Journal of Speech Pathology & Therapy
Open Access

Autism spectrum disorder; Automatic emotion recognition; Emotion recognition technologies; Children with autism; Interdisciplinary research; Modalities

Introduction

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition characterized by challenges in social communication and interactions, as well as restricted and repetitive behaviors [1]. One crucial aspect of social interactions that is particularly challenging for individuals with autism is emotion recognition. Emotion recognition involves the ability to accurately perceive and interpret the emotions expressed by others through facial expressions, vocal cues, body language, and other nonverbal cues. For individuals with autism, deficits in emotion recognition can lead to difficulties in understanding and responding appropriately to the emotions of others, which in turn can hinder their ability to engage in meaningful social interactions. Emotion recognition plays a pivotal role in social development and communication [2,3]. It allows individuals to understand the emotional states of others, enabling empathy, emotional regulation, and effective social bonding. For individuals with autism, who often experience challenges in understanding and expressing their emotions, the ability to recognize and understand emotions in others becomes even more crucial. By improving their emotion recognition skills, children with autism may experience enhanced social interactions, increased social engagement, and improved communication. The emergence of automatic emotion recognition technologies has opened up new possibilities for supporting children with autism in developing their emotion recognition abilities. Automatic emotion recognition involves the use of algorithms and machine learning techniques to automatically detect and classify emotions from various sources, such as facial expressions, speech, physiological signals, and more [4,5]. These technologies offer a non-intrusive and objective way to assess and monitor emotions, which can be particularly beneficial for children with autism who may have difficulty verbalizing their feelings. The objectives of the review article are to explore the current state of automatic emotion recognition technologies applied to children with autism and to provide insights into their potential impact on therapy and intervention programs. The structure of the review article will consist of several sections:

Methodological approach: The review will adopt a systematic literature review methodology to ensure a comprehensive and unbiased analysis of the existing literature. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) approach will be used to guide the review process. This approach involves setting clear research questions, defining search terms and criteria, conducting searches across relevant databases, and selecting and evaluating studies based on predefined inclusion and exclusion criteria.

Emotion recognition in children with autism: This section will delve into an in-depth analysis of emotion recognition deficits in children with autism. It will review existing literature that highlights the specific challenges faced by children with autism in recognizing and interpreting emotions in others [6]. The review will discuss how these deficits may contribute to social communication difficulties and hinder the development of meaningful social interactions.

Automatic emotion recognition technologies: Here, automatic emotion recognition technologies will be defined and their interdisciplinary nature will be explored. An overview of various sources used for emotion recognition, such as speech, facial expressions, physiological signals, and more, will be provided. The section will also discuss the potential benefits and limitations of using these technologies in the context of autism.

State-of-the-art studies: This section will summarize recent studies that have explored the application of automatic emotion recognition technologies to children with autism [7]. Key findings and methodological approaches used in these studies will be highlighted, providing a comprehensive understanding of the current advancements in the field.

Comparison with related reviews: In this section, the present review will be compared with other literature reviews on emotion recognition, autism therapy, and social robots. This comparison will help identify gaps in the existing research and suggest areas for further exploration.

Challenges and limitations: The review will address the challenges and limitations associated with implementing automatic emotion recognition technologies for children with autism. Ethical considerations and potential biases related to using these technologies in vulnerable populations will also be discussed.

Recommendations and future directions: The section will provide recommendations to enhance the effectiveness and usability of automatic emotion recognition technologies for children with autism. Promising areas for future research and development in the field will be suggested to foster continuous advancements.

Implications for autism therapy and child-robot interactions: The potential benefits and implications of integrating emotion recognition technologies into autism therapy and child-robot interactions will be explored. This section will discuss how these technologies can be utilized to create more personalized and effective interventions for children with autism. The article aims to present a comprehensive overview of the current state of automatic emotion recognition technologies for children with autism [8-10]. By examining the existing literature, addressing challenges and limitations, and providing recommendations, the review seeks to contribute valuable insights for researchers, practitioners, and educators working to support children with autism in their social and emotional development.

