Chapter 10 - Visual Imagery

Mental imagery

• Our ability to mentally recreate perceptual experience in the absence of a sensory stimulus.”

• You can also create mental images of stimuli that you have never experienced.

  • e.g “‘Imagine a sidewalk covered in chocolate sauce.

Dual-Coding Theory (Paivio, 1971)

• Human knowledge is stored in two systems:
  • Verbal system: abstract, symbolic, language-based
  • Non-verbal system: modality-specific, sensory-motor based

• Images function as analog codes, meaning they resemble the things they represent.

• Because images preserve perceptual features, they provide a second code that supports memory.

The Imagery Debate Overview

We know that imagery clearly exists and influences cognition (memory, language, decision making).

• The debate focuses on how mental images are represented. What format/code does imagery take in our minds?

• Kosslyn (1994): Images are depictive—they preserve spatial and perceptual properties.

• Pylyshyn (1973): Images are descriptive—symbolic, conceptual codes that do not resemble the real object.

• The question: Does the mind store images like pictures, or like statements?

Depictive vs Descriptive Representations (Details)

• Depictive representations preserve spatial layout and perceptual information.

• Descriptive representations store only conceptual relationships, not perceptual features.

• Some argue images are epiphenomena: by-products of deeper cognitive processes, not functional representations.

Propositional Codes (Pylyshyn)

• Pylyshyn argued cognitive processing relies on propositions, not pictures.

• Propositions are statements that express relationships and can be judged as true or false.

• Example: “The lamp is to the left of the books” is a proposition.

• According to him, propositional codes are sufficient to explain imagery, so picture-like representations are unnecessary.

Evidence used in the Imagery debate

Researchers have conducted experiments to resolve the imagery debate.

• Do people press mental images in the same way they process real stimuli?

  • If images are depictive (maintain perceptual and spatial characteristics) then people should process images and physical stimuli similarly.

  • If images are descriptive, then mental processing would depend on the number of propositions instead of perceptual and spatial characteristics of stimuli.

Mental Scanning (Kosslyn, 1973 & 1978)

• Asked the question, do mental images keep the same spatial characteristics of physical stimuli?

• Kosslyn predicted that if imagery works like a real picture, scanning across a longer mental distance should take longer than a short one.

• Participants memorized simple line drawings with clear structure (top/bottom or left/right).

• During the task they imagined the picture, focused on one part (for example the roots of a flower), then mentally moved to another part and pressed a button when they “reached” it.

• Reaction time increased when the target feature was farther away in the image.

• This pattern supports the idea that mental images preserve spatial relationships.

Kosslyn: Addressing alternative explanations

• The results could also be explained if participants memorized a list of features instead of a spatial image.

• In that case, longer reaction times might reflect searching through a list, not scanning across space.

• This alternative supports the descriptive/propositional view of imagery rather than the depictive view.

Mental Scanning Map Study (Kosslyn, 1978)

• To respond to the “feature list” criticism, Kosslyn used a map with several landmarks.

• Participants memorized the map, then imagined it and mentally traveled from one landmark to another.

• Travel time increased as the real distance between landmarks increased.

• The number of “features” between landmarks was kept constant, so list-searching could not explain the results.

• This provided stronger evidence that mental images preserve spatial distance, supporting depictive imagery.

Mental Rotation (Shepard & Metzler, 1971)

• Shepard & Metzler tested whether people mentally rotate images the same way they rotate real objects.

• Participants saw two 3-D block shapes and decided whether they were the same shape or different.

• When the shapes were the same, one was rotated to different angles before being shown.

• Reaction time increased as the rotation angle increased.

• This produced a linear relationship: the more rotation needed, the slower the response.

• People mentally rotated images at about 60° per second, supporting the idea that imagery uses spatial, depictive processing.

Mental Scaling (Kosslyn, 1975/1978): Size Affects What You “See”

• How much detail you can access in a mental image depends on the imagined object’s size.

• Example: imagine a cat next to an elephant vs a cat next to a butterfly.

• When the cat is imagined beside an elephant (making the cat “small”), participants took longer to answer questions like “Does a cat have claws?”

• They needed to mentally zoom in on the smaller image.

• This shows mental images behave like real perceptual objects — smaller images contain less accessible detail.

Mental Scaling (continued): Kosslyn 1975, 1978

• Participants imagined:

  • an elephant-sized butterfly,

  • a butterfly-sized elephant.

