How does a good representation of the intricate movements of living things look like? Can we turn all movements of biological origin visible and measurable in spite of their fast speed and spatial complexity? These were the most puzzling questions for E-J. Marey (1830-1904), a French medical doctor, the inventor of, e.g., the first electrocardiograph. In his photographic studies on locomotion he wished to generate a description of human motion, and invented a technique to optimize image resolution both in space and time. The resulting pictorial representation of movement (the so- called chronophotographs) showed movement trajectories of a few shiny buttons that were attached to the head and major limb-points of an otherwise invisible human actor.
How does the brain represent dynamic information about moving objects and other moving creatures? Our psychophysical studies show that the representation suggested by Marey might be used by the brain as well. We employed a psychophysical reverse mapping technique (Kovács & Julesz, Nature, 1994) to study the global interaction pattern of a large number of neurons responsible for low-level visual coding in the brain. The pattern of neural interactions revealed an effective and sparse shape-representation, which is optimal for coding in memory and to form associations because it relies on small cell assemblies that can carry information about large, extended objects. It resembles Marey's solution in terms of reducing redundancies in order to optimize space-time resolution.