1 Foundations of Attention

What attention does

Attention is not a single mental “spotlight.” It can involve selecting a location or object, maintaining alertness, resolving conflict, or allocating limited processing capacity. Attending to something does not guarantee conscious awareness, and unattended information is not always completely unprocessed.

Several related functions describe different aspects of attention. Sustained attention means maintaining focus over time; orienting means shifting focus toward a location or event; and executive control means managing conflict and task goals. These functions overlap but are not identical.

Selection and divided attention

Selective attention prioritizes a target while competing stimuli are present, as when following one speaker at a noisy gathering. Divided attention coordinates processing across more than one task or information stream, as when listening for a route instruction while driving. Performance often declines when tasks overlap, especially when both require demanding decisions or responses.

Attention can be inferred from performance rather than measured directly. Researchers compare conditions that differ in target relevance, distraction, or task overlap. No single measure isolates attention completely: results can also reflect perception, memory, decision-making, motor speed, motivation, or participants’ strategies.

How attention is measured

Researchers use several kinds of behavioral measures to investigate attention:

• Reaction time and accuracy: Faster or more accurate responses to attended targets can indicate prioritization. Both measures matter because faster responding may come at the cost of more errors.

• Dichotic listening: Different speech streams are presented to the two ears. Participants may shadow, or repeat aloud, the designated stream. Researchers assess how well they report the attended message and what they notice from the unattended one. This method helped motivate early filtering theories.

• Visual search: Participants find a target among distractors. Changes in response time and accuracy as the number or type of distractors varies help reveal the demands of selection.

• Response-conflict tasks: In a task such as the Stroop, a distracting feature can conflict with the required response. Added time or errors provide an index of interference from competing information.

• Dual-task and psychological refractory period (PRP) tasks: Participants respond to two stimuli presented close together. The delay to the second response, and any accuracy trade-offs, indicate limits on concurrent processing.

Good studies use clear comparisons and consider both speed and accuracy, since observed performance may reflect processes other than attention alone.

Early selection and attenuation

A central question in theories of selective attention is when selection occurs as information is processed, from initial sensory analysis through identification and response. Theories differ in how much unattended information is processed before selection.

Broadbent’s filter model proposes that incoming inputs are briefly analyzed for basic physical properties, such as which ear a sound reaches. A capacity-limited filter then selects one channel for more complete processing, while other channels are largely blocked. The model explains why people can struggle to report an unattended speech stream. However, a strict filter cannot readily explain findings that some unattended information, such as personally significant words, can affect the listener.

Treisman’s attenuation model proposes that unattended inputs are weakened rather than completely blocked. Their processing is reduced, but sufficiently important or primed information can still reach recognition. This helps explain why a listener might notice their name in an unattended stream while usually being unable to report its full contents.

Late selection and perceptual load

Late-selection theories propose that multiple inputs may receive substantial perceptual or semantic analysis, with selection occurring later—for example, when choosing which information guides a response or reaches awareness. These accounts accommodate effects of unattended meaning, but must explain how extensive processing can occur despite limited capacity.

The early-versus-late question does not have a single universal answer. Selection can vary with task conditions and with the kind of processing being measured.

Perceptual-load theory helps explain why distractor effects differ across tasks. When the relevant task uses little perceptual capacity, spare capacity may process distractors and produce interference. When the target task is perceptually demanding, less capacity remains for irrelevant inputs, often reducing their influence. Thus, distractor processing depends partly on the demands of the current task.

Capacity limits and bottlenecks

Attention is limited, but the limit is not necessarily one fixed pool that applies equally to every task. Capacity theories describe performance as depending on the mental effort available and how it is allocated among concurrent activities. A difficult task generally requires more effort; if two tasks together demand more than can be supplied, performance on one or both tends to suffer.

A bottleneck account emphasizes that some operations—particularly central decisions about which response to make—may have to proceed one at a time. In psychological refractory period studies, shortening the time between two stimuli commonly delays the response to the second, even when the first response is not delayed as much. This pattern supports a central processing bottleneck in at least some dual-task situations. Competing perceptual demands or response preparation may also contribute to interference.

Capacity sharing and bottlenecks need not be competing explanations for every task. Performance depends on tasks’ perceptual, memory, decision, and response requirements. Two familiar or well-practiced activities may interfere less than two novel, demanding activities, but practice does not guarantee that tasks can be performed without cost—especially when they require the same processing stage or response channel.

• Psychological refractory period (PRP) task