• When the “target animal” was imagined as smaller, reaction times increased.

• This effect depended on the relative size of the objects, not their real size.

• Again, this supports depictive imagery because image size affects how quickly features can be accessed.

Imagery & Perception: Interaction (Perky, 1910)

• If mental images function like internal pictures, imagery and perception should rely on similar mechanisms.

• Perky asked participants to imagine a lemon while a dim, faint image of a lemon was projected on the screen without their awareness.

• Participants described their mental images in ways that matched the projected image.

• This suggests imagery and perception can interfere with each other because they draw on shared processes.

Imagery Interferes With Perception (Segal & Fusella, 1970)

• Participants either imagined a visual stimulus (a tree) or an auditory one (a phone ringing).

• While imagining, they tried to detect a very faint real visual or auditory signal.

• Detection of the visual signal dropped when participants were imagining a visual image.

• Detection of the auditory signal dropped when imagining an auditory image.

• This selective interference suggests imagery uses the same resources as perception in that modality.

Imagery Facilitates Perception (Farah, 1985)

Imagery and Motion Aftereffects (Winawer et al., 2010)

• Normally, staring at motion in one direction causes a motion aftereffect in the opposite direction.

• Winawer showed that simply imagining motion for 60 seconds could bias later motion perception.

• This demonstrates that imagery can activate motion-sensitive visual areas strongly enough to alter perception, reinforcing overlap between imagery and perceptual neural mechanisms.

Challenges to Depictive Imagery: Shape Identification (Reed, 1974)

• Reed tested whether people can easily extract new shapes from a mental image.

• Participants memorized a complex figure, then judged whether new shapes were part of the original picture.

• In some cases they were accurate, but in many cases accuracy was low.

• Reed argued this is because people were not storing a detailed spatial image.

• Instead, they may have stored verbal labels or descriptions of the picture’s components, which do not support detailed shape extraction.

Challenges to Depiction: Demand Characteristics & Expectancy Effects

• Research on imagery may be influenced by experimenter expectancy.

  • Subtle cues from the researcher can unintentionally signal the “expected” pattern of results to participants.

• Demand characteristics can also affect performance when participants form an idea of the experiment’s purpose and adjust behavior accordingly.

• Pylyshyn argued that some support for depictive imagery may come from these effects rather than from true picture-like mental representations.

arguments against depictive representations continued

• Studies show that experimenter expectations can influence participant behavior in imagery experiments (e.g., Intons-Peterson, 1983).

• These issues do not invalidate all of Kosslyn’s findings.

• Instead, they highlight that both depictive and descriptive theories still have valid arguments.

• Advances in neuroscience allow researchers to study imagery using brain activity, which may help clarify how imagery is represented.

Studying Imagery in the Brain: Patients & Neuroimaging

• Early imagery research examined patients with localized brain damage to infer which brain areas were involved in imagery.

• New imaging techniques (PET, fMRI) allow researchers to directly compare brain activity during imagery and perception in healthy individuals.

• These tools help reveal the structure and function of mental imagery processes.

Patient TC: Loss of Perception and Imagery

• Patient TC suffered cardiac arrest and had damage in the occipital and temporal lobes, resulting in cortical blindness.

• TC was unable to distinguish light from dark, did not blink or move the head in response to moving objects, and had lost conscious vision.

• TC also showed a loss of mental imagery ability, unable to describe familiar objects, places, or tasks from memory.

• This case suggests that imagery and perception share neural systems, because damage affected both abilities.

Patient PB: Loss of Vision but Preserved Imagery (Zago et al)

• Patient PB experienced a stroke that caused occipital lobe damage and cortical blindness, similar to TC.

• However, unlike TC, PB could still perform visual imagery tasks.

• This shows that imagery and perception can sometimes dissociate, meaning they rely on overlapping but not identical neural systems.

Madame D: Impaired Perception but Preserved Imagery. Bartolomeo et al

• Madame D had multiple strokes affecting the area between the occipital and temporal lobes.

• She had partial visual ability but suffered color blindness and struggled to recognize faces, objects, or read.

• She could copy drawings (basic perception intact) but not identify what she copied.

• Despite this, she retained mental imagery and used it to help identify objects at home.

• Her case shows imagery can sometimes support perception, and again suggests partial overlap but not full equivalence between the two systems.

