It is typically thought that filling-in requires neuronal adaptation. When visual input weakens, we might no longer see boundaries between features. We tested this by measuring visual cortical excitability with rapid invisible frequency tagging (RIFT) in human MEG. [3/6]