【2012-02-08】Prof. Janet Hsiao, "Computational modeling of hemispheric lateralization in face and visual word recognition"

Poster:Post date:2012-02-03

Title: Computational modeling of hemispheric lateralization in face and visual word recognition
Speaker: Janet Hsiao
                Department of Psychology, University of Hong Kong

2012/2/8 3:00pm
Room 104

In this talk, I will present a connectionist (neural network) model of visual processing to illustrate how computational models can account for human behavior, tease apart confounding factors that are difficult to control in human studies, and provide theoretical explanations/predictions for psychological data. The model incorporates several known observations about visual anatomy and neural computation; it also implements a theory of hemispheric asymmetry in perception, which posits a low spatial frequency bias in the right hemisphere (RH) and a high spatial frequency bias in the left hemisphere. Through applying this model to face and visual word recognition, I show that the model is able to: (1) account for the left side bias/RH lateralization in face processing; (2) provide a theoretical account for the lateralization difference between English word and Chinese character recognition; (3) account for lateralization difference between different Chinese character types; and (4) illustrate why holistic processing and RH lateralization in the recognition of visual stimuli do not always go together. These results demonstrate that computational modeling is an essential tool in psychological research.
Last modification time:2012-02-03 PM 5:32

cron web_use_log