Why This Matters
You don't perceive everything in your environment. Your brain selects and processes only a fraction of the sensory input hitting you at any moment. Theories of attention explain the mechanisms behind this selection: selective attention, divided attention, visual search, and resource allocation.
These theories appear frequently on exams, and you're expected to distinguish between early vs. late selection models, understand capacity limitations, and explain how attention binds features into coherent perceptions. Don't just memorize theory names. Know what problem each one solves and where it fits in the broader debate about when and how filtering occurs.
The real payoff comes from understanding how these theories build on, modify, or challenge each other. When you can explain why Treisman modified Broadbent, or how Load Theory reconciles the early vs. late selection debate, you're thinking at the level the course demands.
Early Selection Theories: Filtering Before Meaning
These theories propose that attention acts as a filter before semantic processing occurs. The central question: how does the brain decide what gets fully processed when sensory input exceeds our capacity?
Broadbent's Filter Theory
Broadbent proposed the first formal model of selective attention, based largely on dichotic listening experiments where participants hear different messages in each ear.
- Early filtering based on physical characteristics. Information is selected based on features like pitch, loudness, or spatial location before any meaning is extracted.
- Single-channel bottleneck allows only one stream of information through at a time, creating selective attention.
- Unattended information is completely blocked. This explains why participants in dichotic listening tasks can report almost nothing about the content of the unattended ear.
The major limitation: Broadbent's model can't explain why people sometimes notice meaningful information (like their own name) in an unattended channel. That's exactly the problem Treisman set out to fix.
Treisman's Attenuation Theory
- Attenuated rather than blocked. Unattended information is turned down like a volume dial, not completely filtered out. The filter is "leaky."
- Dictionary unit with variable thresholds. Each word or concept has an activation threshold. Most thresholds are high, so attenuated signals can't reach them. But biologically or personally significant stimuli (like your own name) have permanently low thresholds, so even a weakened signal can activate them. This explains the cocktail party effect: hearing your name across a noisy room.
- Modified early selection. Treisman keeps Broadbent's basic architecture (early filtering based on physical features) while accounting for the breakthrough of meaningful stimuli.
Compare: Broadbent vs. Treisman: both propose early selection based on physical features, but Treisman allows important information to break through the filter via attenuation rather than complete blocking. If an FRQ asks about the cocktail party effect, Treisman's attenuation model is your go-to explanation.
Late Selection Theories: Processing Everything First
These models argue that all incoming information receives full semantic analysis before attention selects what reaches consciousness. Selection happens after meaning is extracted, not before.
Deutsch and Deutsch's Late Selection Theory
- Full semantic processing of all inputs. Every stimulus is analyzed for meaning before the attentional filter operates.
- Selection based on relevance and importance rather than physical characteristics. The filter decides what reaches awareness only after comprehension has already occurred.
- More unconscious processing than early selection models assume. This easily explains semantic breakthrough effects, since all inputs are fully processed anyway.
The main criticism: processing every single input for meaning seems metabolically expensive. Why would the brain invest resources analyzing information it will ultimately discard?
Compare: Early selection (Broadbent/Treisman) vs. Late selection (Deutsch & Deutsch): the core debate is when filtering occurs. Early selection is more efficient but struggles to explain semantic breakthrough. Late selection handles meaning-based effects naturally but raises questions about metabolic cost.
Capacity and Resource Models: Attention as Energy
Rather than focusing on when selection occurs, these theories ask how much attention we have and how it gets distributed. They treat attention as a limited pool of mental energy.
Kahneman's Capacity Model
Kahneman (1973) shifted the conversation from filtering to resource allocation. Instead of asking where the bottleneck is, he asked how a finite supply of mental effort gets divided.
- Attention as a flexible, limited resource that can be allocated across tasks based on current demands and arousal level.
- Mental effort varies with task difficulty. Harder tasks consume more of the available pool, leaving less for anything else.
- Central capacity allocation is influenced by three factors: enduring dispositions (things you habitually attend to), momentary intentions (your current goals), and evaluation of task demands (how hard something is).
Multiple Resource Theory (Wickens)
- Separate resource pools for different modalities. Visual, auditory, verbal, and spatial tasks draw from distinct reservoirs rather than one central pool.
- Cross-modal multitasking is easier because tasks using different resource pools (e.g., listening to directions while scanning the road) interfere less than tasks drawing on the same pool.
- Task compatibility determines interference. This is why texting while driving (both heavily visual) is far harder than talking while driving (auditory + visual). The two visual tasks compete for the same limited resource.
Compare: Kahneman vs. Multiple Resource Theory: Kahneman proposes one central pool of attention. Wickens' Multiple Resource Theory proposes several specialized pools. Use Multiple Resource Theory to explain why some dual-task combinations work better than others.
Visual Attention and Search: Finding What Matters
These theories explain how attention operates specifically in visual perception: how you search scenes, combine features, and shift focus across space.
