I propose that sleep deprivation and related techniques are valuable tools for investigating visual attention. I present data from three studies examining different aspects of visual attention: visual search, spatial cueing, and attentional blink (AB) during one such protocol (constant routine). The protocol featured 36-40 hours of continuous wakefulness with no time cues, a semirecumbent posture, and dim lighting.

Visual search: Reaction time (RT) was measured in a standard spatial configuration search for 5s among 2s. Perception/decision time (RT x set size intercept) increased by 400 ms with sleep deprivation, but search time per item (slope) did not change significantly.
Spatial cueing: In a standard Posner cueing task, overall RTs showed little effect of sleep deprivation. However, RT on invalid trials increased substantially, indicating sleepy subjects had difficulty disengaging attention.
Attentional blink: Subjects reported digits in an RSVP letter stream. AB was observed with two digits in the stream. Sleepy subjects showed attenuated AB and decreased overall accuracy.

The three tasks do not show a uniform response to sleep deprivation. By comparing the effects of sleep deprivation on different tasks, we can learn about the relationship between different attentional mechanisms, and speculate about their potential neural substrates.