Cognitive Psychology: Foundations, Brain Systems, and Perception

Posted on Oct 7, 2025 in Psychology

Foundations of Psychological Research

Level and Scope of Psychological Explanations

Psychological phenomena can be explained at different levels:

Some psychological explanations are based on biology.
Some psychological explanations are based on mental states.
Some psychological explanations are based on social/cultural factors.

The scope of psychological research refers to whether it applies to:

All human beings.
Certain groups of people (e.g., people with schizophrenia).
Individual people (e.g., my mother).
Specific actions by a specific individual (e.g., why my mother called me by my brother’s name last week).

Cognitive psychology started as the scientific study of knowledge:

How is knowledge acquired?
How is knowledge retained?
How is knowledge used?

Historical Foundations of Cognitive Psychology

All psychology originated in philosophy.

Epistemology: How do we gain knowledge?

Plato: Founder of *Rationalism* (understanding the world purely by rational analysis, without empirical observation).

We are born with the knowledge inside our mind – we just need to find a way to get it out.
Case in point: Mathematics can predict the world even before we get to observe it.

Aristotle: Founder of *Empiricism* (we need to observe the physical world to understand it).

The Cognitive Revolution

A new approach to the field of psychology emerged in the 1950s and 1960s.

Cognitive psychology arose partly from the limitations of previous research traditions:

Introspection
Behaviorism

Limits of Introspection and Behaviorism

The Limits of Introspection

Wilhelm Wundt and his student Edward B. Titchener began the study of experimental psychology in the late 1800s.

They believed psychology should focus on studying conscious mental events.

They were “Structuralists”: they understood the Mind as a series of discrete units of processing.

The only person who can directly experience or observe someone’s thoughts is that person.

Introspection — “looking within”

Observing and recording your own thoughts and experiences.
Required systematic training.

Problems with introspection:

Some thoughts are unconscious.
It is often impossible to test claims made via introspection.
Cannot directly observe or measure thoughts.
Self-report accuracy is often unknown.

The Gains from Wundt and the Structuralists:

They treated psychology (the study of the mind) as a science.
They used Reaction Time (RT; also called Response Time) to quantify mental processes.
RT is the most frequently used variable in cognitive research.

The Behaviorist Movement

The behaviorist movement dominated psychology in America for more than the first half of the 20th century.

Focused on Learning: observable behaviors and stimuli, not mental events (whether mental processes exist or not, they are irrelevant, because they cannot be scientifically studied).

B.F. Skinner developed “a technology of behavior.”

Problems with behaviorism became apparent by the late 1950s.

Language appears spontaneously, without obvious associative learning (which was a central tenet of Behaviorism).
Stimulus-response accounts, in many cases, are not enough to explain behavior.
The same stimulus can elicit different behaviors.

Intellectual Foundations of the Revolution

From introspection and behaviorism, experimental psychologists learned that:

Introspective methods for studying mental events are not scientific.
We need to study mental events in order to understand behavior.

The Transcendental Method

Transcendental method:

Knowledge transcends sensory experience – also requires understanding of how we process that experience.
Reason backward from observations to determine the cause.
“Inference to best explanation” (the heart of most modern science).
Analogous to a detective using clues left behind to figure out how a crime was committed.

Cognitive psychologists study mental events indirectly.

Measure observable stimuli and responses.
Develop hypotheses about mental events.
Design new experiments to test these hypotheses.

Behaviorists argued that learning was simply a change in behavior.

Edward Tolman argued that learning also involved the acquisition of new knowledge.

Using rats in a maze he demonstrated that learning could occur without apparent changes in behavior.

Tolman demonstrated that rats acquire a “cognitive map” of their locations without changes in behavior.

Days 1–10: Rats were placed in a maze (without food). The rats wandered through the maze. There were no changes in behavior across the days.
Day 11: Food was added to one location in the maze, and the rats found it.
Day 12: The rats immediately ran to where the food had been the day before.