Imagery Lost but Perception Intact

• Some patients with closed-head injuries lost the ability to form mental images while keeping normal performance on tests of visual perception, memory, and language.

• They could not draw animals or objects from memory even though they could see normally.

• These cases show the reverse dissociation: perception without imagery.

• Taken together, patient studies reveal that imagery and perception rely on related but not identical brain systems.

Neuroimaging Supports Shared Mechanisms (Kosslyn, 1999)

• Neuroimaging studies generally show that imagery and perception use overlapping brain systems, though not perfectly identical.

• Kosslyn used PET to measure brain activity while people viewed and imagined black-and-white stripes.

• Both viewing and imagining the stripes activated primary visual cortex (V1).

• Using TMS to disrupt V1 made people less accurate during imagery tasks.

• This provides causal evidence that V1 contributes to visual imagery, not just perception.

Specialized Brain Areas Activate in Both Imagery and Perception (O’Craven & Kanwisher, 2000)

• Researchers tested whether imagery activates the same specialized regions that perception activates.

• When participants viewed or imagined faces, the fusiform face area (FFA) showed higher activity.

• When participants viewed or imagined buildings, the parahippocampal place area (PPA) showed higher activity.

• Brain activity patterns were strong enough to classify what category a person was imagining.

• suggests shared areas of perception and imagery

Imagery vs Perception: Overlap and Differences (Ganis et al., 2004)

• Used newer neuroimaging techniques to re-examine brain activity in imagery vs perception.

• Front of brain (planning, cognitive control, attention, memory) showed strong similarity for both imagery and perception tasks.

• There was limited similarity in V1 because imagery happens with no actual visual stimulus.

• Perception: stimulus hits receptor cells and travels up to early visual areas.

• Imagery: a re-enacted perceptual experience where the same neurons are driven from frontal brain areas, not from input at the eyes.

Amedi et al., 2005 (Imagery is fragile)

• During visual imagery, other nonvisual sensory areas (like auditory cortex) are deactivated.

• During perception, these other sensory areas are not turned down.

• Interpretation: imagery is more fragile than perception, so the brain “turns down” other senses to reduce interference with visual imagery.

Hub-and-Spoke Model: Neural Representation of Knowledge

• Neuroimaging research shows that knowledge is represented through both general and modality-specific systems.

• The anterior temporal lobe (ATL) acts as a semantic memory hub that stores generalized, abstract knowledge.

• Modality-specific “spokes” store context-dependent sensory and motor details across the cortex.

• This model helps explain why imagery studies sometimes show overlap and sometimes differences between imagery and perception.

MVPA (Multivoxel Pattern Analysis): Mind-Reading Imagery. Ragni et al. (2020)

• Ragni et al. (2020) tested whether brain activity patterns during imagery can reveal what a person is imagining.

• Participants viewed line drawings of lowercase letters, simple shapes, and objects.

• Later, they imagined these same stimuli.

• MVPA successfully identified which object people were imagining based on patterns of brain activity from both imagery and perception trials.

• Shows that imagery produces distinctive and decodable neural patterns.

GANs (Generative Adversarial Networks)

• GANs are a type of artificial neural network consisting of two competing parts:

  • Generator: trained to create new images that match presented examples.

  • Discriminator: trained to distinguish between real images and generated ones.

• The two networks improve by competing with each other, leading the generator to create highly realistic images.

• Goal: produce images that are indistinguishable from real ones to the point that they can fool human observers

• “Deepfakes” use GAN-based systems to impersonate people in videos.

• These videos can depict people doing or saying things they never actually did, raising ethical and social concerns.

Imagery Impacts Cognition and Behavior

• The imagery debate is not fully resolved, but most researchers agree imagery is not stored only as propositions.

• Imagery arises from brain mechanisms that overlap with perception.

• Regardless of format, imagery influences cognitive functions and behaviors.

• Examples include memory improvements such as the Picture Superiority Effect.

Picture Superiority Effect & Imagery Benefits for Memory

• Imagery improves memory compared to verbal processing alone.

• People remember pictures better than words.

• This effect is especially strong when images are interactive, as shown by Bower (1970).

• In Paivio & Csapo (1973), participants either:

  • read words or named pictures (verbal condition), or

  • formed visual images of words and pictures (imagery condition).

• After a surprise memory test, performance was higher in the imagery condition than the verbal condition.

• This demonstrates that imagery enhances encoding and recall.

