Neurologically normal individuals show a leftward spatial bias and tend to collide with objects on the right side more frequently than on the left. The upper visual field is associated with extrapersonal space, and mediated by the ventral stream through parvocellular projections.The lower visual field is associated with peripersonal space, and mediated by the dorsal stream through magnocellular projections. Upper and lower visual field differences have been observed in perceptual asymmetries but results have been mixed. Object-and space-based coordinates also both influence the leftward bias; however their relative contributions are unknown as similar spatial conditions are often collapsed across. More left-side collisions emerged on a route following task in the lower visual field and more right-side collisions were seen in the upper visual field (Thomas, Stuckel, Gutwin, & Elias, 2009). Left-handers made more right-side collisions in the central condition, whereas right-handers showed no bias. Leftward biases on the greyscales task were stronger in the lower visual field; however no distance-based differences were observed (Thomas & Elias, 2010). A stronger spatial bias was found on the greyscales task, whereas a stronger object-based bias was found on the object luminosity task (Thomas & Elias, in press). When individual spatial conditions were examined, the image chosen most often was always located in the lower field. Stimulus type and spatial location interacted to determine which coordinate type contributes more strongly to leftward biases. We also found that the leftward bias on the greyscales task was stronger in the lower visual field during prolonged presentation and in the upper visual field during brief presentation. A global motion task was created to preferentially engage magnocellular projections to the dorsal stream. Isoluminant red/green and blue/yellow colour tasks, which preferentially engage parvocellular projections to the ventral stream, were also created. Leftward biases were seen on the greyscales and motion tasks. On an isoluminant colour task, biases were significantly weakened, suggesting