Behaviorists argued that language acquisition and use could be understood in terms of behaviors and rewards (i.e., associative learning of actions and outcomes).

Noam Chomsky argued that this does not explain the creativity of language.

How do we produce and understand new sentences?

European Roots and Gestalt Psychology

Gestalt psychologists in Europe argued that mental processes and behaviors cannot be understood without considering the “whole.”

The notion that perceivers shape their own experience is a central theme in modern cognitive psychology.

Bartlett suggested that people spontaneously use schemas to interpret experiences and aid memory.

Computers and Information Processing

Psychologists began to consider that the human mind might use processes and procedures like a computer.

Data were explained in terms of computer terminology (e.g., “buffers,” “central processors”).
This led to the Information-processing approach.

Research Methods in Cognitive Psychology

Process of research in cognitive psychology:

Form a hypothesis.
Derive predictions from the hypothesis.
Collect data to test predictions.
Confirm hypothesis or modify (or reject) the hypothesis.

Examples of methodologies:

Performance or accuracy measures.
Response time (RT) measures.
Neuroimaging techniques.
Cognitive neuroscience.
Clinical neuropsychology: Study of brain function based on damaged brain structures. For example, Patient H.M., studied by Brenda Milner (McGill University).

Neural Basis of Cognition

Neurons, Glia, and Synaptic Transmission

Basic parts of a neuron:

Dendrites: detect incoming signals from other neurons.
Cell body: contains the nucleus and cellular machinery.
Axon: transmits signals to other neurons through the specialized axonal terminals, which form synaptic connections with follow-up neurons.

The Synapse

Neurotransmitters change the postsynaptic membrane by triggering the flow of charged atoms (ions) through.

If there is sufficient ionic flow to surpass the neuron’s threshold, an action potential is produced.
All-or-none law: an action potential (colloquially called “spike”) is always of the same magnitude.
Signal frequency (in spikes/sec) can vary, depending on the stimulus.
Typically, action potentials are generated at the cell body and travel down the axon. An action potential always travels in one direction along the axon.
Myelination of axons (akin to cable insulation) makes the spike propagation very fast, but not instantaneous (about 100 m/sec). How does this relate to RT?

More than 100 neurotransmitters have been identified so far. Their functional effect mostly depends on which brain circuits they operate in. E.g., different neurotransmitters play a role in mood, in anxiety, or in attention.

Synaptic transmission:

Allows a neuron to receive (at dendrites) and integrate (at cell body) information from many other neurons.
Neurons can compare incoming signals and adjust responses to input as per other incoming signals.
Strengths of synaptic connections are adjustable: Neural plasticity.
Crucial for learning and memory.
Altered by experience, development, aging, pathology.
Can also be altered pharmacologically: e.g., anxiolytic, antidepressant, antipsychotic medication and others.

Glia

Guide development of nervous system.
Repair damage in the nervous system.
Control nutrient flow to neurons, including oxygen.
Electrical insulation (myelination of axons).

Brain Structures: Hindbrain, Midbrain, Forebrain

The human brain has three main structures:

Hindbrain
Midbrain
Forebrain

Hindbrain

Brainstem: Medulla and pons (Key life functions).
Cerebellum: largest region of the hindbrain (Movement coordination and balance).

Midbrain

Midbrain functions include:

Coordinating precise eye movement.
Relaying auditory information from ears to forebrain.
Regulating pain experiences.

Forebrain

The forebrain surrounds the midbrain and most of the hindbrain.

Includes the cerebral cortex (cerebrum), four lobes, and subcortical structures.

Cortex: outer surface of the forebrain; approximately 80% of the brain.

Divided into two cerebral hemispheres by the longitudinal fissure.

Subcortical Structures

The subcortical parts of the forebrain include:

Thalamus: sensory relay station.
Hypothalamus: controls behaviors that serve specific biological needs (e.g., eating).
Limbic system:
- Amygdala: emotional processing.
- Hippocampus: learning and memory.

