Uncinate fasciculus disruption relates to poor recognition of negative facial emotions in Alzheimer's disease: a cross-sectional diffusion tensor imaging study

Background: Recognising facial emotions involves visual and emotional information processing. Patients with dementia, including dementia of Alzheimer's type (DAT), are known to poorly recognise facial emotions, especially negative facial emotions. In this study, we aimed to assess if DAT patients exhibit poor facial emotional recognition, and to identify a neural basis for how poor facial emotional recognition might occur. Methods: Magnetic resonance imaging and diffusion tensor imaging (DTI) analysis were conducted in 20 DAT patients and 15 cognitive normal (CN) subjects. The uncinate fasciculus (UF), inferior longitudinal fasciculus, and inferior fronto-occipital fasciculus were delineated by deterministic tractography. DTI parameters were calculated for each fibre. Facial emotion recognition was evaluated with the Facial Emotion Selection Test (FEST). The relationships between FEST scores and DTI parameters in each fibre were measured by partial correlation analyses with age, gender, and the Mini-Mental State Examination as covariates. Group-wise comparisons between DAT and CN subjects were performed for each DTI parameter in each fibre. Results: DAT patients showed lower FEST negative emotion scores than CN subjects (P < 0.05). The score of negative emotion subscale was negatively correlated (r = −0.770, P < 0.001) to mean diffusivity of the left UF in DAT patients. There were no relationships between negative emotion subscale and the other fibre tracts. DAT patients showed no differences in the DTI parameters for each fibre compared to CN subjects. Conclusions: DAT-related prefrontal-limbic network dysfunction is associated with poor recognition of unpleasant emotions; consequently, worse facial recognition of negative emotion is observed in DAT patients.