Learn more about our exciting upcoming events!
ELSC-ICNC Seminar: Isabel Gauthier
ELSC-ICNC cordially invite you to the seminar given by:
Professor of Psychology Vanderbilt University, Nashville
On the topic of:
Holistic processing can help us understand face perception and the origins of its specialization
Thursday, June 02, 2011 at 17:00
at the ELSC-ICNC lecture hall
(Silverman Bldg., Wing 3, 6th floor - Edmond J. Safra Campus)
The concept of holistic processing (HP) is a cornerstone of face recognition research, and an important construct in efforts to link neural and performance measures in the cognitive neuroscience of perceptual expertise. Yet central questions having to do with HP remain unanswered and debates fail to reach a resolution despite accumulating empirical evidence. Is HP specific to faces? Is HP caused by expertise? Is HP of faces even related to performance in face recognition? Are inverted faces processed holistically? I will argue that a considerable source of confusion in this literature stems from a methodological problem. One of the most popular measures of HP comes from the composite task (Young et al., 1987). Critically, there are two versions of the composite task to measure HP for unfamiliar faces currently in use in the literature and they often provide incompatible results, such that one, or both, method(s) must be rejected. One approach (which I call partial design) confounds manipulations of interest with complex response biases that are influenced both by stimulus properties and by top-down strategies. The alternative approach to measuring HP (which I call complete design) is robust to these influences and captures important differences in the processing of objects and faces, both behaviorally and in neural measures. I will make the strong claim, based on several experiments, that the partial design measure of HP should be abandoned. Using the complete design, we found that HP of faces is related to face recognition performance, which converges with recent work where HP of objects is related to expertise for objects. Just as abandoning phrenology did not lead us to reject the value of cortical specialization of functions, this appears to be a case where abandoning a flawed measure can save a useful construct and promote empirical convergence and theoretical resolution.