Lateralization and Hemispheric Specialization

The brain is divided into roughly symmetrical left and right hemispheres.

Contralateral cortical organization:

E.g., left visual hemifield goes to Right hemisphere. Right hand is controlled by left hemisphere.

Connected by commissures (thick bundles of fibers that carry information).

Largest commissure is the corpus callosum.

Split-brain patients:

Severed corpus callosum.
Was a last-resort treatment for severe epilepsy.
Severely limits communication between the hemispheres.

Evidence for some hemispheric specializations of functions:

Language mostly controlled by the left hemisphere.

Cognitive Neuroscience Methods

Cognitive neuroscience relies on a variety of methods to study the brain and nervous system.

Neuropsychology
Neuroimaging
Electrical recordings
Manipulation of brain function

Neuropsychology and Neuroimaging

Neuropsychology: the study of the brain’s structures and their relation to function.

Clinical neuropsychology
Lesions

Structural Neuroimaging Techniques

Computerized axial tomography (CT) scans: fast; inexpensive.
Magnetic resonance imaging (MRI) scans: finer detail; not for everyone.

Functional Neuroimaging Techniques

Positron emission tomography (PET) scans: use radioactive substances to trace them in the brain, including oxygen.
Functional magnetic resonance imaging (fMRI) scans: trace flow of oxygenated blood without radioactive substances.
- Fusiform Face Area gets activated when you perceive a face.
- Parahippocampal Place Area gets activated when you perceive a place.

Electrical Recording and Brain Manipulation

Data from Electrical Recording

Communication between neurons is chemical. Neurons communicate with one another via neurotransmitters.

Communication within a neuron is electrical.

“Input” end of a neuron receives neurotransmitters; “output” end releases neurotransmitters.
An electrical impulse (action potential) conveys the signal from the input end to the output end.

Electroencephalogram (EEG):

Recording of the electrical communication within neurons.
Used to study:
- Broad rhythms (e.g., sleep stages).
- Event-related potentials (ERPs): a brief response to any sensory or motor event.

Manipulations of Brain Function

Brain function can also be studied through techniques that manipulate the functions.

Chemical effects on neurotransmitters.
Electrical stimulation.
Gene manipulation.

Every method has strengths and weaknesses.

EEG: Strength: temporally locating neural activity (when?); Weakness: spatially locating neural activity (where?).
fMRI scans: Strength: spatially locating neural activity (where?); Weakness: temporally locating neural activity (when?).
MRI scans: detect brain structures, not activity.

Researchers can overcome limitations by combining techniques.

Problem: Most neuroimaging techniques used to study brain activity and structures provide only correlational data.

Sources of causal data:

Brain lesions.
Transcranial magnetic stimulation (TMS): Magnetic pulses activate neurons; produces temporary lesions.

Localization of Function

Effort to determine the function of specific brain structures. What’s happening where within the brain?

The Cerebral Cortex and Projection Areas

Largest portion of the human brain (in volume).
Thin layer of tissue covering the cerebrum (i.e., forebrain).

Regions of the cortex:

Primary motor projection areas: Frontal lobe; departure points in the motor cortex for signals that control muscle movement.
- Contralateral control.
- Topographical organization.
- More cortical coverage reflects greater motor precision.

Sensory Areas

Primary sensory projection areas: arrival points of sensory signals.

Somatosensory area: parietal lobe; skin sensations.
Primary auditory cortex: temporal lobe; auditory sensations.
Primary visual cortex: occipital lobe; visual sensations.
- Contralateral topographical organization.
- Cortical space assigned based on acuity.

Association Areas

Approximately 75% of the cerebral cortex.

Contain many specialized subregions, damage to which can result in:

Apraxia: problems with the initiation or organization of voluntary action.
Agnosia: problems identifying familiar objects, typically in one modality.
Unilateral neglect syndrome: problems in which one visual hemifield is ignored.
Aphasia: problems producing or understanding language.

