Innovative Investigations of Language in Autism Spectrum Disorder

Letitia Naigles, Letitia Naigles

• Editorial: De Gruyter Mouton
• Año: 2016
• ISBN:
  9783110409956
• Edición: 1
• Categoría: Filología
• Categoría: Lingüística

In recent decades, a growing number of children have been diagnosed with autism spectrum disorder (ASD), a condition characterized by, among other features, social interaction deficits and language impairment. Yet the precise nature of the disorder's impact on language development is not well understood, in part because of the language variability among children across the autism spectrum. The contributors to this volume — experts in fields ranging from communication disorders to developmental and clinical psychology to linguistics — use innovative techniques to address two broad questions: Is the variability of language development and use in children with ASD a function of the language, such that some linguistic domains are more vulnerable to ASD than others? Or is the variability a function of the individual, such that some characteristics predispose those with ASD to have varying levels of difficulty with language development and use? Contributors investigate these questions across linguistic levels, from lexical semantics and single-clause syntax, to computationally complex phonology and the syntax-pragmatics interface. Authors address both spoken and written domains within the wider context of language acquisition. This timely and broadly accessible volume will be of interest to a broad range of specialists, including linguists, psychologists, sociologists, behavioral neurologists, and cognitive neuroscientists.

Courtenay Frazier Norbury

1Eye-Tracking as a Window on Language Processing in ASD

Introduction

In our everyday lives, the language that we hear and produce is situated within a rich environmental experience. The conversational context is shaped by the physical and visual landscape, our previous experiences and knowledge about the world, and our understanding of our social partners. In contrast, clinical assessment of language abilities most often occurs in the absence of rich contextual information. It is therefore unsurprising that scores on structured and standardized tests of language may not always reflect an individual’s ability to use language in socially meaningful contexts (Norbury, Nash, Baird, & Bishop, 2004). This may be especially true in the case of autism spectrum disorder (ASD), in which there is variation in core language abilities and qualitative differences in how core language skills are used for the purposes of social communication.

In this chapter, I consider how innovations in eye-tracking technology can elucidate differences in language processing within ASD. I begin by outlining the rich variation that exists in language ability within ASD and then explain how eye-tracking methods work. Many eye-tracking studies of ASD have focused on precursor skills, such as visual and social attention, and I evaluate the evidence that early perturbations of attention may negatively affect language learning. I then provide a selective review of studies that have used eye-tracking methods with typically developing (TD) language users. Application of these techniques in studies of language processing within ASD is just beginning but includes studies of lexical processing, sentence comprehension, language production, and discourse processing. Taken together, these studies have highlighted multiple factors that contribute to individual differences in language development within ASD.

Language Variation Within ASD

Impairments in using language for social purposes is a cardinal feature of ASD (American Psychiatric Association, 2013). Impairments in language content and structure, for example, vocabulary, phonology, and grammar, are considerably more variable (Norbury, 2013a). For example, a significant minority of children with ASD have limited expressive verbal skills (Wodka, Mathy, & Kalb, 2013). In contrast, Loucas et al. (2008) found that almost half of children with ASD with normal nonverbal reasoning skills scored within the normal range on standard tests of language structure and content. The remaining children in this sample had a profile in which language skills were significantly below expectations for age and nonverbal ability; grammar was strikingly impaired and vocabulary and nonsense word repetition were in the low-average range, while articulation skills were generally preserved. Such a profile is reminiscent of children without autism but with specific language impairment (SLI; Kjelgaard & Tager-Flusberg, 2001).

Historically, language impairments have been considered sequelae of core deficits in social interaction and social cognition. However, the identification of distinct language phenotypes within ASD (Tager-Flusberg & Joseph, 2003) has challenged this view and given rise to an influential theory positing that ASD with additional language impairment (ALI) represents a comorbid condition that shares neurobiological risk with SLI (Tager-Flusberg & Joseph, 2003). Similarities between the SLI and ALI groups are often mirrored by similar performance between children with ASD and language scores within the normal range (ALN) and their TD peers (for review, see Norbury, 2013a).

