Abstract
Are recognition deficits in prosopagnosia characterized by a lack of holistic face processing? Some reports have shown reduced or no holistic processing in prosopagnosics (Ramon & Rossion, 2010, Avidan, 2011, Palermo, 2011) whereas others demonstrate normal holistic processing patterns (Le Grand, 2006; Bukach, 2006). We addressed this debate more thoroughly with a larger sample of developmental prosopagnosics (DPs; N=30) and a comparable number of healthy controls (N=34). Face recognition ability was assessed using the Cambridge Face Memory Test (CFMT; Duchaine & Nakayama, 2006) and holistic processing through the part-whole test (PW; Tanaka & Farah, 1993). The PW not only allows for overall holistic advantage analysis, but also includes separate eyes, nose, and mouth trials, allowing us to measure how each feature is holistically processed. To explore DP's individual differences in holistic processing we contrasted the use of regression, which is more valid in cases where the aim is to statistically remove control task variance (Peter, 1993), against the usual method of subtraction, which creates a measure correlated with both primary and control tasks. Our results show, similar to healthy controls, intact holistic processing of the mouth and nose regions in DPs. However, surprisingly, they show a part over whole advantage for the eyes - a reverse pattern from controls. This indicates that DPs are capable of holistic processing, but only for certain sections of the face. When analyzing individual differences in DPs, we found a significant correlation between CFMT score and holistic processing of the mouth using regression, and a weaker association when using subtraction. Overall our results show that DPs are unable to integrate the eyes into a holistic representation, that their spared holistic processing of the mouth may provide them some small amount of face recognition ability, and that using regression over subtraction, when appropriate, provides a better measure of individual differences.
Meeting abstract presented at VSS 2012