Visual experience and sensitivity to motion sequences, from human movement patterns to animated shapes

Humans effortlessly interpret complex motion sequences and are especially attuned to the kinematics and dynamics of human actions and embedded social cues. Remarkably, as demonstrated by Heider and Simmel animations (H&S, 1944), even very simple interacting geometric shapes evoke rich social narratives. The extent to which sensitivity to interaction sequences relies on high-level inferences versus more bottom-up visual processes remains unclear. Exploring the role of visual experience in the development of this sensitivity can provide valuable insights. In prior work with patients treated for congenital cataracts as a part of Project Prakash, we probed their post-surgery capacity to recognize and interpret motion sequences, revealing a diverse spectrum of abilities. On the one hand, we observed recognition of human point-light displays immediately upon eye-opening following surgery (Ben-Ami, Neuropsychologia, 2022). On the other hand, we found compromised extraction of goal-directed inter-shape interactions from H&S sequences, even years after treatment (Verma, JOV, 2023). This study addressed this observed disparity by examining sensitivity to point-light displays of dyadic physical interactions (e.g., hugging, fighting). Like H&S animations, these displays capture coordinated movements between two interacting individuals. Additionally, they encompass the structural and kinematic information of naturalistic human movement found in human walker point-light displays but missing from H&S animations. Patients were tested before and up to one year after sight-restoring surgery (longitudinal, n=10) or five years post-surgery (cross-sectional, n=8). Employing a free-report design, patients described observed interactions and extracted their valence (‘aggressive’ or ‘friendly’). Cross-sectional findings reveal fully developed recognition and valence extraction, while longitudinal results indicate a gradual onset of these abilities over time after surgery. The distinct dependency of point-light versus H&S interaction recognition on visual experience implies potentially divergent underlying mechanisms. We shall offer hypotheses to account for this intriguing distinction between simple visual agents and those anchored in biological kinematics.