Visual Perception and the Visual System

Perception seems fast, easy, and automatic, but consider people with akinetopsia (e.g., Patient L.M.):

Unable to perceive motion.
See “nothing between” location changes.
For example, parked cars versus moving cars; pouring coffee.

Many separate and complicated processes are involved in producing cohesive perception.

Sensitivity, Acuity, and Light

Vision is the dominant sense in humans.

More brain area is devoted to vision than the other senses.
If visual inputs conflict with other sensory inputs, visual inputs generally dominate final perception. For example, ventriloquism.

Two Properties of Seeing: (A) Sensitivity (reflects the dimmest light you can detect).

Two Properties of Seeing: (B) Acuity.

What is light?

Light is a form of electromagnetic radiation.
It behaves as a wave (when light moves around the world).

Photoreceptors

The cornea and lens focus the light on the retina.

The retina contains two types of photoreceptors:

Rods
Cones

What is wavelength?

The spatial distance between two consecutive positive (or negative) peaks.
The lower the frequency, the longer the wavelength.

The Visual Pathway: Retina to V1

A series of neurons and pathways communicate information from the retina to the visual cortex.

From the eye:

Photoreceptors
Bipolar cells
Ganglion cells

To the thalamus via the optic nerve:

Lateral Geniculate Nucleus (LGN)

Then to the occipital lobe:

Area V1, the primary visual projection area, or primary visual cortex, which is located in the occipital lobe.

Lateral Inhibition and Edge Enhancement

Mach bands demonstrate that cells linking the retina and brain already perform computations analyzing the visual input.

Inhibitory synaptic transmission: not all neurotransmitters drive generation of action potentials in postsynaptic neurons – sometimes neurotransmitters inhibit (i.e. block) postsynaptic neurons from firing.
Lateral inhibition: when cells are stimulated, they inhibit the activity of neighboring cells. (Winner take all architecture).
Results in edge enhancement, which allows us to detect the boundaries of objects.

Visual Coding and Receptive Fields

Coding: the relationship between activity in the nervous system and the stimulus that is somehow represented by this activity.

Single Neurons and Single-Cell Recording

Knowledge of the visual system partly comes from single-cell recordings.

Each cell in the visual cortex has a receptive field (RF).
Neuron firing rates depend on the stimulus presented within the neuron’s RF.
Which patterns or stimuli drive that neuron to fire spikes?

Multiple Types of Receptive Fields: Retina and LGN

Center-surround cells (“dot detectors”):

A stimulus in the center of the receptive field leads to faster firing rates.
A stimulus in the surrounding areas of the receptive field leads to slower, below-baseline firing rates (because lateral inhibition).
A stimulus covering the entire receptive field has the same effect as no stimulus.

Multiple Types of Receptive Fields: visual cortex.

Parallel Processing and the Binding Problem

Advantages of parallel processing:

Speed and efficiency.
Mutual influence among multiple systems.
Resolves contradictory demands.

The What and Where Systems

What system (ventral pathway):

Pathway connecting the occipital lobe and inferotemporal cortex.
Aids in identification of visual objects.
Damage leads to visual agnosia.

Where system (dorsal pathway):

Pathway connecting the occipital lobe and posterior parietal cortex.
Aids in perception of an object’s location.
Damage leads to difficulties reaching for objects.

Putting the Pieces Back Together

Characteristics of a stimulus are processed separately, but perception is coherent and integrated. How?

The binding problem: the task of reuniting elements of a stimulus that were addressed by different systems in different brain regions. For example, how are the what and where systems coordinated?

Elements that help solve the binding problem:

Spatial position: Overlay map of “which forms are where” with map of “which colors are where,” “which motions are where,” etc.
Neural synchrony: Attributes are registered as belonging to the same object if the neurons detecting these attributes fire in synchrony.
Attention: With insufficient attention, conjunction errors are common.

Form Perception and Gestalt Principles

Perception of stimuli goes “beyond the information given.”

Gestalt psychologists: the perceptual whole is often different than the sum of its parts.

