Investigation in neural computation and circuitry of human visual motion perception.

摘要

Motion is an important cue in the everyday lives of visual creatures. Motion information facilitates the separation of figure from background, aides in seeing objects that would otherwise be effectively camouflaged and surfaces that would be otherwise imperceptible. The research presented is an investigation of the neural correlates of complex motion stimuli. Experiment 1 is a psychophysical investigation of a complex motion phenomenon, called biological motion. Previous research has shown the resilience of this stimulus under highly degraded conditions, but by creating stimuli that favor the "form system", we measured the reliance of biological motion perception on the "motion system". We challenge form-based biological motion research, and we conclude that motion perception is necessary (but not sufficient) for perceiving biological motion. We conjecture that this insufficiency is due to another mechanism, in addition to those involved in simple motion discriminations. Experiment series 2 is a neuroimaging investigation of biological motion as a function of contrast modulations, which seeks to find the neural correlate of the effect found in Experiment 1. We specifically targeted the human middle temporal complex (hMT+, the motion-sensitive human homologue to monkey MT), a region implicated in motion perception and historically important in neuroscience research. We find the responses in hMT+ to be stimulus-dependent and to be a part of network of brain regions supporting complex motion perception. Experiment series 3 is a neuroimaging investigation of another form of complex motion perception, a phenomenon called motion transparency. When the visual system encounters two overlapping motion vectors, it resolves them by segmenting them into different surfaces (or objects). We attempt to uncover the neural basis of object segmentation defined by motion vectors. We find the hMT to house competing motion vectors with mutual inhibition, including a local competition between motion vectors as well as a global competition between motion-defined surfaces. These results add to the expansive literature on motion processing and depart from a more traditional depiction of the neural underpinnings of motion perception.