Competing unconscious reference-frames shape conscious motion perception

Visual processing starts with retinotopic encoding: neighboring points in the real world are projected onto neighboring points in the retina. However, perception is usually non-retinotopic. For example, the motion trajectory of a reflector on a bike appears to be circular (non-retinotopic motion), whereas it is cycloidal in Euclidian and retinotopic coordinates. We perceive the circular motion because we subtract the horizontal motion of the bike (reference-frame) from the cycloidal one. It is impossible to perceive the retino-topic cycloidal motion. It seems that reference-frames for non-retinotopic perception can be generated by low-level stimulus-driven motion-based and higher-level internally-driven attention-based reference frames. Here, we studied how the visual system selects a reference-frame by manipulating the salience of stimulus parts that reinforced specific attention-based coordinates in a Ternus-Pikler Display (TPD). When only retinotopic motion is perceived (reflector presented without the bike), it is impossible to perceive the non-retinotopic motion. Likewise, only 2 disks of the TPD are presented at the same location separated by an ISI of 200ms. In the left disk, a dot rotates clockwise and counterclockwise in the right disk (retinotopic motions). It is impossible to see the non-retinotopic rotation, which would occur when focusing on the left and right disk in alternation. This result surprises because our stimulus is within the capacity of attentional tracking. We propose that the retinotopic motion hinders attentional tracking, but it can be recovered by increasing the salience of stimuli thereby providing a stronger signal to attention-based reference-frames. Hence, our results suggest that the competition among multiple reference-frames is resolved by a variety of mutually suppressive, unconscious mechanisms to create the conscious percepts we eventually perceive.