Impact of augmented reality on narrative-related outcomes in children with Autism Spectrum Disorder: a within-subject design

Recent developments in assistive technology have opened innovative avenues for supporting children with neurodevelopmental conditions, particularly those diagnosed with Autism Spectrum Disorder (ASD). A novel study conducted by researchers at the Catholic University of Valencia San Vicente Mártir in Spain has evaluated how augmented reality (AR) technology influences narrative production, comprehension, and socio-cognitive functioning in children with ASD. Published in Frontiers in Psychology, the research utilizes a rigorous within-subject design to compare traditional paper-based storytelling with an interactive, tablet-based AR approach.

The findings indicate that while fundamental language structures remain relatively stable, the integration of AR serves as a powerful cognitive and visual scaffold. Specifically, the technology significantly boosts children’s behavioral engagement, attentional focus, and ability to express complex socio-cognitive concepts such as thoughts, emotions, and intentions.

Background and Context of the Study

Autism Spectrum Disorder is a neurodevelopmental condition characterized by persistent challenges in social communication, reciprocal social interaction, and restricted or repetitive patterns of behavior. Among the various pragmatic competencies affected, narrative discourse is widely recognized as a critical benchmark. Constructing or retelling a story demands the seamless integration of linguistic, cognitive, and social abilities. Individuals must organize sequential events logically, tailor their communication to the listener’s perspective, and maintain global thematic coherence.

For children with ASD, narrative impairments—particularly at the macrostructural level—are well-documented. These difficulties frequently manifest as an inability to maintain global coherence, organize causal sequences, and interpret or verbalize the internal mental states of story protagonists. These limitations are closely tied to challenges in Theory of Mind (ToM), which is the cognitive capacity to attribute mental states, such as beliefs, desires, and emotions, to oneself and others in order to explain and predict behavior.

Traditional narrative assessments often rely on static, two-dimensional picture books. However, children with ASD frequently exhibit a local processing bias, often described within the framework of Weak Central Coherence theory, where they fixate on isolated details rather than integrating elements into a unified whole. To address these hurdles, researchers and clinicians have increasingly turned to digital and multimedia interventions. Augmented reality, which overlays interactive virtual content—such as three-dimensional animations, audio, and visual pop-ups—onto the physical world, offers a promising medium to bridge this gap by highlighting salient cues and guiding attentional focus.

Methodology and Experimental Design

To investigate the practical applications of this technology, the research team recruited a sample of 15 children diagnosed with ASD from clinical and educational settings across Valencia, Spain. The participant group comprised 12 boys and 3 girls, with a mean age of 6 years and 6 months and a mean composite intelligence quotient (IQ) of 90.23, ensuring that participants did not have an intellectual disability. Furthermore, all children possessed language abilities corresponding to Phase 4 of Tager-Flusberg’s developmental classification, meaning they could successfully combine words into clauses, use emerging grammatical markers, and deploy vocabulary to meet everyday communicative needs.

The study employed a carefully controlled, counterbalanced within-subject design. Every participant completed storytelling and retelling tasks under two distinct conditions: a traditional treatment-as-usual (TAU) modality using printed picture books, and an experimental digital modality using an AR application hosted on a tablet.

The stimulus materials consisted of two structurally matched stories detailing everyday events in the life of a fictional child named Ismael ("Ismael in the Countryside" and "Ismael at the Beach"). Both narratives contained identical numbers of illustrations, episodes, and direct references to emotional and intentional states. To eliminate shared prior knowledge, children listened to the story delivered by a specialized researcher and subsequently retold the narrative to an unfamiliar third-party teacher.

Researchers evaluated participant performance across multiple dimensions:

• Microstructural production: Measured via the Systematic Analysis of Language Transcripts (SALT) software, assessing the number of utterances, Mean Length of Utterance (MLU), subordination index, and lexical diversity (number of different words and type-token ratio).
• Macrostructural and socio-cognitive indicators: Evaluated through the recall of main ideas and the frequency of mental state terms (perceptual, emotional, desire, and belief-related vocabulary).
• Comprehension: Assessed using a ten-question rubric testing both literal and inferential comprehension.
• Reading orientation behaviors: Captured using the Children’s Orientation to Book Reading (COB) Rating Scale, which measures dimensions such as overall engagement, persistence, posture, eye gaze, distractibility, verbal communication, and response to adult support.

Key Findings and Statistical Outcomes

Statistical analyses using paired-samples t-tests and Holm-Bonferroni corrections revealed a distinct pattern of effects across the evaluated domains. The most robust and statistically significant advantages of the AR condition emerged in the realms of behavioral engagement and socio-cognitive expression.

Participants demonstrated significantly higher scores on the Children’s Orientation to Book Reading scale during the AR sessions, registering a mean score of 29.00 compared to 23.40 in the traditional paper condition ($t(14) = 3.35, p = 0.005$, with a large effect size of $d = 0.86$). This indicates that the interactive, three-dimensional nature of the technology substantially improved sustained attention, reduced distractibility, and enhanced active participation.

An even more pronounced effect was observed in the deployment of socio-cognitive language markers. Children in the AR condition produced a significantly greater number of mental state terms, averaging 4.27 references compared to 2.13 in the printed book condition ($t(14) = 3.76, p = 0.002$, yielding a large effect size of $d = 0.97$). This finding remained statistically significant even after strict correction for multiple comparisons, suggesting that the dynamic visualization of characters’ expressions and actions served as an effective scaffold for interpreting internal mental states.

Conversely, improvements in microstructural language features—such as utterance productivity, number of different words, and type-token ratio—showed initial significance in uncorrected analyses but did not survive conservative statistical corrections. Structural measures like MLU and the subordination index, along with overall macrostructural content recall, showed no statistically significant differences between the two conditions. Total comprehension scores exhibited a positive trend toward significance with a moderate effect size ($d = 0.47$) favoring the AR format, hinting that extended exposure could potentially yield deeper semantic gains.

Clinical Implications and Educational Applications

The insights gained from this investigation carry substantial weight for speech-language therapists, educational psychologists, and classroom educators working with neurodivergent youth. By demonstrating that AR can reliably foster behavioral engagement and prompt the use of mental state vocabulary, the study highlights the utility of assistive technology as an active cognitive scaffold.

In clinical assessment environments, the elevated engagement fostered by AR can mitigate common barriers such as task avoidance, low motivation, and high distractibility. Consequently, clinicians may obtain a more accurate representation of a child’s underlying communicative capacities. In educational settings, integrating augmented reality into shared reading activities can facilitate meaningful participation among students with ASD, helping them navigate social concepts without requiring radical modifications to standard curricula.

However, the authors emphasize that AR should not be viewed as a standalone solution for comprehensive language development. Because higher-level discourse organization, syntactic complexity, and global narrative coherence remained stable across formats, AR is best deployed as a complementary resource alongside evidence-based interventions like dialogic reading and explicit pragmatic instruction.

Limitations and Future Research Directions

While the study provides promising preliminary evidence, several methodological limitations warrant consideration. The modest sample size of 15 participants restricts the broader generalizability of the findings and limits statistical power. Additionally, researchers noted the possibility of a "novelty effect," wherein heightened engagement might partially stem from the children’s initial fascination with tablet technology rather than the augmented features alone.

Future research directions will need to incorporate larger, more diverse cohorts, longitudinal tracking to measure sustained developmental changes, and non-augmented digital comparison groups (such as standard e-books) to isolate the unique contributions of three-dimensional augmented reality from generic digital interactivity. Furthermore, analyzing individual cognitive and sensory profiles will help determine which specific subsets of children with ASD derive the maximum benefit from technological scaffolding.