User Adaptive Text Predictor for Mentally Disabled Huntington's Patients

Area of Science:

• Assistive Technology
• Computational Linguistics
• Neurodegenerative Diseases

Background:

• Huntington's disease (HD) significantly impairs motor control and communication.
• Existing text predictors may not adequately address the specific needs of individuals with HD.
• Efficient text input is crucial for maintaining quality of life and independence.

Purpose of the Study:

• To design and evaluate a specialized text predictor for patients with Huntington's disease.
• To improve text input rate by limiting phrase options.
• To enhance the predictor's ability to handle unknown words using semantic relationships.

Main Methods:

• Development of a specialized text predictor with constrained phrase options.
• Comparative analysis of input rates against a standard general-purpose text predictor.
• Integration of a semantic database (synonyms, hypernyms, hyponyms) to expand vocabulary.
• Utilizing semantic data for predicting unknown words.

Main Results:

• The specialized text predictor significantly improved text input rate compared to general-purpose predictors.
• Initial versions of the specialized predictor presented challenges in user expression.
• Semantic data integration successfully enabled predictions for words outside the initial training data.
• The enhanced predictor demonstrated improved flexibility in word prediction.

Conclusions:

• Specialized text predictors can enhance communication efficiency for individuals with Huntington's disease.
• Balancing input speed with expressive freedom is a key design consideration.
• Leveraging semantic databases is an effective strategy for improving text predictor adaptability.
• Further research can refine these predictors for optimal user experience and communication.