Is interference control in children with specific language impairment similar to that of children with autistic spectrum disorder?

Abstract

Aims

The purpose of the study was to examine resistance to proactive interference, which is strongly associated with working memory (WM) performance and language processing, in children with specific language impairment (SLI), with autism spectrum disorder (ASD), and with typical development (TD).

Methods

Sixty children (eight to ten years; matched in age and nonverbal IQ) participated in the study. Resistance to proactive interference was measured using a verbal conflict paradigm.

Results

Children with SLI and ASD show a deficit in resistance to proactive interference compared to their TD peers, but the source of the problem appears to be different for the two clinical groups. The interference problem exhibited by the children with SLI is related to a more complex deficit involving different cognitive-linguistic functions, whereas the children with ASD show a specific problem in cognitive flexibility.

Implications

The theoretical implications are that poor resistance to interference may be caused by weaknesses in different WM functions, such as a deficit in updating or responses based on familiarity rather than recollection. The clinical implications are that children with SLI and ASD show distinct patterns of performance; therefore they need different types of intervention to strengthen their resistance to proactive interference.

Introduction

Individual variations in inhibition-related functions are associated with different cognitive skills, including working memory (WM; Hasher, Lustig, & Zacks, 1988), reading comprehension (Gernsbacher, 1993), and language processing (Van Dyke, Johns, & Kukona, 2014), as well as with different disorders, such as ADHD (Barkley, 1997, Nigg, 2003), autism spectrum disorder (Adams and Jarrold, 2012, Bishop and Norbury, 2005), and specific language impairment (Marton, Campanelli, Scheuer, & Yoon, 2014; Spaulding, 2010). Although the terms inhibitory control and interference control are often used interchangeably in the clinical literature, most theorists distinguish these two functions from each other (Friedman & Miyake, 2004; Mazuka, Jincho, & Oishi 2009; Wilson & Kipp, 1998).

The present paper is based on Friedman and Miyake (2004) model of inhibitory control, which distinguishes between inhibition of prepotent responses, resistance to distractor interference, and resistance to proactive interference. While prepotent response inhibition is about blocking an automatic response, distractor interference refers to the suppression of irrelevant stimuli in the external environment that compete with the target. The focus in this paper is on resistance to proactive interference, which is the ability to suppress the intrusion of previously relevant but currently irrelevant information.

From a developmental perspective, the different inhibitory components show distinct developmental trajectories. Response inhibition develops typically during the preschool years (Mazuka et al., 2009), whereas interference control develops through adolescence (Bjorklund & Harnishfeger, 1990). Although there are shared underlying mechanisms behind automatic response inhibition, resistance to distractor interference and resistance to proactive interference, these three functions are separate (Dempster & Corkill, 1999; Friedman & Miyake, 2004). Further, it is resistance to proactive interference that plays a critical role in language processing and WM performance (Martin-Rhee & Bialystok, 2008; Van Dyke & Johns, 2012). The relationship between WM and interference control strengthens as children become older. An increase in the association between WM and resistance to interference is related to the development of stronger – more active – WM representations, more conscious differentiation between relevant and irrelevant items, and intentional suppression of irrelevant memory traces (Roncadin, Pascual-Leone, Rich, & Dennis 2007). In order to perform well on a WM or language task, current task goals have to be active and previous memory traces need to be suppressed as relevant and irrelevant items compete for the same limited capacity (Unsworth, Brewer, & Spillers 2013).

In general, children with SLI exhibit typical sensory and intellectual skills but perform below average on different language tasks and on WM measures (Leonard, 2014, Marton and Schwartz, 2003, Montgomery, 2003). Several terms have been used in the literature to classify these children (primary language impairment, developmental language disorder, expressive and receptive language impairment, etc.).¹ We refer to our participants as children with specific language impairment in this paper because this is the most widely used term within the research community. Although there are only a few studies focusing on interference control in children with SLI, there is evidence for a deficit in resisting interference in this population (Pauls & Archibald, 2016).

The first studies that reported problems with resistance to interference in children with SLI did not focus specifically on interference control but on WM capacity. Their error analysis data, however, revealed a weakness in resistance to interference in children with SLI (e.g., Gillam, Cowan, & Day, 1995; Marton, Kelmenson, & Pinkhasova 2007; Weismer, Evans, & Hesketh 1999). These children produced a relatively large number of perseverative errors, which is an indication of poor resistance to proactive interference.

A second line of research involved traditional executive function tasks, such as the Wisconsin Card Sorting test (WCST-64; Kongs, Thompson, Iverson, & Heaton, 2000) and variations of the Stroop test. Although the outcomes are somewhat mixed, the overall results show that children with SLI exhibit problems in tasks that require the selection of relevant information and the suppression of irrelevant stimuli or memory traces. In the Wisconsin Card Sorting test, children with SLI produced more perseverative errors than their peers. These children had difficulty switching from one sorting principle to another as the conditions changed (Marton, 2008). One reason for these perseverative errors may be a deficit in resisting proactive interference.