Demonstrated by reversible (or ambiguous) figures. For example, the Necker cube or neutral figure/ground organization.

Gestalt Principles

Many stimuli (not just reversible figures) are ambiguous and require interpretation.

We assume more than what is definitively present.

Most often we are not conscious of the assumptions our brain makes.

Our ability to interpret ambiguous scenes is governed by a few basic principles.

Organization and Features

Perception involves multiple activities going on in parallel:

Information gathering (along multiple dimensions).
Interpretation.

Neither step in the perceptual process goes first.

Perceptual Constancy and Depth

Perceptual constancy: we perceive constant object properties (sizes, shapes, etc.) even though sensory information about these attributes changes when the viewing circumstances change.

Brightness constancy (we correctly perceive the brightness of objects whether they’re illuminated by dim or strong light).
Size constancy (we correctly perceive an object’s size despite the changes in retinal-image size created by changes in viewing distance).
Shape constancy (we correctly perceive an object’s shape despite changes in its shape on the retina).

Constancy is partly influenced by relationships within the retinal image.

The relationship between objects in the visual field stay the same on the retina regardless of your viewing distance.

Distance cues also contribute to constancy.

As object distance increases, the activated area of retina decreases, same as if the object was smaller. And yet, we do not mistake the object size.

Unconscious inference: As object size decreases by half, our brain doubles the size of retinal input.

The Perception of Depth

We need to know distance to be successful at size judgments and to interact successfully with the surrounding world.

Perception of distance depends on various distance cues:

Binocular disparity.
Monocular cues.
Motion cues.

Binocular disparity: the difference between each eye’s view of a stimulus. Can lead to perception of depth even in the absence of other distance cues.

Monocular distance cues: depth cues that depend only on what each eye sees by itself.

Lens adjustment.
Pictorial cues, including interposition.
Linear perspective.
Texture gradients.

Motion also provides cues for judging distance and depth.

Motion parallax (objects cross our visual field at a different pace, depending on their relative distance).
Optic flow (motion toward or away from an object changes the pattern of stimulation across the entire visual field).

The Role of Redundancy:

There is some redundancy across different cues. However, different cues become more or less important under different circumstances. For example, binocular disparity is only informative when objects are relatively close to the viewer.

Object Recognition Processes

Patients with apperceptive agnosia can perceive an object’s features but not the object in its entirety.

Patients with associative agnosia do see the object but cannot name it.

Cannot link the visual input to visual knowledge.
They can draw well from memory.

Recognition: Bottom-Up and Top-Down Processing

The process of object recognition is complex. Complicated by:

Variations in “stimulus input.”
Contextual influences.

Bottom-up processing: processes that are directly shaped by the stimulus (Data driven).

Top-down processing: processes shaped by knowledge (Concept-driven; Example: context effects).

The Importance of Features

Recognition begins with identifying visual features in the input pattern.

Dots, vertical lines, curves, diagonals, etc.
Evidence for feature detectors in the visual system.

Larger units (i.e., more complex features) can then be detected by assembling the smaller units.

Visual search tasks: tasks in which participants examine a display and judge whether a particular target is present.

Efficient when target is defined by a simple feature.
Slow when target is defined by a combination of features.

The difference in RT suggests that feature analysis is a separate step from the step in which detected features are combined.

Simple features are detected first, followed by complex features.

Evidence suggests that object recognition begins with feature detection. But how are these features assembled into complete objects?

Word Recognition and Context Effects

Can people recognize briefly presented stimuli?

Brief presentation of stimuli (approx. 20–30 ms) using a Tachistoscope (device used to present stimuli for precise amounts of time, now often simulated with modern computers).

The Word-Superiority Effect

Word-superiority effect (WSE):

It’s easier to perceive and recognize letters-in-context (entire words) than letters in isolation.

The word-superiority effect can be demonstrated using a “two-alternative, forced-choice” procedure.

