’s academic and social emotional development. In general, research indicates that children with specific language impairment (SLI) perform worse on EF tasks than typically developing (TD) children. The literature, however, is not consistent as to whether the EF deficits in children with SLI are domain specific to language tasks, or non-domain specific, affecting both verbal and nonverbal tasks. There are differing views regarding the development of EF. From adolescence, EF involves three semi-independent factors, working memory, inhibitory control, and cognitive flexibility (shifting), that are assessed with different types of tasks. For younger children, Lee, Bull, and Ho (2013) reported that EF is best viewed as a two-factor system, and some researchers have suggested that EF is a single factor for young children (Brydges, Reid, Fox, & Anderson, 2012). Garon, Bryson, and Smith (2008) proposed a hierarchical relationship between EF components in which more basic and earlier developing skills, such as attention, underlie more complex components such as working memory, inhibition, and shifting. Garon and colleagues proposed that lower level skills allow children to achieve the more complex EFs. If an EF hierarchy exists, then one would predict that deficits on lower level EF tasks would result in even greater deficits in higher order EF tasks.
This study had two goals:
The authors wanted to compare performance of preschoolers with and without SLI on a battery of EF tasks that measured both verbal and nonverbal abilities on four EF components.
The authors also wanted to use Garon and colleagues’ EF developmental integrative framework to examine the EF differences between preschoolers with and without SLI.
Information gained from this study can influence the way speech-language pathologists evaluate and treat children with SLI.
The Study
Participants
Fifty-two 4- and 5-year-olds participated in the study. All children received a standard score of 75 or above on a nonverbal intelligence measure—the Kaufman Assessment Battery for Children, Second Edition (Kaufman & Kaufman, 2004)—and had no diagnosis of motor, behavioral, cognitive, or neurological disorders. Half of participants were classified as having typical language development, and the other half were classified as having SLI. Children in the typical language group scored 87 or higher on the Structured Photographic Expressive Language Test–Preschool, Second Edition (SPELT-P 2; Dawson et al., 2005). These children also scored within normal limits on the Goldman-Fristoe Test of Articulation–Second Edition (Goldman & Fristoe, 2000) and had no history of speech or language therapy. Children in the SLI group had standard scores below 87 on the SPELT-P 2, and a speech-language pathologist judged their conversational language abilities to be atypical during informal conversational interactions across approximately 10 testing sessions over the course of the study.
EF Tasks
Each participant completed a total of eight EF measures designed to assess verbal and nonverbal performance on each of the four EF components of interest: sustained selective attention, working memory, inhibition, and shifting. For the purposes of this research, tasks were categorized as verbal if they included verbal stimuli and/or required children to produce a verbal response. Tasks were classified as nonverbal if they included primarily visual stimuli and did not require a verbal response from participants.
Sustained selective attention
Verbal and visual sustained selective attention were assessed using two similar computerized continuous performance tasks presented. In both tasks, participants were required to press a button in response to a target and to withhold responding to nontargets. In the verbal sustained selective attention task, the target was a phrase, “It’s a fly,” and nontarget phrases differed in the final word (e.g., “It’s a rock,” “It’s a star”). The visual sustained selective attention task included an image of a pig as the target and images of other animals (e.g., horse, snake) as nontargets.
Working memory
Verbal working memory was assessed using a backward word-span task, which required children to repeat lists of common one-syllable words (e.g., dog, foot, train) after the experimenter in reverse order. The total number of lists correctly repeated served as the verbal working memory score. The nonverbal working memory measure was a backward block-tapping task. This task required children to tap a set of six identical blocks in the reverse order as the experimenter. The nonverbal working memory score was total number of spans correctly repeated.
Inhibition
Two computerized measures that were based on Diamond, Kirkham, and Amso’s (2002) day–night task were used to assess children’s inhibition. Each task required the child to respond to day or night stimuli by pressing either a button labeled with an illustrated sun image or a button labeled with an illustrated moon image. Participants were required to inhibit prepotent responding of sun for day and moon for night and instead provide the opposite response pairings. In the verbal day–night task, children pressed the buttons in response to auditory stimuli of a woman saying day or night. In the nonverbal task, children pressed a button in response to visual stimuli: a photograph of the daytime sky or a photograph of the nighttime sky.