The results across studies using the Stroop task show inconsistent findings in children with SLI. Most researchers used a version of the color-word task in which children need to name the color of the ink of different color words (e.g., the word blue is printed in red and the correct response is red). To perform well on this task, children have to suppress their automatic response (i.e. reading the word) as they name the color of the ink. In one study, children with SLI exhibited a larger Stroop-effect than their typically developing (TD) peers (Im-Bolter, Johnson, & Pascual-Leone, 2006), whereas in a different study, children with SLI performed similarly to the TD controls (Lukács, Ladányi, Fazekas, & Kemény, 2016). In addition to variations in participants, these two studies also differed in their scientific approach. Im-Bolter and colleagues used the Stroop task to examine inhibition as the interruption of mental attention, whereas Lukács and colleagues viewed inhibition as one of the executive functions within the model of Miyakeet al., (2000). The authors of the two studies also performed different data analyses that might have contributed to the contrasting outcomes.

Response inhibition has been found intact in children with SLI by other researchers as well (Laloi, De Jong, & Baker, 2017; Marton, Campanelli, Scheuer, Yoon, & Eichorn, 2012). As indicated by the inhibition and interference control model (Friedman & Miyake, 2004), these two functions are related to each other but are not the same. Based on previous findings, children with SLI perfom better in tasks that require withholding an automatic response than in tasks where resistance to interference is needed (Laloi et al., 2017, Marton et al., 2014, Spaulding, 2010).

Henry, Messer, and Nash (2012) developed a different type of Stroop task. First, children were asked to repeat the experimenter’s words (e.g., doll, car), then they were instructed to say doll when the experimenter said car and say car when the experimenter said doll. Blocks of copying and inhibiting followed each other. Children with SLI performed more poorly than the TD children and this outcome reflected a broader executive function deficit that involved both verbal and nonverbal domains and various functions (e.g., shifting, response inhibition, and working memory updating).

Findings from direct measures of interference control suggest that children with SLI show difficulty resisting both distractor and proactive interference (Marton et al., 2012, Marton et al., 2014, Spaulding, 2010). In Spaulding’s study, preschool-age children with SLI performed more poorly than their TD peers in resisting distractor interference across modalities and domains (visual, nonlinguistic auditory, and linguistic). School-age children showed similar problems in a visual matching-to-sample test. The task was to find a matching pair of a visual stimulus among similar items. Children with SLI performed more poorly than their age-matched peers; they achieved even lower scores than younger, language-matched children (Marton et al., 2012). Notably, in the delayed condition, in which there were no distractor items but children had to rely on their visual short-term memory, there was no difference between the children with SLI and their TD peers.

Results from studies using variations of the picture-word interference paradigm also indicate differences between children with SLI and their peers (Schwartz, 2010, Victorino and Schwartz, 2015). Children are required to view and name a picture while interfering words or pictures are presented. Distractor items are either phonologically or semantically similar to the target, or are unrelated words. The interference effect is measured as the difference between response times to related (interfering) and unrelated items. Children with SLI show early semantic, early phonological, and late semantic interference (Seiger-Gardner & Brooks, 2008; Seiger‐Gardner & Schwartz, 2008). In a cross-sectional study that involved children with SLI, children with hearing impairment, cochlear implant users, and TD controls, children with SLI performed more poorly than any of the other groups across all experimental conditions (phonological, semantic, unrelated) independent of the time course of distractor item presentation (de Hoog, Langereis, van Weerdenburg, Knoors, & Verhoeven, 2015). Children with SLI had difficulty selecting the relevant items and simultaneously suppressing the irrelevant ones.

In a simple categorization task, children with SLI performed similarly to their age-matched peers in the baseline condition where all distractor items were new (i.e. not seen previously), but showed a deficit in the interference condition where distractor items were stimuli that had been targets in previous trials (Marton et al., 2014). Thus, children with SLI showed poor resistance to proactive interference. Unlike their TD peers, these children seemed to respond based on item familiarity, instead of memory recollection. They showed difficulty resisting memory traces. One possible reason for this finding was that children with SLI did not update their WM efficiently as new trials were presented.

In summary, despite some controversies in the literature, most of the data indicate a deficit in resistance to interference in monolingual children with SLI compared to typically developing controls. There is limited research focusing specifically on resistance to proactive interference in this population, despite the fact that interference control plays an important role in WM and language processes.

Children with ASD demonstrate restricted and repetitive behaviors and interests, and weaknesses in social interaction and communication from early childhood (DSM–V, 2013). This study did not include nonverbal children with autism or children with intellectual disabilities, only high functioning children with ASD (see more details under Participants).

