This thesis encompasses three experiments that sought to clarify the role of evoked gamma-band activity in human EEG. Although gamma-band activity has been attracting the interest of numerous researchers for several years, our knowledge about the basic functions of this signal appears still incomplete. The experiments were also aimed at determinig optimal stimulation for the detection of gamma-band responses in general, and optimal task design for the detection of top-down effects. Specifically, it was investigated how evoked gamma-band activity is influenced by properties of stimulation, whether top-down processes may also influence this signal and under which conditions such interactions can occur, and finally offer an integrative explanation for the various experimental findings on gamma-band activity and its role in information processing. The experiments revealed that early evoked gamma-band responses show a strong dependency on parameters of stimulation and are probably involved in the sensory representation of the stimulus. At the same time, however, this representation is also subject to top-down influences like expectation in a target detection experiment or matches with representations in visual long-term memory. The data also suggested that bottom-up processes result in a phase-locking of evoked gamma-band activity while top-down influences increase the power of gamma-band activity within the same neural assemblies that are activated by stimulation. The results thus suggest that the evoked gamma-band response is an early interface between bottom-up and top-down processes. We could thus confirm and extend previous reports from human and animal electrophysiological studies that demonstrated the relevance of gamma activity for perceptual and cognitive processes.