Participants are shown a word (e.g., DARK) or a single letter (e.g., E).
Question: Did the display contain an E or a K?

Participants are more accurate when the original stimulus was a word rather than a single letter.

Degree of Well-Formedness

Well-formedness: how closely a letter sequence conforms to the typical patterns of spelling in the language.

The more well-formed a letter sequence is, the easier it is to recognize the sequence, and the greater the context effects produced by the sequence on recognition (e.g., HZYQ < FIKE < HIKE).

Well-formedness also influences errors.

A likely error: DPUM misread as DRUM.
An unlikely error: DRUM misread as DPUM.

People can perceive stimuli as being more regular than they actually are (e.g., TPUM misread as TRUMPET).

Feature Nets and Distributed Knowledge

The Design of a Feature Net:

Each detector in the network has an activation level.

With input, this activation level increases.
Some detectors will be easier to activate than others.

Detectors fire when their response threshold is reached (similar to a neuron’s threshold for firing an action potential).

Individual detectors are probably complex assemblies of neural tissue, not individual neurons or groups of neurons.

Starting activation levels depend on:

Recency: detectors that have fired recently will have higher activation levels.
Frequency: detectors that have fired frequently will have higher activation levels.

Priming will depend on frequency and recency. The network is biased to respond to inputs similarly to how it has responded previously.

Recognition Errors:

The bias to recognize frequent or primed words can result in errors, but it also results in correct recognition more often than incorrect recognition. It helps more than it hurts!

Distributed Knowledge: Knowledge is not locally represented. Rather, feature nets contain distributed knowledge. Knowledge in a network is reflected by relationships across detectors.

Efficiency versus Accuracy:

The same mechanisms that enable the network to resolve ambiguous inputs (e.g., sloppy handwriting), and recover from errors, can also result in recognition errors. Perfect accuracy is sacrificed for efficiency.

Recognition by Components (RBC) Model

Descendants of the Feature Net:

McClelland & Rumelhart model: Emphasizes the role of inhibitory connections among detectors. Information flows bottom-up, top-down, and within the same level. Higher-level word detectors can influence lower-level detectors (top-down flow of information).
Recognition by Components (RBC) model: Applies the feature net model to recognition of 3-dimensional objects.

Hierarchy of detectors in RBC:

Feature detectors (e.g., curves, edges).
Geon detectors.
Geon assemblies representing relations between geons.
Object model: a representation of the complete, recognized object.

A geon can be identified from virtually any angle. The object model can be activated, and the object can be recognized.

Most objects can be recognized from just a few geons. Partial occlusion of objects does not necessarily prevent recognition.

Face Recognition and Holistic Perception

Object Recognition and the Brain

Cells in the Inferotemporal (IT) cortex might be the biological foundation for word detectors or object detectors. Viewpoint-dependent cells may trigger viewpoint-independent cells.

Faces Are Special

Recognizing faces specifically seems to involve specialized neural structures.

Prosopagnosia: an inability to recognize individual faces (including their own) despite otherwise normal vision.

Some people are super-recognizers, with extremely accurate face recognition (No other perceptual or memory-based advantages).

Faces show a powerful inversion effect. The effect is much larger than for other types of stimuli (e.g., houses).

Some researchers suggest that other types of recognition are special in the same way as facial recognition. Example: A bird-watcher who developed prosopagnosia lost the ability to distinguish faces and types of warblers.

The fusiform face area (FFA) is particularly responsive to faces, but high levels of activation in an expert’s brain can also be produced by tasks requiring subtle distinctions.

In contrast to other object recognition, face recognition seems to depend more on holistic perception: perception of the overall configuration rather than an assemblage of parts.

Evidence from the composite effect: easier to recognize when the two halves are misaligned.

Top-Down Influences on Object Recognition

Limits of feature nets:

Some target objects depend on configurations, not individual features.
Knowledge that is external to object recognition may still influence recognition.

The Benefits of Larger Contexts:

Some top-down effects can be explained by feature nets (e.g., the word-superiority effect).