Previous research indicates difficulty in inhibition related functions in children with ASD but the findings are inconsistent (e.g., Adams and Jarrold, 2012, Bishop and Norbury, 2005; Christ, Kester, Bodner, & Miles, 2011; Lopez, Lincoln, Ozonoff, & Lai, 2005). Reasons for the mixed results are inconsistent use of terminology, test impurity, and individual differences. Despite all these limits, there is growing interest in studying interference control in children with ASD. Most of the previous studies, however, focused on response inhibition and/or on resistance to distractor interference but not on resistance to proactive interference.

A widely used interference paradigm in the autism literature is the local-to-global interference task, in which participants observe a large number or letter that is composed of smaller numbers or letters (e.g., Rinehart, Bradshaw, Moss, Brereton, & Tonge, 2000). Findings across researchers are mixed because different studies included different experimental manipulations (e.g., variations in exposure time; congruent, incongruent, and neutral conditions) and participants varied in their intellectual abilities. Overall results, however, indicate that children with ASD perform more poorly than TD peers on global tasks and display local interference while responding globally. Thus, children with ASD exhibit a local intrusion on global processing.

Several studies used some of the traditional interference control tasks, such as the flanker task, the Stroop task, and the Wisconsin Card Sorting task. Results from studies using the flanker task are also mixed. In the traditional flanker task, participants are presented with rows of arrows that point to the right or to the left. Children are instructed to respond to the arrows in a given position by pressing the corresponding right or left button. Most researchers used at least three different conditions: congruent, incongruent, and neutral. While Van Eylen, Boets, Steyaert, Wagemans, and Noens (2015), found no group difference between children with ASD and TD, Adams and Jarrold (2012) reported that children with ASD exhibited a deficit in resistance to interference. The two studies differed in their stimuli and data analyses. Adams and Jarrold introduced several manipulations in size and in distance among the arrows of the flanker task. Their main finding was that unlike TD children, individuals with ASD produced more errors for incongruent than congruent trials even in the easier conditions. The authors concluded that children with ASD exhibit impaired resistance to distractor interference.

Results from a metaanalysis (Geurts, van den Bergh, & Ruzzano, 2014) suggest significant heterogeneity among studies employing the Stroop task. One reason for this finding may be the large number of variants of this task. Changes in stimulus type, presentation modality and domain may all influence the outcomes. Nevertheless, there are numerous studies suggesting that children with ASD perform more poorly than their peers as indicated by a larger Stroop effect (e.g., Andersen, Skogli, Hovik, Egeland, & Øie, 2015; Sanderson & Allen, 2013).

Another widely used task is the WCST, which is a complex measure of interference control. Children and adolescents with ASD produced more perseverative errors than their TD peers in this test (e.g., Barneveld, Swaab, de Sonneville, van Rijn, van Engeland, & Swaab 2013; Van Eylen et al., 2015). Although these findings are consistent across studies, perseverative responses might reflect different underlying problems. One of them may be poor resistance to interference. Furthermore, individuals with ASD also showed difficulty maintaining sets, a problem that may be associated with a weakness in sustained attention.

In summary, despite some inconsistencies in the autism literature on interference control, there is considerable amount of evidence suggesting a deficit in this area in children with ASD. These children produce more perseverative errors and exhibit larger Stroop effect than their peers. Variations in participants and task type contribute to the heterogeneity of outcomes across studies. Taken together the findings from the SLI and the ASD literature, despite the inconsistencies in findings across studies, both clinical groups show weaknesses in resisting interference but the causes behind these problems are unknown. Problems with working memory updating, monitoring and attention control may all contribute to these children’s perseverative behaviors and to their weaknesses in interference control.

The aim was to examine resistance to proactive interference in children with SLI, ASD and TD using a simple verbal categorization task. Based on the findings in the literature, we formulated the following hypotheses:

• Children with SLI and ASD will show larger interference effects than TD children (TD < SLI = ASD) as indicated by their accuracy scores.

• Children with SLI will be slower in responding to interfering items than children with ASD and TD. As task complexity increases, children with SLI typically show slower response times than their peers (e.g., Miller, Kail, Leonard, & Tomblin, 2001).

• Both clinical groups will show a deficit in resisting interference but their pattern of performance will differ. Children with SLI will show more errors and slower RTs than their TD peers, particularly with the interference items. This prediction was based on the outcomes of Marton et al. (2014). In contrast, children with ASD will perform similarly to their TD peers with all new items –both targets and distractors- but will show a significant decrease in accuracy with the interference items (based on findings by Norbury, 2005). An increase in RT will only occur for the interference items in children with ASD.

• Children’s language skills will have a direct effect on their ability to resist interference (based on findings by Norbury, 2005).