Use of Multi-Modal Feedback to Facilitate Word and Syllable Combinations

Friday, May 18, 2012
Sheraton Hall (Sheraton Centre Toronto)
9:00 AM
L. DeThorne1, K. Karahalios2, J. Halle3, K. Lyons1 and M. Aparicio Betancourt4, (1)Speech & Hearing Science, University of Illinois, Champaign, IL, (2)University of Illinois at Urbana-Champaign, Urbana, IL, United States, (3)Special Education, University of Illinois, Champaign, IL, (4)Neuroscience Program, University of Illinois, Champaign, IL

Although communication difficulties represent a core feature of autism, literature documenting the effectiveness of speech therapy in this population is relatively sparse.  Given documented challenges in auditory processing, imitation, and motor planning, the use of multimodal feedback (visual, tactile, and auditory) in speech therapy holds particular promise.


The purpose of this interdisciplinary work, currently funded by Autism Speaks, is to examine the use of multimodal feedback to facilitate multisyllabic speech production in children with speech-language impairments, including autism. The specific research questions are as follows:

    Does an integrated speech therapy approach involving multimodal feedback improve children’s production of multisyllabic targets?
    Does use of a computer-based voice visualization system, VocSyl, facilitate multisyllabic productions to a greater extent than traditional methods?


Eighteen children with a variety of developmental disabilities (including autism), age 3-8 years, have been recruited to this work in progress, with completed treatment data available for 10 children at the time of submission. An initial assessment battery indicated that all enrolled participants were at the single-word stage of development.

 Participants were systematically assigned six at a time to the following conditions: (a) speech intervention with traditional multimodal feedback, (b) speech intervention with Vocsyl, a computer-based voice visualization system, and (c) playgroup intervention focused on facilitating social interaction, serving as a control condition. Thirty multisyllabic targets were selected for all participants based on initial assessment and divided into two lists of 15 targets (treated v. untreated). Target lists were counterbalanced for semantic category, phonological complexity, and number of syllables. Neither list is treated in the control playgroup condition.

The two speech interventions consist of both an explicit motor practice segment and a period of developmental play. Motor practice is completed with either the pacing board or Vocsyl depending on the condition. The children assigned to the control group receive a comparable number of playgroup sessions without focus on any of the 30 multisyllabic targets.

All participants are systematically assessed on their 30 selected multisyllabic targets at five time points throughout the course of intervention using both a card-labeling and an object-play task.


Analyses focus on group differences across conditions as well as within group differences on treated versus untreated targets. The four children who have completed the control playgroup condition to date increased their production of targets to 20% and 15% in the card-labeling and object-play tasks respectively, thereby providing a comparison for our explicit speech interventions. In contrast, the six children in the traditional speech treatment group increased their accuracy on treatment targets by 40% and 39% respectively. Additionally, the children in the traditional speech treatment condition averaged 40% accuracy on treated targets compared to 16% on the untreated targets.


Preliminary results based on six children in the traditional treatment condition and four in the control group suggested a modest but positive impact for the traditional speech intervention. Data collection should be complete by May, allowing us to present complete group comparisons, including a direct comparison of the traditional and computerized treatment conditions.

| More