AbstractBackgroundDeficits in face perception and face processing have been reported in Alzheimer’s disease (AD) (Kawagoe et al., 2017). However, conflicting findings were presented in these functions in amnestic mild cognitive impairment (aMCI) compared to older adults (Lim et al., 2011). This study aimed to examine eye movement patterns during visual scanning of faces with neutral expressions in AD, aMCI, and healthy controls (HC).MethodNinety‐five participants, including 29 patients with AD (mean age 72.5±7.9), 34 patients with aMCI (mean age 70.6±6.7 years), and 32 HC (mean age 68.5±6.2 years) were participated in the study. Eye movements were recorded from the right eye with the EyeLink 1000 Plus video‐based eye tracker. In the face scanning task, neutral face images without emotional expressions produced by artificial intelligence were used. In this task, which consists of a practical trial and 20 trials, each stimulus was shown for 3000 ms. Participants were instructed to examine the neutral face images presented. Number of saccade, fixation counts, mean fixation duration, and first fixation duration were analysed separately on the regions of eyes, mouth, nose, and face.ResultThe mean number of fixation, the number of saccades, and mean fixation duration across the face were significantly reduced in AD patients compared to HC. The mean number of fixation and fixation durations in the eye were significantly reduced in AD compared to HC. No significant difference was observed in all groups in eye movements in the nose and mouth. In addition, aMCI and HC exhibited similar eye movement patterns in the face‐scanning task.ConclusionStudies with healthy individuals have reported typical eye movement patterns focusing primarily on the eyes, but also on the nose and mouth regions. It is stated that the decrease in the gaze towards the face and eye region in AD patients may be related to face processing difficulties. To our knowledge, this is the first study to examine eye movement patterns during neutral face scanning in AD and aMCI.This work was supported by the TUBITAK project number 119S560.
Read full abstract