Long-term training, transfer, and retention in learning to lipread

Area of Science:

• Speech Perception Research
• Auditory and Visual Neuroscience
• Cognitive Psychology

Background:

• The fuzzy logical model of perception (FLMP) successfully describes bimodal speech perception, integrating auditory and visual cues.
• Previous research has validated the FLMP's ability to explain joint contributions of audible and visible speech.

Purpose of the Study:

• To evaluate the FLMP's predictive accuracy for lipreading confusion matrices and performance across skill levels.
• To compare the FLMP's predictions against a prelabeling integration model (PRLM).
• To investigate the FLMP's capacity to explain improvements in bimodal speech perception due to lipreading training.

Main Methods:

• A long-term lipreading training paradigm was implemented.
• Subjects learned to lipread consonant-vowel syllables, with training including discrimination and identification tasks.
• Testing involved visual, auditory, and bimodal presentation of syllables, words, and sentences at normal and accelerated rates.

Main Results:

• Subjects demonstrated significant improvement in lipreading abilities across all test item types.
• The FLMP provided a more accurate description of confusion matrices compared to the PRLM, both initially and after training.
• The FLMP successfully accounted for performance gains in bimodal speech perception as lipreading and listening skills improved.

Conclusions:

• The FLMP is a robust model for understanding lipreading and bimodal speech perception.
• Long-term training can lead to substantial improvements in lipreading performance.
• The FLMP's predictive power extends to explaining skill acquisition in speech perception.