Other top-down effects, however, require more explanation:

Larger priming effects are evident for words viewed in a sentence rather than in isolation – but only if the sentence makes sense.
Context and expectations influence perception. E.g., it’s more likely to mishear “I got scared when I saw what it’d done to him” as “. . . when I saw what I’d done to him,” if we think it’s a crime suspect talking versus a job candidate talking.

Attention and Selective Processing

Patients who suffer from unilateral neglect syndrome are unable to attend to inputs coming from one side of the body. Typically, this results from damage to the right parietal cortex (due to stroke), causing contralateral neglect.

Selective attention refers to the skill through which we focus on one input or one task while ignoring other stimuli.

Dichotic listening tasks: different audio inputs presented to each ear via headphones.

Participants were instructed to pay attention to one input (the attended channel) and ignore the other input (the unattended channel).
Shadowing: repeat out loud the information from the attended channel. Confirmed participants were paying attention to the attended channel.

Sometimes the effects of selective attention are so strong that we fail to see stimuli that are directly in front of our eyes.

Participants are generally clueless about the semantic content of the unattended channel.

Not all aspects of the unattended channel are ignored, though. Not ignored:

Physical attributes (e.g., speech versus music; speaker’s gender).
Personally important semantic content (e.g., your name).

How can we explain general insensitivity to the unattended channel, along with the fact that some information does leak through?

One proposal is that we block unattended inputs with a filter. This filter blocks potential distractors, while attended inputs are not filtered out.

Theories of attention need to be able to explain how we:

Inhibit new or unexpected distractors.
Promote the processing of desired stimuli.

Selective Attention and Filtering

Inattentional Blindness and Change Blindness

Inattentional blindness: the failure to see a prominent stimulus, even if one is staring right at it, because you don’t expect that stimulus, or because you’re focused on something else.

Inattentional deafness: the auditory corollary.
Inattentional numbness: the haptic corollary.

Change blindness: the inability to detect changes in a scene despite looking at it directly, possibly even while looking for changes.

Early vs. Late Selection Hypotheses

Inattentional blindness and change blindness could result from:

A failure to perceive the stimulus (perception limit).
A failure to remember the stimulus (memory limit).

Early selection hypothesis:

Only the attended input is analyzed and perceived.
Unattended information receives little or no analysis (Never perceived).

Late selection hypothesis:

All inputs are analyzed.
Selection occurs after analysis.
Selection may occur before consciousness or later.
Unattended information might be perceived, but is then forgotten.

There is evidence both for the early selection hypothesis and for the late selection hypothesis.

Late selection—stimuli that are not attended to (don’t reach consciousness) can nevertheless affect perception.
Early selection—electrical brain activity for attended inputs differs from activity for unattended inputs within 80 ms.

Selection via Priming and Biased Competition

Selection may be a consequence of priming based on your expectations.

Perceiver anticipates the attended channel.
Detectors that are needed for the (now expected) input are primed.
Primed detectors fire more readily.

Regardless of expectation, some high-frequency or salient information is already primed (Example: our name).

Biased competition theory: attention creates a temporary bias in neurons’ sensitivity.

Neurons receive input from attended stimuli and distractors.
Attention adjusts neurons’ priorities.
More responsive to input with desired properties.
Less responsive to everything else.
Desired inputs receive further processing.
Distractor inputs do not.

Repetition priming: priming produced by a prior encounter with the stimulus (Stimulus-driven; Requires no effort or cognitive resources).

Expectation-driven priming: detectors for inputs you think are upcoming are deliberately primed (Effortful; Not done for unexpected inputs or inputs in which you have no interest).

Spatial Attention and the Spotlight Model

Effects of repetition priming and expectation-driven priming come from studies of spatial attention: our ability to focus attention on a specific location in space.

Posner et al. (1980):

Goal: press a button as soon as the target appears, while focusing on the central fixation mark.

