Reference Frames for Spatial Cognition: Different Brain Areas are Involved in Viewer-, Object-, and Landmark-Centered Judgments About Object Location

Functional magnetic resonance imaging was used to compare the neural correlates of three different types of spatial coding, which are implicated in crucial cognitive functions of our everyday life, such as visuomotor coordination and orientation in topographical space. By manipulating the requested spatial reference during a task of relative distance estimation, we directly compared viewer-centered, object-centered, and landmark-centered spatial coding of the same realistic 3-D information. Common activation was found in bilateral parietal, occipital, and right frontal premotor regions. The retrosplenial and ventromedial occipital-temporal cortex (and parts of the parietal and occipital cortex) were significantly more activated during the landmark-centered condition. The ventrolateral occipital-temporal cortex was particularly involved in object-centered coding. Results strongly demonstrate that viewer-centered (egocentric) coding is restricted to the dorsal stream and connected frontal regions, whereas a coding centered on external references requires both dorsal and ventral regions, depending on the reference being a movable object or a landmark.