Shifting
Two tasks that required switching attention between different stimulus dimensions were used to measure children’s verbal and nonverbal shifting abilities. The verbal shifting task used a computerized format that required children to switch between labeling different properties of stimulus images. The stimuli were illustrated images of a girl with a cat and a boy with a dog. Children began the task with the preswitch phase, or “pet game,” during which each stimulus picture was preceded by a visual and auditory cue (collar image and bell sound), and the correct response to each image was labeling the animal (e.g., cat or dog). After completing two practice trials that included experimenter feedback and six test trials, the task switched to the postswitch phase, or “kid game,” in which the correct response was to label the child (e.g., boy or girl). Postswitch stimuli were preceded by a different visual and auditory cue (t-shirt and gong sound). The advanced task phase required children to switch randomly between the pet game and the kid game using the two cue types to guide their responding. Experimenters recorded children’s first response to each item, and overall accuracy for each task phase.
Nonverbal shifting was tested using the dimensional change card sort (DCCS; Zelazo, Frye, & Rapus, 1996). Two sorting bins were placed in front of the participant, each labeled with a target card (e.g., blue rabbit and red car). In the preswitch phase of the task, children sorted the test cards on the basis of one dimension (color or shape). After children sorted six cards, the experimenter informed them that the game changed, and in this postswitch phase, children sorted the remaining six cards on the basis of the opposite dimension (shape or color).
Results
Hypothesis 1: Children with SLI would perform worse than TD children on all EF tasks.
On each EF component—sustained selective attention, working memory, inhibition, and shifting—the typical language group outperformed the SLI group in either or both verbal and nonverbal domains.
SLI group performed equally as well as the typical group on the nonverbal inhibition measure and verbal shifting measure.
Hypothesis 2: Group differences would be larger on verbal measures.
There was not support for this hypothesis. Group differences were similar on both verbal and nonverbal EF tasks.
These results indicate that children with SLI have EF deficits at all levels and the EF deficits are domain-general because they are not limited to verbal tasks.
Hypothesis 3: On the basis of the hierarchical development of EF described by Garon et al. (2008), the magnitude of group differences would increase from lower to higher level skills within the hierarchy.
The largest group difference was found for the measure of sustained selective attention, which was in fact the lowest level EF component in the model.
The other three EF components followed the expected pattern with the smallest group difference between the working memory measure, a greater difference on the inhibition task, and the largest difference on the shifting measure.
The authors had predicted that the EF differences would be smallest on the attention tasks, when in fact they were greatest for this task. They suggest that their attention tasks were more complex than they had intended. The continuous performance tasks used in the present study required two aspects of attention: both selective attention to a specific stimulus and sustained attention across a period of time. Such complex tasks may have required use of a higher level EF such as working memory and inhibition (Fisher & Kloos, 2016), and therefore, the continuous performance tasks required the EF components tapped by the other measures.
Further Questions
The authors address the need for further research in the following areas:
The direction of the relationship between language and EF: There is some evidence from TD children for there being a bidirectional association between EF and language, with EF skills improving language abilities and language skills mediating EF. This relationship has been less explored in the SLI population in which both language and EF are affected. The relationship may be bidirectional or it may be the case that impaired language is leading to deficits in EF. For example, children with SLI may be less efficient in using language to mediate their responding on EF tasks (Eichorn, Marton, Campanelli, & Scheuer, 2014), or children’s language impairment may be caused, or exacerbated, by their underlying EF difficulties. For example, poorer attention skills may make it more difficult for children with SLI to recognize underlying grammatical rules in language input, or working memory deficits may disrupt word learning.
Possible interventions to improve EF abilities in children with SLI: Among TD children with typically developing language, a number of methods have been designed to improve executive functioning (e.g., Diamond & Lee, 2011; also see the Resource Review in this issue of Word of Mouth). If EF deficits contribute to language impairment, then it follows that improving EF may in turn benefit children’s language abilities. Even if EF does not directly influence language development, the fact that children with SLI have broad-based EF deficits, and that EF abilities are critical in academic and social-emotional development, it makes sense to seek ways to improve the EF abilities of children with language impairments.