Optimized and Predictive Phonemic Interfaces for Augmentative and Alternative Communication.

Area of Science:

• Augmentative and Alternative Communication (AAC)
• Human-Computer Interaction
• Speech Pathology

Background:

• Individuals with severe motor speech disorders often rely on AAC devices for communication.
• Interface design significantly impacts the efficiency and usability of AAC systems.
• Computational methods offer potential for optimizing AAC interface performance.

Purpose of the Study:

• To evaluate the impact of computational optimization and prediction on communication interface performance for individuals with severe motor speech disorders.
• To compare the effectiveness of random versus optimized layouts and static versus predictive target sizes.
• To assess user preferences and communication rates with different interface configurations.

Main Methods:

• Developed four communication interfaces varying in layout (random/optimized) and target prediction (static/predictive).
• Evaluated interfaces with 36 unimpaired users over 12 sessions and 6 users with motor impairments.
• Measured communication rates and user preferences using alternate access methods.

Main Results:

• Prediction significantly increased communication rates during the initial learning phase for unimpaired users.
• Interface optimization significantly enhanced communication rates after users were trained on target locations.
• Users with motor impairments could effectively use the interfaces and preferred the predictive option.

Conclusions:

• Both computational optimization and prediction enhance communication rates in AAC interfaces.
• Predictive features are preferred by users with motor impairments, indicating their clinical relevance.
• Further research is needed to implement these optimized and predictive interfaces in clinical practice for individuals with speech disorders.