Eye tracking-based assessment of eye movements in the differential diagnosis of neurodegenerative diseases

ABSTRACT Background: Alzheimer's disease (AD) is the leading cause of dementia worldwide, followed by Lewy body dementia (DLB) and frontotemporal dementia (FTD). Dementia affects approximately 4-5% of individuals over 65 and up to 20-30% of those over 80, posing significant health and social challenges, particularly with increasing life expectancy. Currently, diagnoses of AD, DLB, and FTD rely on clinical criteria and are often confirmed only through post-mortem pathological evaluation. Early diagnosis is crucial, and research into early diagnostic markers is expanding. Recent studies have indicated significant alterations in oculomotor functions and viewing behaviors in AD patients compared to healthy controls, with fewer studies focusing on DLB and FTD. Objective: To analyze differences in oculomotor performance using eye-tracking methodologies as potential early diagnostic and progression markers in patients with AD, DLB, and FTD. Methods: Eye movement data were analyzed from 206 subjects: 80 with AD, 26 DLB, 23 with FTD, and 75 healthy controls. Diagnoses were based on established criteria, with 28 AD patients, 5 DLB patients and 9 FTD patients also showing positive biological markers (PET, CSF, DaT Scan, genetic exams). Neuropsychological assessments were conducted using the Montreal Cognitive Assessment (MoCA). Eye movements were recorded with the EyeLink 1000 Plus, a video-based eye tracker with a binocular sampling rate up to 2000 Hz, under free-viewing conditions. Participants freely explored a sequence of 20 images without prior instructions. Results: Patients with AD and DLB were comparable for demographic and clinical variables (AD= mean age: 71.16 ± 7.06, mean education: 11.83 ± 5.16, mean MoCA score: 14.15 ± 4.47; DLB= mean age: 75.57 ± 8.30, mean education: 8.64 ± 4.46, mean MoCA score: 13.68 ± 4.23; FTD= mean age: 68.57 ± 9.08, mean education 12.17 ± 3.76, mean MoCA: 18.50 ± 3.27; CTRL= mean age: 71.12 ± 7.06, mean education: 11.83 ± 5.16, mean MoCA score: 24.68 ± 2.88). As for the eye tracker measures all patients in the AD, DLB, and FTD groups, compared to the controls, exhibit a free gaze characterized by fewer fixations (AD vs. HC, p < .001; DLB vs. HC, p = .003; FTD vs. HC, p = .006) and lower viewing entropy (AD vs. HC, p = .001; DLB vs. HC, p < .001; FTD vs. HC, p = .004). Furthermore, we have observed a specific pattern of free gaze for the three different conditions. Subjects with AD are characterized by both a lower number of fixations and a shorter duration of those fixations. This is also associated with increased blinking rate. The DLB group is characterized by a reduction in the number of fixations but an increase in fixation time. Additionally, there is a reduction in entropy, evaluated by SGE, which is greater than that in the AD group and similar to the FTD group. This is also associated with a decrease of blink rate compared with AD and FTD groups. In addition, there is a slight increase in gaze predictability, expressed by GTE, both compared to healthy controls and to the AD and FTD groups. FTD patients are characterized by a lower number of fixations but a longer fixation time, like the DLB group and worse than the AD group. Additionally, there is an increase in SGE, in line with the DLB group, but with gaze predictability, evaluated by GTE, similar to that of the controls and the AD group. Conclusions: Eye tracker studies revealed the potential to obtain reliable eye movement data for screening diagnosis of neurocognitive disorder, even in the initial state, and could be useful for differential diagnosis of AD, DLB and FTD.