Discussion

This systematic literature review has provided valuable insights into the promising advancements in automatic emotion recognition technologies and their potential to significantly impact the lives of children with autism. The findings underscore the importance of continued research and development in this vital field to address the unique challenges faced by children on the autism spectrum and to optimize the effectiveness of intervention strategies. The reviewed studies demonstrate that automatic emotion recognition technologies offer a novel and objective approach to understanding and interpreting emotions in children with autism. These technologies have the potential to bridge the gap in emotion recognition deficits often observed in this population, thereby facilitating improved social communication and emotional regulation skills [11-13]. By accurately recognizing emotions in real-time, these technologies can enhance the overall emotional well-being of children with autism and contribute to their social development. Moreover, the diversity of observation channels and modalities employed in the reviewed studies highlights the interdisciplinary nature of automatic emotion recognition. This interdisciplinary approach has paved the way for innovative research, incorporating multiple sources of data to improve the precision and robustness of emotion recognition systems. As we advance, it is essential to continue exploring and integrating new data sources to further enhance the accuracy and reliability of these technologies, ensuring their applicability across a wide range of contexts and populations [14,15]. However, this review also highlights several challenges and limitations that must be addressed to fully harness the potential of automatic emotion recognition for children with autism. The disturbances in observation channels and potential biases in the data can impact the accuracy and generalizability of the results. To overcome these challenges, future research should focus on developing adaptive and inclusive algorithms that can account for individual differences in emotional expressions and responses. The findings of this review underscore the need for the creation of open datasets specifically focused on children with autism. Such datasets would facilitate the comparison and replication of studies, encouraging collaborative efforts among researchers to build a comprehensive knowledge base in this domain. Additionally, efforts should be made to involve diverse participant groups, taking into account factors such as sex, developmental age, and level of functioning, to ensure that the technology is applicable to a broad spectrum of children with autism. As we look ahead, it is evident that continued research and development in automatic emotion recognition technologies hold significant promise for revolutionizing the field of autism therapy and child-robot interactions. These technologies can play a pivotal role in designing personalized and adaptive interventions that cater to the unique needs of children with autism. Furthermore, the integration of automatic emotion recognition with social robots has the potential to create highly engaging and interactive interventions, empowering children with autism to develop vital social and emotional skills in a supportive and non-threatening environment.

Conclusion

In conclusion, this comprehensive review highlights the transformative potential of automatic emotion recognition technologies in supporting children with autism. By contributing to a deeper understanding of the current state-of-the-art, challenges, and future directions in this field, this review aims to inspire researchers, practitioners, and educators to continue pushing the boundaries of innovation in the quest for effective and inclusive interventions for children with autism. With concerted efforts, we can foster a brighter and more inclusive future for children on the autism spectrum, empowering them to navigate the complexities of emotions and social interactions with newfound confidence and autonomy.

Acknowledgement:

Not applicable.

Conflict of Interest:

Author declares no conflict of interest.

1. Kindlon D, Garrison W (1984) The Boston Naming Test: Norm data and cue utilization in a sample of normal 6-and 7-year-old children. Brain and Language 21:255-259.

Indexed at, Google Scholar , Crossref

2. Lansing AE, Ivnik RJ, Cullum CM, Randolph C (1999) An empirically derived short form of the Boston naming test. Arch Clin Neuropsychol 14:481-487.

Indexed at, Google Scholar, Crossref

3. Guilford AM, Nawojczyk DC (1988) Standardization of the Boston Naming Test at the kindergarten and elementary school levels. Lang Speech Hear Serv Sch 19:395-400.

Indexed at, Google Scholar, Crossref

4. Albert ML, Goodglass H, Helm NA, Rubens AB, Alexander MP (1981) Dysphasia without repetition disturbance. J Commun Disord 92-106.

Google Scholar

5. Albert M, Goodglass H, Helm N, Rubens A,  Alexander M (2013) Clinical aspects of dysphasia. Springer Science & Business Media.

Indexed at, Google Scholar

6. Martielli TM, Blackburn LB (2016) When a funnel becomes a martini glass: Adolescent performance on the Boston Naming Test. Child Neuropsycho 22:381-393.

Indexed at, Google Scholar, Crossref

7. Ali S, Elliott L, Biss RK, Abumeeiz M, Brantuo M,  et al. (2022) The BNT-15 provides an accurate measure of English proficiency in cognitively intact bilinguals - a study in cross-cultural assessment. Appl Neuropsychol Adult 29:351-363.

Indexed at, Google Scholar, Crossref

8. Cohen M, Town P, Buff A (1988) Neurodevelopmental differences in confrontational naming in children. Dev Neuropsychol 4:75-81.

Indexed at, Google Scholar, Crossref

9. Tsang HL, Lee T M (2003) The effect of ageing on confrontational naming ability. Arch Clin Neuropsychol 18: 81-89.

Indexed at, Google Scholar, Crossref

10. Albert MS, Heller HS, Milberg W (1988) Changes in naming ability with age. Psychol Aging 3:173-178.

Indexed at, Google Scholar

11. Riva D, Nichelli F, Devoti M (2000) Developmental aspects of verbal fluency and confrontation naming in children. Brain Lang 71:267-284.

Indexed at, Google Scholar, Crossref

12. Ly KH, Ly AM, Andersson G (2017) A fully automated conversational agent for promoting mental well-being: a pilot RCT using mixed methods. Internet Interv.

Indexed at, Google Scholar, Crossref

13. Magyar-Moe JL (2009) Therapist’s guide to positive psychological interventions. Academic press.

Google Scholar

14. Hamberger MJ, Seidel WT, MacAllister WS, Smith ML (2018) Auditory and visual naming tests for children. Child Neuropsychol 24:903-922.

Indexed at, Google Scholar, Crossref

15. Nicholas LE, Brookshire RH, Maclennan DL, Schumacher JG, Porrazzo S (1989) Revised administration and scoring procedures for the Boston Naming Test and norms for non-brain-damaged adults. Aphasiology 3:569-580.

Indexed at, Google Scholar, Crossref