Eye movement in face recognition in elderly with mild cognitive impairment

Abstract

Studies had confirmed the significant relationship between eye movement strategies adopted in face recognition task and underlying cognitive abilities in young adults and healthy elderly. Taking steps forward, the current study aimed to examine the utility of eye movement patterns in face recognition task in reflecting early signs of pathological neurocognitive decline in elderly, i.e. those with mild cognitive impairment (MCI). Meanwhile, with the inclusion of a delay face recognition trial in addition to the immediate recognition trial that had long been used, the current study also attempted to examine what cognitive abilities were reflected by individuals’ eye movement patterns in tasks that had different cognitive demands.

Participants who were aged 60 or above were recruited either from the memory clinic in Queen Mary Hospital and the community. They were classified into one group with established diagnosis of MCI (N = 15) and the other group with normal cognition (N = 19). They were required to perform a face recognition task and various cognitive assessments tapping into their general cognition function, verbal memory, semantic fluency as well working memory capacity. Their individual eye movement patterns during the face recognition trials were captured using the Hidden Markov Model.

Significant group difference between elderly with MCI and those with normal cognition in their adoption of eye movement strategies in face recognition task was revealed. The result was consistent with previous studies that individuals with better cognition function tended to employ more active and organized visual exploration of facial features (i.e. the analytic pattern). Further analyses revealed that the likelihood of adopting a less active eye movement (i.e. the holistic pattern) over the analytic pattern in the delay face recognition trial could differentiate the participants with MCI with accuracy of 82%, sensitivity of 93%, specificity of 79%. The result supported the utility of eye movement analysis as an alternative way of MCI screening, supplementing the existing cognitive screening tools which were heavily confounded by participants’ education level. On the other hand, it was found that the use of eye movement was correlated with ones’ working memory capacity in the immediate face recognition trial but it was correlated with individuals’ general cognitive function, verbal memory and semantic fluency in the delay face recognition trial.