Dual-Coding Explanation for Picture Superiority

• Pictures automatically generate both an image code and a verbal label, giving two memory traces.

• Words usually create only a verbal code, giving just one trace.

• More memory codes → better recall.

• Example from the slide: seeing a picture of mountains, clouds, and a hot air balloon produces both imagery and verbal labeling.

Concreteness Effect: Concrete Words Are Remembered Better

• Concrete words (e.g., “apple”) are easier to remember than abstract words (e.g., “justice”).

• Concrete words naturally elicit mental images, providing both image and verbal codes.

• Abstract words are harder to visualize, so memory relies mostly on verbal encoding.

• Parker & Dagnall (2009): participants heard lists of abstract and concrete words and were told they would later be tested on them.

• Memory performance showed the concreteness effect under normal conditions.

Dynamic Visual Noise (DVN) Disrupts Imagery

• Parker & Dagnall also tested memory while participants viewed either:

  • static visual noise, or

  • dynamic visual noise (DVN), which disrupts mental image formation.

• DVN interfered with participants’ ability to create mental images.

• Under DVN, the concreteness effect disappeared → concrete and abstract words were remembered equally well.

• This supports the idea that imagery contributes to the memory advantage for concrete words.

Imagery and Mental Health (Holmes et al., 2005–2008)

• Participants listened to short descriptions of events with either positive or negative outcomes.

• Two groups:

  • Imagery group: created vivid visual images of the scenario.

  • Meaning group: focused on the meaning of the words only.

• The imagery group reported stronger emotional reactions, including both higher anxiety for negative events and more positive emotion for positive events.

• This shows imagery intensifies emotional experience because imagining an event engages perceptual-like systems.

Imagery and PTSD: Intrusive Negative Images

• People with PTSD often experience involuntary, intrusive mental imagery of traumatic events.

• Flashbacks include vivid visual and auditory imagery and can feel like the event is happening again.

• These episodes trigger changes in the autonomic nervous system (e.g., heart rate increases, sweating).

• Individuals with more vivid mental imagery are more likely to experience intrusive images after negative events.

Anxiety Disorders and Imagery

• Anxiety disorders involve persistent and intense worrying that interferes with daily functioning.

• People with anxiety often experience increased negative imagery of future events.

• They believe these feared events are more likely to occur, which increases anxiety.

• Excessive worrying is sometimes used as a coping strategy to reduce unwanted intrusive images.

Depression and Imagery

• Depression includes persistent sadness and loss of interest.

• Depressed individuals experience more negative imagery, especially imagery linked to suicidal thoughts.

• They also experience reduced positive imagery, which may contribute to low mood.

Imagery Rescripting: Treating PTSD, Anxiety, and Depression

• Imagery rescripting is a therapeutic technique for disorders involving harmful or intrusive mental imagery.

• Patients revisit distressing memories and imagine acting differently, such as protecting their younger selves or changing the outcome.

• The goal is to replace negative emotional responses with more adaptive ones.

• Imagery-focused treatments have some of the highest success rates for PTSD and also help with anxiety and depression.

Individual Differences: Imagery Ability Varies Widely

• People differ greatly in how vivid or detailed their mental images are.

• Galton (1880) showed that some individuals describe vivid imagery while others report almost none.

• Imagery ability can be measured using:

  • Self-report questionnaires (e.g., VVIQ),

  • Objective tasks like the Paper Folding Test (PFT), which tests spatial imagery.

• The PFT requires mentally unfolding paper with holes punched in it to determine hole placement.

Aphantasia: Little or No Visual Imagery

• Some individuals, such as patient MX, report no mental imagery at all following brain injury.

• MX scored at the lowest level on imagery vividness tests, and fMRI showed inhibited activity in visual cortex regions.

• After MX’s case was publicized, others reported lifelong absence of imagery, known as congenital aphantasia.

• Aphantasia affects about 1–3% of the population.

• Despite low imagery vividness, individuals with aphantasia often perform normally on spatial tasks, suggesting imagery vividness and spatial reasoning rely on different processes.

Hyperphantasia: Extremely Vivid Imagery

• Hyperphantasia refers to extremely vivid and intense visual imagery.

• It is likely much rarer than aphantasia.

• Occupational trends:

  • Individuals with aphantasia are more likely to enter fields like mathematics or science.

  • Individuals with hyperphantasia are more likely to pursue creative professions.