Eye movements abnormalities during free-viewing task in frontal lobe syndrome: a possible biomarker for frontotemporal lobar degeneration?

Introduction: Frontotemporal lobar degeneration (FTLD) encompasses a spectrum of disorders characterized by cognitive and behavioral disturbances, often confounded with late-onset primary psychiatric disorders (PPD). Reliable biomarkers are essential for diagnostic precision. Eye movement analysis has emerged as a promising physiological marker of brain function. (Behler et al. 2021) (Russell et al. 2021). Aim: To investigate eye movement dynamics during a free-viewing task in FTLD and PPD compared to healthy controls (HC) and assess correlations with cognitive and behavioral measures. Materials and Methods: A total of 129 participants were recruited: 39 FTLD (19 bvFTD, 4 PSP, 4 CBD, 4 PPA), 18 PPD, and 72 HC. Eye movements were recorded using the Eyelink 1000 Plus system during a free-viewing task involving 20 neutral images displayed for 10 seconds each. Diagnostic classification followed established criteria, and participants underwent clinical and cognitive assessments (MoCA). Results: Nine patients were discarded for poor data quality. The final sample consisted of 120 participants: 31 FLTD (19 bvFTD, 4 PSP, 4 CBD, 4 PPA), 17 PPD, 72 HC. Mean age in FTLD population was 69 (50-83) while in PPD was 65 (49-82). Mean MoCA score was 19 for FTLD (11-25), 22 for PPD (13-29). FTLD patients exhibited fewer, longer fixations and fewer saccades compared to controls, with the bvFTD subgroup showing, in addition, an increased blink per trial. PPD patients mirrored the spatial exploration alterations of FTLD (reduced fixation and saccade number, increased fixation duration) and demonstrated a higher total blink number and larger saccade amplitudes. A longer duration of saccades differentiated PPD from bvFTD. The more apathetic bvFTD patients, identified through the FBI questionnaire, displayed patterns similar to PPD, except for lower saccade velocity in the latter. Disinhibited patients showed no significant differences from PPD. In FTLD, cognitive decline correlated with oculomotor metrics: worsening delayed recall was associated with fewer saccades and fixations (p<0.05) and longer fixation durations (p=0.057), while a decline in FAB performance corresponded to reduced blink rates (p=0.05). In PPD, language impairments correlated with an increased main sequence, while poorer recall and naming performance were linked to reduced blink rates per trial (p<0.05 and p=0.052, respectively). An increase in the main sequence was observed with worsening apathy (FBI A) and dementia severity (CDR plus-SB) in FTLD, while in PPD, depressive symptoms (MADRS) correlated with greater saccade amplitude. A subgroup of FTLD patients classified as “static viewers” by PCA (N=14) exhibited higher apathy levels, greater dementia severity, and worse cognitive performance (language and recall) than FTLD “dynamic viewers”. In contrast, PPD "static viewers" showed no significant differences from PPD “dynamic viewers”. Linear regression identified a model based on saccade duration that could distinguish bvFTD from PPD, though this finding requires validation in larger samples. Conclusions: Eye movement analysis during free-viewing tasks revealed distinctive oculomotor signatures in bvFTD, with static viewing patterns reflecting disease severity and apathy. Despite limited differentiation between bvFTD and PPD, saccadic metrics hold promise as diagnostic biomarkers. Free-viewing tasks offer a practical, non-invasive approach to capture disease-relevant information in FTLD, even in patients with executive and attentional deficits.