Abstract
Objective
Recent models of anorexia nervosa (AN) emphasise the role of reduced emotion recognition ability (ERA) in the development and maintenance of the disorder. However, methodological limitations impede conclusions from prior research. The current study tries to overcome these limitations by examining ERA with an audio-visual measure that focuses strictly on multimodal nonverbal cues and allows to differentiate between ERA for different emotion categories.
Method
Forty women with AN and 40 healthy women completed the Geneva Emotion Recognition Test. This test includes 83 video clips in which 10 actors express 14 different emotions while saying a pseudo-linguistic sentence without semantic meaning. All clips contain multimodal nonverbal cues (i.e., prosody, facial expression, gestures, and posture).
Results
Patients with AN showed poorer ERA than the healthy control group (d = 0.71), particularly regarding emotions of negative valence (d = 0.26). Furthermore, a lower body weight (r = 0.41) and longer illness duration (ρ = −0.32) were associated with poorer ERA in the AN group.
Conclusions
Using an ecologically valid instrument, the findings of the study support illness models emphasising poor ERA in AN. Directly addressing ERA in the treatment of AN with targeted interventions may be promising.