Our understanding of the sources of language variation within ASD is hampered by the limitations of traditional assessments of language competence; they can only ever reflect the outcome of a complex information processing chain and cannot elucidate where in that chain language processing breaks down. It is therefore possible that different clinical populations experience different vulnerabilities with the various skills that underpin language processing. The net result might be similar performance on a given task, but for very different reasons. For instance, Norbury (2005) asked children with ALI, ALN, and SLI and TD peers to complete a picture verification task. In this task, participants heard a simple declarative sentence and 1,000 ms later, a picture appeared on the screen. Participants were required to indicate whether the picture was an appropriate match for the sentence they just heard. In some instances, the sentences included an ambiguous word such as bank, which could be followed by a picture of a riverbank or a picture of a financial institution. Some sentences included a neutral verb (e.g., “John ran by the bank”) in which either picture would be acceptable, and other sentences contained a verb that biased a particular interpretation of the ambiguous word (e.g., “John fished by the bank”). Overall, accuracy rates for the TD and ALN groups were much higher in the biasing condition relative to SLI and ALI groups, as children with SLI and ALI continued to accept either picture, even if the verb in the sentence context was inconsistent with the picture. The similarity in performance between SLI and ALI is theoretically and clinically interesting, but there are numerous reasons why children may not succeed at this task. Children may not understand the verb or understand the multiple meanings of the ambiguous words. Alternatively, they may understand the individual words in the sentence but fail to integrate those words or fail to remember the whole sentence when it is time to make a decision. It could be that understanding of sentences is intact, but problems with sustaining attention result in a focus on the final word rather than the entire sentence. It is therefore possible that the similar performances of the ALI and SLI groups do not reflect comorbidity but arise from different deficits in underlying mechanisms. Similarly, the “typical” performance of children with ALN may be characterized by subtle differences in the time course of language processing, which may contribute to differences in using language in everyday contexts.

Online methods of measuring social or language processing may take us closer to the source of linguistic errors, highlighting when and why language breaks down and whether the time course of language processing differs as a function of diagnostic status. Eye-tracking paradigms allow fine-grained analysis of language processing in real time and have enormous potential to elucidate the factors that influence language variation within the autism spectrum (Norbury, 2013b).

Eye-Tracking as a Window on Language Processing: Studies of Typical and Atypical Language Learners

How Eye-Tracking Studies Work

The world around us is packed with visually complex information from which we must filter out distractions and focus on the relevant aspects of any scene to understand what is happening and how best to pursue our goals. The sequence and duration of gaze is fundamental to gathering information because high-quality visual detail is only available from a limited spatial region surrounding the fovea (Henderson, 2003). Our visual attention requires a combination of top-down and bottom-up processes. For example, our eyes are drawn to movement, social beings, and otherwise visually salient objects (Henderson, 2003). However, classic experiments by Yarbus (1967) demonstrated that our internal goals also significantly influence the order and duration of gaze to different aspects of a visual scene. Yarbus presented adults with a painting to inspect and then asked viewers to report specific details about the picture. If the task was to establish the age of characters depicted in the scene, viewers increased looking time to faces relative to baseline. In contrast, if the goal was to establish how wealthy the characters were, looks to faces decreased while looking time to clothes and possessions increased.

The application of eye-tracking techniques to the study of language processing was pioneered by Cooper (1974) and Tanenhaus, Spivey-Knowlton, Eberhard, and Sedivy (1995) using the visual world paradigm. In this paradigm, viewers are presented with a visual scene, and their eye movements are recorded while they listen to spoken language because looking times to particular objects are influenced by what the viewers hear. For example, in Cooper’s (1974) work, viewers began to gaze at a picture of a snake as the word snake unfolded. Interestingly, if the picture “snake” was not on screen, viewers would gaze at semantically related items, providing a window on the implicit semantic associations made as language is processed. Thus, by recording the eye movements of viewers as they perform language-related tasks, we can increase understanding of what information is available to the viewer, when that information becomes available, and how that information is related to spoken language production and comprehension.

The application of eye-tracking methods to explore language processing in young TD children or children with atypical language acquisition is a new and rapidly emerging field. To date, only a handful of studies that have applied these methods to children with ASD, yet these studies illustrate the potential of such methods to yield new insights into qualitative differences in foundational skills (visual and social attention) and in the timing of language processes that serve comprehension and production. Most eye-tracking studies of language processing in ASD to date have involved older, cognitively able children and have measured aspects of language comprehension. This largely reflects the attention demands of experimental tasks and the technical demands of linking spoken language with visual fixations. As methods for presentation (including unrestricted head movements and mobile eye-tracking) and analysis continue to develop, there is great potential for extending eye-tracking research to a wider variety of participants in ever more naturalistic environments.

What Are the Dependent Variables in an Eye-Tracking Study?

Eye movements are typically divided into fixations, the points at which gaze pauses at a certain position, usually for at least 100 ms, and saccades, rapid movements to a new position. The temporal order of fixations and saccades yields a scanpath, which provides important information about when and for how long the viewer gazed at different aspects of the visual display. Fixation duration to a particular area of interest is typically the first variable researchers include in their analyses. In the context of language and communication, there is much interest in the extent to which children with ASD fixate the eye and mouth regions of the human face, as these are purveyors of social cues important for language. Latency of fixation is also a key variable; for example, how long it takes a participant to look at a picture of CAKE on hearing the phrase “Jane will eat the . . .” (Nation, Marshall, & Altmann, 2003). We may also be interested in first fixations because these can tell us what is of immediate interest to the viewer. We can then investigate how long it takes to move the eyes from the first fixation to another area of interest. Using the pattern and timing of fixations, we can begin to infer a number of cognitive processes, such as cognitive intent, interest and attention, salience, and cognitive load (Aslin, 2007).

