Temporal characteristics of neuronal sources for implied motion perception

Abstract

Viewing photographs of objects in motion evokes higher fMRI activation in human MT+ than similar photographs without this implied motion. MT+ is traditionally considered to be involved in motion perception. Therefore, this finding suggests feedback from object-recognition areas to MT+. To investigate this hypothesis, we measured EEG, which provides information on the involvement of different brain areas over time. Pictures of biological agents with and without implied motion, and random-dot patterns (stationary, coherently moving, and flickering) were displayed. The difference potential between responses to pictures with and without implied motion was maximal at 260 to 320 ms after stimulus onset. Source analysis of this difference revealed one bilateral, symmetric dipole pair in the border region of the occipital-temporal-parietal lobe. In contrast, differences between the three types of random-dot patterns occurred from 100 ms onwards, and corresponded to cascades of dipole activation in a larger area, including that activated by implied motion. Latencies of the implied motion activations are consistent with a feedback projection onto MT+ following object recognition in higher-level temporal areas.