Pupil size as an indicator of cognitive activity in mild Alzheimer's disease

It is well established that pupil activity indexes cognitive processing. For instance, research has consistently demonstrated that the pupil reacts to working memory span task performance. However, little is known about pupil reaction to cognitive processing in Alzheimer's Disease (AD). We thus investigated whether span tasks can modulate pupil size in patients with AD. We invited 24 patients with AD and 24 healthy older adults to perform backward and forward spans, as well as to count aloud in a control condition, while their pupil activity was recorded with eye tracking glasses. In patients with AD, analysis demonstrated larger pupil size during backward spans ( = 2.12, = .39) than during forward spans ( = 1.98, = .36) [ (23) = 3.22, = .004], larger pupil size during forward spans than during counting ( = 1.67, = .33) [ (23) = 4.75, < .001], as well as larger pupil size during backward spans than during counting [ (23) = 10.60, < .001]. In control participants, analysis demonstrated larger pupil size during backward spans ( = 3.36, = .49) than during forward spans ( = 2.85, = .68) [ (23) = 5.82, < .001], larger pupil size during forward spans than during counting ( = 2.09, = .62) [ (23) = 5.42, < .001], as well as larger pupil size during backward spans than during counting [ (23) = 9.70, < .001]. Results also demonstrated a significant interaction effect between groups and conditions [ (2,92) = 16.63, < .001]; in other words, patients with AD have shown fewer variations on the pupil size across the conditions compared to the control participants. The larger pupil size during backward spans, compared with forward spans or counting, can be attributed to the high cognitive load of backward spans. The modulation of pupil size, as observed across backward/forward spans and counting, can possibly be attributed to sympathetic/adrenergic and parasympathetic/cholinergic activities. Our study demonstrates the value of pupillometry as a potential biomarker of cognitive processing in AD.