Artificial Intelligence for Enhancing Emotion Expression in Augmentative and Alternative Communication Systems: A Preliminary Study Including Children Both With and Without Autism.
Purpose: Artificial intelligence is increasingly being explored in augmentative and alternative communication (AAC) to support personalized and responsive communication. Integrating emotion recognition into AAC systems may help users convey emotion content in real time, enhancing social interaction...
| Publicado en: | American Journal of Speech-Language Pathology Vol. 35; no. 4; pp. 1713 - 1730 |
|---|---|
| Autores principales: | , , , , , |
| Formato: | pictorial research tables/charts Journal Article |
| Publicado: |
American Speech-Language-Hearing Association
Jul2026
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=195279697&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 195279697 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 10580360 5E7 jtl: American Journal of Speech-Language Pathology issn: 10580360 maglogo: N pubinfo: dt: Jul2026 vid: 35 iid: 4 pid: 42 pub: American Speech-Language-Hearing Association place: Rockville, Maryland artinfo: ui: 195279697 195279697 195279697 10.1044/2026_AJSLP-25-00469 195279697 ppf: 1713 ppct: 17 formats: tig: atl: Artificial Intelligence for Enhancing Emotion Expression in Augmentative and Alternative Communication Systems: A Preliminary Study Including Children Both With and Without Autism. aug: au: Pitt, Kevin M. Ousley, Ciara Gibbons, Christopher Martinez, Eli Kirkpatrick, Ciera E. Bubak, Austin affil: Department of Special Education & Communication Disorders, University of Nebraska--Lincoln sug: subj: Autism Spectrum Disorder In Infancy and Childhood Alternative and Augmentative Communication In Infancy and Childhood Artificial Intelligence Emotions Facial Expression Face Perception Funding Source Human Children with Disabilities Psychosocial Factors Child Adolescence Persons with Mental Disorders Psychosocial Factors Scales Happiness Sadness Task Performance and Analysis Consumer Satisfaction Pilot Studies Validation Studies Precision Wilcoxon Signed Rank Test Intelligent Systems Comparative Studies Consumer Attitudes Descriptive Statistics Child Behavior Cues Confidence Male Female Data Analysis Software Child: 6-12 years Adolescent: 13-18 years Male Female ab: Purpose: Artificial intelligence is increasingly being explored in augmentative and alternative communication (AAC) to support personalized and responsive communication. Integrating emotion recognition into AAC systems may help users convey emotion content in real time, enhancing social interaction and communication quality. Method: This preliminary study utilized DeepFace, a facial emotion recognition tool, as a starting point for integrating emotion recognition into AAC systems. Twelve neurotypical children and four children with autism, aged 8-13 years, completed two tasks: an imitated (copy) emotion condition using static images and a semispontaneous emotion-elicitation condition. Similar to existing AAC devices, each participant completed an individualized calibration procedure to help tailor the model to their facial expressions. Number scale ratings of the user experience were also collected. Results: Overall, the system demonstrated promising provisional accuracy, particularly for happiness, sadness, and calm (neutrality), although performance varied across emotions, participants, and task types. User experience ratings were generally positive, suggesting satisfaction and engagement with the system, with variability observed. Conclusions: Findings highlight the value of calibration and personalization in AAC. Future work should expand training data sets to better represent children's facial expressions, incorporate multimodal inputs, and explore the alignment of synthetic speech output with detected emotions. Broader studies including active AAC users, diverse developmental and cultural backgrounds, and more spontaneous real-world interactions are needed. These results support the preliminary utility of DeepFace as a possible starting point for integrating emotion recognition into AAC systems, but further research is warranted. Directions for user-centered development are also discussed. pubtype: Academic Journal doctype: pictorial research tables/charts Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|