Related Experiment Video
Updated: Jan 15, 2026
Measuring Engagement of Spectators of Social Digital Games
Published on: July 3, 2021
Measuring Apathy Through Computational Lenses: Behavioral and Physiological Signatures of Engagement in Virtual
Daniel Harlev1, Ramit Ravona- Springer2, Or Koren3
1Department of Physical Therapy (DH, NW), The Stanley Steer School of Health Professions, Gray Faculty of Medical & Health Sciences, Tel Aviv University, Tel Aviv, Israel; Department of Psychiatry (DH), Rambam Health Care Campus, Haifa, Israel; Sagol School of Neuroscience (DH, RRS, SZ, MP, NW), Tel Aviv University, Tel Aviv, Israel.
Objectives:
To develop an objective method for assessing apathy by identifying behavioral clusters based on gaze behavior during an emotionally salient virtual reality (VR) task, and examine associations with apathy severity, cognition, and physiological reactivity.
Design:
Cross-sectional observational study.
Setting:
Memory clinic at a university-affiliated hospital.
Participants:
Eighty-five older adults (mean age 74 ± 5.7 years) with varying cognitive and affective profiles.
Measurements:
Participants completed a naturalistic VR task presenting personalized positive images (e.g., grandchildren), aversive scenes, and neutral stimuli. Eye-tracking data were collected and clustered using unsupervised K-means based on time to first fixation, total fixation duration, and fixation variability. Apathy was assessed using the Lille Apathy Rating Scale (LARS), cognition via the Montreal Cognitive Assessment (MoCA), and depression with the Geriatric Depression Scale (GDS). Physiological reactivity (heart rate variability measured as root mean square of successive differences [RMSSD], galvanic skin response [GSR], respiration) was analyzed using linear mixed-effects models.
Results:
Two gaze-based subgroups were identified: "Engagers" (n = 48) showed faster, more sustained, and less variable gaze behavior; "Nonengagers" (n = 37) exhibited delayed and more variable gaze responses. Nonengagers were older, had lower MoCA scores, and higher apathy levels. RMSSD was lower overall in nonengagers but increased selectively in response to positive stimuli (grandchildren images). No group differences emerged in GSR or respiration.
Conclusions:
Unsupervised clustering of gaze behavior revealed ecologically valid apathy subtypes, linked to cognition and selective RMSSD reactivity. VR-based gaze metrics may offer a scalable, objective tool for detecting motivational impairments in aging.
