Holistic processing of the mouth but not the eyes in developmental prosopagnosia

Abstract

Because holistic processing is a hallmark of normal face recognition, we ask whether such processing is reduced in developmental prosopagnosia (DP), and, if so, what the sources are of this deficit. Existing literature provides a mixed picture, with face inversion effects showing consistent holistic processing deficits but unable to locate their source and with some composite face studies showing reduced holistic processing and some not. We addressed this issue more thoroughly with a very large sample of DPs (N = 38) performing the part–whole task, a well-accepted measure of holistic processing that allows for the separate evaluation of individual face parts. Contrary to an expected overall reduction in holistic processing, we found an intact holistic advantage for the mouth and a complete absence of a holistic advantage for the eye region. Less severely impaired prosopagnosics showed significantly more holistic processing of the mouth, suggesting that holistic processing can aid them in recognizing faces.

Keywords:

• Developmental prosopagnosia
• Holistic face processing
• Part–whole task

Acknowledgments

We would like to thank all the developmental prosopagnosics and control participants for contributing their time and effort. We would also like to thank Sam Anthony, Anne Grossetete, and Long Ouyang for programming the tasks. Finally, we would like to acknowledge funding support from a Veterans Affairs Career Development Award for J.D. and a grant from the National Institutes of Health 5R01EY013602-07 awarded to K.N. Contributions: J.D. performed data analysis, wrote the manuscript, and contributed funding. S.C. recruited and tested developmental prosopagnosics, performed data analysis, and contributed to manuscript preparation. R.M. recruited and tested control subjects, performed data analysis, and contributed to manuscript preparation. J.W. helped with data analysis and contributed to manuscript preparation. K.N. contributed to manuscript preparation and contributed funding.

Notes

¹ We use the inclusive term developmental prosopagnosia instead of the more specific term congenital prosopagnosia, because we cannot rule out that prosopagnosia was caused by developmental abnormalities. We exclude individuals with acute brain damage from our definition of developmental prosopagnosia.

² However, holistic processing and computing second-order relations may not be independent processes. The enhanced ability to judge the spacing between features in faces could be a direct consequence of these features being efficiently combined into a holistic representation (e.g., Farah et al., 1998; Rossion, 2008). In other words, making feature spacing judgements is easier for faces because holistic processing allows one to consider multiple features at once.

³ We chose –1.7 standard deviations (SDs) for two reasons. First, we recognize the limitations of the CFMT in that some people report symptoms of severe prosopagnosia, but score close to the normal range on the CFMT. Because of this, we wanted to be slightly more inclusive than a –2.0-SD cut-off. Our second reason is that we wanted to examine individual differences in prosopagnosia and avoid restriction of range effects. All group effects remain significant when using a –2.0 cut-off; see supplementary materials.