Early theories of attention tried to explain why we cannot consciously process everything we sense. Donald Broadbent's 1958 filter theory proposed an early-selection bottleneck: information held in a sensory buffer is filtered by a physical-property gate such as pitch or loudness before deeper semantic processing, which is why unattended channels cannot normally be consciously reported. Ann Treisman's attenuation theory modified this by proposing that unattended channels are weakened rather than fully blocked, so that highly meaningful stimuli like one's own name can still break through the attenuator, accounting for the cocktail party effect. Later late-selection theories argued that all incoming information is processed semantically before filtering occurs, so the bottleneck sits after comprehension rather than at an early sensory gate. Together these views refine the idea of an attentional bottleneck, the limited-capacity filter that explains why we cannot consciously attend to all stimuli at once despite massive sensory input.
More recent theories emphasize capacity and competition rather than a single filter. Lavie's perceptual load theory proposes that attention selection is determined by the load of the task: high-load tasks consume all attentional capacity and prevent distraction, while low-load tasks leave spare capacity that allows irrelevant stimuli to be processed. Treisman and Gelade's feature integration theory proposes that individual features such as color, shape, and orientation are processed in parallel across separate maps, but combining them into a unified object requires focused attention, which is why searching for a conjunction of features is slower than searching for a single feature. In Posner's framework, exogenous cues like a sudden flash automatically draw attention to a location, while endogenous cues like an arrow or verbal instruction require conscious effort to deploy attention, with endogenous attention slower but more flexible and exogenous attention faster but more reflexive.
Other influential frameworks describe attention as a regulated resource under emotional and environmental pressure. Eysenck's attentional control theory proposes that anxiety impairs the goal-directed attention system while increasing reliance on the stimulus-driven system, which is why worry and test anxiety fragment study focus. Kaplan's attention restoration theory proposes that voluntary directed attention is mentally fatiguing and that natural environments replenish attentional capacity through soft fascination, so that even brief exposure to nature can restore focus. Global workspace theory models consciousness as a broadcast of information from a central workspace to many specialized brain modules, with attention acting as the gatekeeper selecting what enters that workspace. Attention schema theory goes further, suggesting the brain constructs a simplified internal model of its own attention, similar to a body schema, allowing us to report on and control our own focus and to attribute attention to others.
Laboratory tasks have been designed to probe these mechanisms. The Stroop effect demonstrates the automaticity of reading: naming the ink color of a word that spells a different color produces interference and slowed reaction time, showing the limits of selective attention when an automatic process competes with a controlled one. The Flanker task measures selective attention and inhibitory control by requiring participants to identify a central stimulus while ignoring flanking distractors, with slower reactions on incongruent trials revealing how efficiently irrelevant input is suppressed. The attentional blink is a brief lapse in attention that occurs when two targets appear in rapid succession, around \(200\text{–}500\) milliseconds apart, with the second target often missed because attention is still blinking from processing the first. Inhibition of return biases attention away from previously attended locations, encouraging exploration of new areas of the environment.