Klinische Neurophysiologie 2004; 35 - 271
DOI: 10.1055/s-2004-832183

Constructive Brain Processes in Motion Perception

P Sterzer 1
  • 1Frankfurt

Bistable perceptions allow the dissociation of perception- from stimulus-related neural processes. We investigated the neural mechanisms underlying bistable apparent motion (AM) perception using functional magnetic resonance imaging (fMRI). In a first set of experiments, we asked how perceptual reversals during bistable AM are coded in extrastriate visual areas. We assessed event-related activations associated with perceptual reversals during continuous presentation of two different ambiguous AM stimuli, the 'spinning wheel illusion' and the 'dynamic dot quartet' (see figure). Contrasting with previous studies that used bistable paradigms involving objects or faces, we detected no reversal-related activations in ventral extrastriate regions, but instead in the human motion complex (hMT+/V5) with both ambiguous AM stimuli used. This suggests an attribute-specific coding of perceptual reversals in extrastriate visual cortex. In the spinning wheel illusion but not in the dynamic dot quartet, we observed additional reversal-related activations in the „kinetic occipital“ area, which has been implicated in the processing of kinetic contours. Since the spinning wheel stimulus, in contrast to the dynamic dot quartet, evokes the impression of a moving object contour in front of a stationary background, the different response patterns indicate a fine-tuning of reversal-related activations to stimulus features. In a second set of experiments, we were interested in the neural correlates of 'correspondence', which refers to the question of how the visual system determines which object in one frame corresponds to which object in the next in ambiguous AM displays. To investigate the influence of intrinsic object properties on the path of AM, we used dynamic dot quartets with differences in color or luminance serving as correspondence cues. We observed a suppression of neural activity in the calcarine cortex (V1) whenever perceived AM paths violated color and luminance cues. In contrast, when AM paths matched the correspondence cues, V1 activity was the same as during perception of uncued displays. This finding might reflect regulatory mechanisms that flexibly gate early visual feature processing in accord with an overruling perceptual decision. Taken together, our findings indicate that reversal-related activation patterns in bistable visual perception are not only attribute-specific but depend on stimulus characteristics in a finely tuned way suitable to encode perceptual content in the absence of sensory input changes. Moreover, perceptual decisions seem to be associated with adjustments of activity levels at early visual processing stages.