Targets presented to the left or right of the fixation mark.
Prior to each trial:
- Neutral cue to signal the start of the trial, or
- Arrow indicating location of the upcoming target (with 80% accuracy).

Differences in RT across the different validity condition of the cues:

Repetition priming does not have a cost.
Expectation-based priming does have a cost.
Participants perform worse in trials when they are misled than when they have no expectations.
Requires mental resources.

The costs of expectations reveal the presence of a limited-capacity system.

Attention as a Spotlight

Spatial attention is sometimes thought of as a spotlight beam.

Can be moved anywhere in the visual field, independently of the eyes.
Scope can be widened or focused.
The area outside the spotlight is not completely in the dark.

Movement of attention – not movements of the eyes.

Benefits of attention can appear without eye movements (covert attention), or prior to any eye movement:

Eye movements take approximately 180 ms.
Shifts in attention to primed stimuli are detected within 150 ms.

The control system for attention (sources of attentional signals) includes:

Orienting system (flexibly shifting attention, as when driving): Disengage attention from one target, shift attention to a new target, engage attention on the new target.
Alerting system (sustaining attention, as when studying): Maintain alert state in the brain.
Executive system: Control voluntary actions.

Which factors influence what people attend to?

Visual prominence.
Level of interest.
Importance.
Beliefs and expectations.
Ignore highly predictable items.
Ignore ultra-rare items (“if you don’t find it often, you often don’t find it.”).
Culture (individualistic cultures focus more on individual people and objects; collectivist cultures focus more on how people and objects relate to each other).

Endogenous control of attention (goal-directed): Example: searching for your phone.

Versus

Exogenous control of attention (stimulus-driven): Example: ambulance sirens.

Attending to Objects or Attending to Positions?

We generally pay attention to objects and positions in space.

Evidence from unilateral neglect syndrome for spatial attention deficit.
Experiment with rotating barbell suggests object attention does not terminate when object enters neglected space.

Feature Binding

Attention plays a key role in solving the binding problem.

Feature integration theory involves two stages:

Preattentive stage: Parallel processing of the stimulus (Efficient).
Focused attention stage: Expectation-based priming creates processing advantages for the stimulus.

Priming facilitates the processing of desired inputs. Also helps to prevent perception of unwanted inputs.

Priming—and thus your ability to pay attention—largely depends on your ability to anticipate the upcoming stimulus.

Divided Attention and Resource Limits

Divided attention: the skill of performing multiple tasks simultaneously.

Our limited mental resources restrict how well we can multitask. Divided attention will fail if the combination of tasks exceeds our resources.

The Specificity of Resources

Dividing attention is generally easier if the concurrent tasks are different from each other. Similar tasks (e.g., two spatial tasks) compete for the same resources.

The Generality of Resources

Even very different tasks compete with each other for mental resources.

Tasks vary in their perceptual load.
Executive control can be devoted only to one task at a time.

Executive control refers to mechanisms that allow us to:

Control our thoughts.
Keep goals in mind.
Organize mental steps.
Shift plans and change strategy.
Inhibit automatic responses.

Damage to the prefrontal cortex (PFC) can impair executive control.

Perseveration error: tendency to produce the same response over and over when the task clearly requires a change in response.
Goal neglect.

Practice Diminishes Resource Demand

Practiced skills require fewer resources or less frequent use of resources. E.g., consider a novice driver vs. an experienced driver. Practice leads to a decrease in interference between tasks.

Automaticity describes tasks that are well practiced and require little or no executive control.

Not requiring control means that they are executed without control – on auto-pilot. This may make it harder to suddenly control them when necessary.

An example of automaticity is an effect known as Stroop interference.

Where Are the Limits?

Tasks require resources, and you cannot use more resources than you have.

Some resources are task-specific and others are task-general.

If two tasks make demands upon the same resources, the result will be interference. (On the flip side, if there is interference, it suggests they share resources.)

Performance is limited by different factors in different settings:

Nature of the task.
Task novelty.
Practice.