Eye-Tracking Studies of Visual and Social Attention in ASD: Links to Language Development

Visual Attention

To make inferences about underlying cognitive processes, we must first establish that individuals with ASD do not differ from peers with respect to basic oculomotor control processes. In general, studies of children and adults with ASD report intact basic oculomotor control (Kelly, Walker, & Norbury, 2013; Luna, Doll, Hegedus, Minshew, & Sweeney, 2007). However, studies of infant siblings at genetic risk for ASD have reported atypicalities in both visual orienting (disengagement from central fixation to peripheral cue) and oculomotor efficiency (fixation to peripheral cue in the absence of competing stimuli; Elsabbagh et al., 2009). Such studies have suggested that reduced attentional control may be an early marker of ASD that has consequences for how young children with ASD gather information about the world around them. It may also adversely affect social-communicative development because the child may not recognize or follow social-communicative cues such as eye gaze (cf. Brenner, Turner, & Müller, 2007). However, longitudinal studies directly linking early oculomotor anomalies to later socio-communicative function are currently lacking. In addition, early perturbations of visual attention are evident in other neurodevelopmental disorders, such as Down syndrome, Fragile X, and attention-deficit/hyperactivity disorder (Cornish, Scerif, & Karmiloff-Smith, 2007).

Measures of volitional oculomotor control are equally vital to understanding the interaction of vision and language. Most studies that use eye-tracking methods to explore language processing strip the visual scene to a bare minimum to avoid visual distraction. This allows a fairly “pure” estimate of how language influences eye movements (and vice versa) but is far removed from the challenges of language processing in everyday contexts. In the real world, we must actively select visual targets from a visually cluttered world and filter out extraneous visual information to focus on the task at hand. An important question, therefore, is to what extent are individuals with ASD able to do the same?

Volitional oculomotor control is typically assessed using an antisaccade task, a well-known measure of top-down attention control. A cue appears peripheral to a central fixation, but in this task participants are required to ignore the cue and fixate the opposite side of the screen. Kelly, Walker, and Norbury (2013) used this task in school-age children with ALI and ALN and comparison groups of TD age-matched peers and peers with SLI. ASD diagnosis was not predictive of task performance; only those children with language impairments showed significant impairments on the antisaccade tasks, as both the ALI and SLI groups had greater difficulty suppressing reflexive shifts of gaze to the cue. A similar pattern of performance was found on a visual search task in which children had to fixate a prespecified target and maintain fixation on that target while ignoring visual distractors. Although the ALI and SLI groups were just as quick as ALN and TD peers to find the target, they made significantly more fixations to distractor items than the groups without language difficulty. These findings echo Takarae, Luna, Minshew, and Sweeney (2008), who reported increased deficits on visual motor tasks in individuals with ASD with early language delay, relative to peers with ASD who met early language milestones.

These findings highlight the importance of parsing heterogeneity within ASD but do not speak to the direction of causation. It has been argued that early deficits in top-down visual attention can be deleterious to language development because such deficits alter the child’s ability to learn from visual social cues and disrupt the establishment of joint attention, with cascading effects on language learning (Brenner et al., 2007). However, for older children and adults, performance on tasks of visual attention control may benefit from the ability to verbalize task rules (e.g., “look to the opposite side”). Longitudinal studies of infant siblings at risk for ASD may clarify the causal relations between early perturbations of visual attention and later language development.

S...

Índice

• Cover
• Title Page
• Copyright
• Dedication
• Contents
• Contributors
• Foreword
• Acknowledgments
• Introduction: Perspectives on Language in ASD
• 1 Eye-Tracking as a Window on Language Processing in ASD
• 2 Sentence Processing in Young Children With ASD
• 3 Looking Through Their Eyes: Tracking Early Language Comprehension in ASD
• 4 Learning Words in a Social World: Impairments Associated With ASD and Fragile X Syndrome
• 5 Parental Input to Children With ASD and Its Influence on Later Language
• 6 The Effect of Computational Complexity on the Acquisition of French by Children With ASD
• 7 Advanced Syntax and Primary Pragmatics in Children With ASD
• 8 Connections Among Complementation Sentences, Executive Functioning, and Theory of Mind in Autism
• 9 Language Acquisition in ASD: Beyond Standardized Language Measures
• 10 Recall, Structure, and Complexity in Story Retellings by Children With ASD
• 11 Language Representation and Language Use in Children With Optimal Outcomes From ASD
• Index
• About the Editor
• Endnotes