In our daily environment, visual motion provides a rich source of information, referred to as optic flow. Using psychophysics and functional Magnetic Resonance Imaging (fMRI), we investigated the mechanisms and neuronal substrate involved in visual motion tasks underlying visually guided navigation, such as heading, time-to-passage (TTP) and object trajectory perception.

First (Chapter 3), we studied the effect of directionally constrained noise on heading perception to investigate the contribution of spatial and temporal integration mechanisms. We found that spatial integration mechanisms play a critical role for tolerating high amounts of directionally constrained noise, while temporal integration mechanisms improve the accuracy of heading judgments. Through Ideal Observer Models (IOM), we provided computational evidence that the human visual system is not limited by 2D information available on the retina, but rather it recovers 3D scene information.

Second (Chapter 4), in a psychophysical experiment on TTP perception, observers made judgments on relative arrival times for approaching objects (of constant size), whose trajectories were not on a collision path with the observer. In the absence of local expansion cues, observers mainly rely on the image velocities and the presence of global cues, i.e. optic flow, affects TTP judgments. The results of an fMRI study involving the same TTP task, revealed significant blood-oxygenated-level-dependent (BOLD) signal in motion processing areas such as human middle temporal area and intraparietal sulcus in addition to premotor and motor cortices and somatosensory association areas. The distribution of activation within these areas varied in proportion with the employed motion information cue complexity.

Third (Chapter 5), we used fMRI to study the neuronal substrate for object trajectory perception during self-motion in healthy control subjects and four stroke patients with posterior cortical lesions sparing human middle temporal area. These patients were impaired on lower-level motion tasks but they had normal performance on tasks involving optic flow including object trajectory perception. The cortical activation pattern in these patients was not different from that found in healthy controls, providing confirmatory neural substrate for the hypothesis that the precise computation of lower-level motion tasks is not critical for higher-level visually guided navigation tasks.