Startle eyeblink modification was examined as a measure of information processing. College students were presented with tones of 5 and 7 s duration of either high or low pitch, followed by startle-eliciting stimuli at lead intervals of 120, 2.000, 4,500, or 6,000ms. Attention to tones was manipulated by instructing the task group to count the longer tones of either pitch. The no-task group had no instructed task. Startle eyeblink was inhibited at the short lead interval and facilitated at the long lead intervals in both groups. The task group showed greater inhibition and facilitation during attended than during ignored tones, indicating that early and late controlled processing was occurring. In the task group, I he degree of facilitation appeared to reflect the degree of cognitive demands of the task. Startle eyeblink modification may provide a sensitive measure of the nature and timing of stapes of processing in active and passive attentional conditions.