The integration of local elements to coherent forms is at the core of understanding visual perception. Accumulating evidence suggests that both early retinotopic and higher occipitotemporal areas contribute to the integration of local elements to global forms. However, the spatiotemporal characteristics of form analysis in the human visual cortex remain largely unknown. The aim of this study was to investigate form analysis at different spatial (global vs. local structure) and temporal (different stimulus presentation rates) scales across stages of visual analysis (from V1 to the lateral occipital complex-LOC) in the human brain. We used closed contours rendered by Gabor elements and manipulated either the global contour structure or the orientation of the local Gabor elements. Our rapid event-related fMRI adaptation studies suggest that contour integration and form processing in early visual areas is transient and limited within the local neighborhood of their cells' receptive field. In contrast, higher visual areas appear to process the perceived global form in a more sustained manner. Finally, we demonstrate that these spatiotemporal properties of form processing in the visual cortex are modulated by attention. Attention to the global form maintains sustained processing in occipitotemporal areas, whereas attention to local elements enhances their integration in early visual areas. These findings provide novel neuroimaging evidence for form analysis at different spatiotemporal scales across human visual areas and validate the use of rapid event-related fMRI adaptation for investigating processing across stages of visual analysis in the human brain.