Feature Integration Theory (Treisman & Gelade)
This theory addresses the binding problem: how does the brain combine outputs from separate feature detectors (color, shape, orientation, motion) into a unified perception of a single object?
- Two-stage processing model. In the preattentive stage, basic features are detected automatically and in parallel across the visual field. In the attentive stage, focused attention is required to bind those features together into coherent objects.
- Pop-out vs. conjunction search. A red target among green distractors "pops out" in the preattentive stage (only one feature distinguishes it). But finding a red circle among red squares and green circles requires serial, attentive search because you need to bind color and shape.
- Illusory conjunctions occur when attention is divided or overloaded. Features get incorrectly combined (you might "see" a red circle when the display actually contained a red square and a green circle). This is direct evidence that binding requires focused attention.
Spotlight Theory of Visual Attention
- Attention as a moveable spotlight that enhances processing of stimuli within its beam while leaving areas outside the beam less processed.
- Flexible in size and location. The spotlight can narrow for detailed analysis (reading fine print) or widen for broader monitoring (scanning a crowd).
- Enhanced processing within the beam explains faster reaction times and better accuracy for stimuli at attended locations compared to unattended ones.
Guided Search Theory (Wolfe)
- Combines bottom-up and top-down guidance. Attention is directed by both stimulus salience (a bright flash or sudden movement) and your current goals (looking for your keys on a cluttered desk).
- Parallel feature processing feeds serial search. Basic features are detected simultaneously across the scene, generating an activation map. Attention is then guided to locations with the highest activation (most likely to contain the target).
- Prior knowledge shapes search efficiency. This explains why a radiologist spots a tumor on an X-ray faster than a novice. Top-down expectations tune the activation map.
Compare: Feature Integration Theory vs. Guided Search Theory: both involve parallel feature detection followed by focused attention, but they emphasize different things. Feature Integration Theory focuses on binding (how features combine into objects). Guided Search focuses on finding (how attention is directed to likely targets using top-down knowledge).
Selective Attention and Competition: What Wins?
These theories address how attention handles competing stimuli and distractors. What determines which information gets processed and which gets suppressed?
Load Theory of Selective Attention (Lavie)
This theory is especially important because it offers a resolution to the early vs. late selection debate.
- Perceptual load determines distractor processing. Under high perceptual load (a demanding visual task that uses up capacity), there are no resources left to process distractors, so they're effectively filtered early. Under low perceptual load, spare capacity automatically spills over to process distractors, producing late-selection-like effects.
- Reconciles early vs. late selection by showing that both can occur depending on task demands. Early selection isn't wrong; late selection isn't wrong. The answer depends on the load.
- Cognitive load works differently. High cognitive load (heavy working memory demands) actually increases distractor interference because executive control resources needed to maintain focus are consumed. Don't confuse perceptual load with cognitive load on an exam.
Biased Competition Theory (Desimone & Duncan)
- Stimuli compete for neural representation. When multiple objects fall within the same receptive field, their neural responses mutually suppress each other.
- Attention biases the competition by enhancing the neural signal for the relevant stimulus while inhibiting competitors. This has been directly observed in single-cell recordings and fMRI studies.
- Dynamic and context-dependent. Which stimulus wins the competition shifts based on goals, stimulus salience, and task demands.
Compare: Load Theory vs. Biased Competition: Load Theory emphasizes capacity limits determining what gets filtered. Biased Competition emphasizes active neural suppression of losing stimuli. Both explain why distractors sometimes break through and sometimes don't, but from different levels of analysis (cognitive vs. neural).
Quick Reference Table
|
| Early selection (filter before meaning) | Broadbent's Filter Theory, Treisman's Attenuation Theory |
| Late selection (filter after meaning) | Deutsch and Deutsch's Late Selection Theory |
| Attention as limited capacity | Kahneman's Capacity Model |
| Multiple resource pools | Wickens' Multiple Resource Theory |
| Feature binding and visual search | Feature Integration Theory, Guided Search Theory |
| Spatial attention | Spotlight Theory |
| Distractor processing | Load Theory of Selective Attention |
| Neural competition | Biased Competition Theory |
Self-Check Questions
-
Both Broadbent's Filter Theory and Treisman's Attenuation Theory propose early selection. What key modification did Treisman introduce, and what phenomenon does it explain?
-
How would you use Load Theory to reconcile the debate between early and late selection models? Under what conditions would each type of selection occur?
-
Compare Feature Integration Theory and Guided Search Theory: what do they share in their approach to visual attention, and how do their emphases differ?
-
If someone can easily listen to a podcast while cooking but struggles to text while driving, which theory best explains this difference, and why?
-
An FRQ asks you to explain why you might hear your name spoken across a noisy room even when you weren't listening to that conversation. Which theories are relevant, and how would you structure your response?
-
A participant in an experiment makes illusory conjunctions (reporting a "blue X" when the display contained a blue O and a red X). Which theory predicts this, and what does it tell you about the role of attention in perception?