Multilevel Modeling of Gaze From Listeners With Hearing Loss Following a Realistic Conversation

Area of Science:

• Audiology
• Human-Computer Interaction
• Psychology

Background:

• Effective communication is crucial for individuals with hearing loss.
• Studying real-world communication abilities in hearing-impaired individuals requires innovative tools.
• Gaze analysis presents a novel method for understanding listener behavior in conversations.

Purpose of the Study:

• To develop and evaluate a method for analyzing eye-gaze patterns in people with hearing loss during conversations.
• To determine when and how much listeners with hearing loss direct their gaze toward a new speaker.

Main Methods:

• Twenty-two older adults with hearing loss participated.
• Participants viewed a prerecorded two-person audiovisual conversation with background babble noise.
• Multilevel logistic regression (MLR) was used to analyze eye-gaze direction relative to conversation events and speaker transitions.

Main Results:

• Conversation events did not predict gaze shifts.
• Listener gaze toward a new speaker during silence transitions was time-dependent, increasing in an s-shaped curve before and after speech onset.
• Individual conversation events accounted for more variance than individual listener differences.

Conclusions:

• MLR modeling of eye-gaze during talker transitions is a viable approach for studying realistic conversational perception in hearing-impaired individuals.
• This methodology provides valuable insights for future research in this area.