| Muriel Boucart , Jean-Francois Dinon, Pascal Despretz, Thomas Desmettre, Katrine Hladiuk, & Aude Oliva. (2008). Visual Neuroscience, 25, 1-7.
| Age-related macular degeneration (AMD) is a major cause of visual impairment in people older than 50 years in
Western countries, affecting essential tasks such as reading and face recognition. Here we investigated the mechanisms
underlying the deficit in recognition of facial expressions in an AMD population with low vision. Pictures of faces
AU4 displaying different emotions with the mouth open or closed were centrally displayed for 300 ms. Participants with
AMD with low acuity (mean 20/200) and normally sighted age-matched controls performed one of two emotion tasks:
detecting whether a face had an expression or not (EXNEX task) or categorizing the facial emotion as happy, angry, or
AU5 neutral (CATEX task). Previous research has shown that healthy observers are mainly using high spatial frequencies in an EXNEX task while performance at a CATEX task was preferentially based on low spatial frequencies. Due to impaired processing of high spatial frequencies in central vision, we expected and observed that AMD participants failed at deciding whether a face was expressive or not but categorized normally the emotion of the face (e.g., happy, angry, neutral). Moreover, we observed that AMD participants mostly identified emotions using the lower part of the face (mouth). Accuracy did not differ between the two tasks for normally sighted observers. The results indicate that AMD participants are able to identify facial emotion but must base their decision mainly on the low spatial frequencies, as they lack the perception of finer details.