Do we foveate targets during smooth pursuit?

Saccades have been shown to land at the center of mass of the target (He & Kowler 1991). Here, we examined if eye position is similarly centered on a target during typical smooth pursuit. Pursuit is considered a foveating behavior, even though it is driven by velocity and not eye position (Rashbass 1961), and has been demonstrated in the periphery (Winterson and Steinman, 1978). We recruited 8 untrained participants (age: 25-35, 3 males). Targets were displayed in a scanning laser ophthalmoscope (SLO) and moved in a modified step-ramp paradigm, stepping 6° from the center and moving through a 12° trajectory, at 6°/sec in 8 directions (0-315° in 45° increments). Targets were ring-shaped and either small (0.6°) or large (1.7°). Monocular fixation stability of each participant was measured to determine the 68% bivariate contour ellipse area (BCEA). Eye and target positions were compared directly on each frame during the longest period of continuous pursuit. Across participants, the target center rarely fell within the BCEA for either small (0.24±0.10) or large (0.16±0.12) targets. However, instances where any part of the target fell within the BCEA were substantially more common (small: 0.69±0.13; large: 0.79±0.22). In some cases, pursuit loci were up to 3-4° from BCEA boundary, arguing against a centering strategy. We also examined whether displacement around the center was related to the quality of pursuit gain. There was no relationship between gain and target placement. Furthermore, the non-centering trend persisted even for trials with near-perfect gains (within 1±0.15), arguing against retinal slip. Our results show that eye position is not centered on the target in pursuit. Unlike in Kowler & Blaser (1994), the distribution of eye position around the target cannot be explained by a bias from target center. This outcome raises the question of whether pursuit is indeed a foveating behavior.