1. Visual Imagery
One of the greatest problems confronting psychology is the nature of mental representation.
Part of this debate is the nature of representations of mental images.

2. Study of Imagery Banned by behaviorists
Possible subject of study in cognitive psychology
Cognitive psychology is distinguished from the earlier behaviorism by its claim that there are internal representations of knowledge on which the mind operates
However, this is a difficult area of study
Mental images are subjective
How can we show that images are used?
How are they represented?
Kosslyn and Thompson (2003, p. 723)
“Visual mental imagery occurs when a visual short-term memory (STM) representation is present but the stimulus is not actually being viewed; visual imagery is accompanied by the experience of ‘seeing with the mind’s eye’.”
Mental images are subjective; thus developing a cognitive understanding of Imagery has been fraught with difficulties.

3. How are mental images represented?
The analog vs. propositional debate
analog: the representation has the same structure as the thing represented
propositional: a sentence-like description of the image, non-spatial
Most studies mentioned seem to argue for analog representations (e.g., mental rotation, brain imaging studies). Yet mental images are not processed exactly the same as visual images

4. Images vs. Propositions
Imagine: The can is on the box. The can is black
Analog
Propositions
on( can, box )
black( can )

5. Some Questions about Mental Images
What is the relationship between imagery and perception?
How are mental images processed and transformed?

6. Imagery = perception in reverse?

7. Imagery & Perception
If the mechanisms used to perceive stimuli are also used to generate mental images, then we should predict:
Mental images should be quasy pictorial
Mental images should activate some of the brain areas involved with visual processing

8. Just as in visual images, level of detail in mental images can vary
Imagine a bee next to a rabbit
Imagine a elephant standing next to a rabbit
It is an obvious perceptual fact that when something is close up and large in the visual field, it is easy to recognize, but when it is far away and small, the task is not so easy. You have no trouble recognizing a friend standing just a few feet away, but recognizing your friend would be much harder if the two of you were at opposite ends of a football field. The investigator used this fact about perception to demonstrate that people have mental images.
Participants were asked to visualize a target object (for example, a goose) next to one of two reference objects, a fly or an elephant. Each pair of objects was to fill the frame of a participant’s mental image, and in each case the proportional size of the target object relative to that of the reference object was to be maintained. (Thus, the image of the goose would be larger when paired with the fly than when paired with the elephant.) While holding one of these two pairs of images in mind, such as goose-and-fly or goose-and-elephant, participants heard the name of a property (for example, “legs”) and had to decide as quickly as possible whether or not the target animal has that property by referring to their image; the participants were told that if the animal has the property, they should be able to find it in the image.
Participants were an average of 211 milliseconds faster to verify properties when they imagined the tar­get objects next to the fly than when next to the elephant. In a control condition, in which participants visualized enormous flies and tiny elephants next to the normal-sized animals, the results were reversed— the participants were faster when the queried animal was visualized next to a tiny elephant. So, it wasn’t the fly or elephant per se that produced the results, but rather their size relative to that of the queried animal.
The finding parallels the motivating observation, namely, that recognizing a friend is easier up close than across a football field. When a given object was imaged as relatively large (next to a fly), it was easier to process visually than when it was imaged as relatively small (next to an elephant). As the property named became larger in the image, it was easier to identify. From this result, the investigator concluded that the participants used images to answer the questions asked of them and to verify the properties named.
Does a rabbit have eyebrows?
Does a rabbit have eyebrows?

9. Shrinking mental images with brain damage
Before surgery
After surgery

10. Visual Imagery and Hemispatial Neglect
Mental images from opposite sides of an imagined public landmark
(Bisiach and Luzzatti, 1978)

11. Finke and Kosslyn (1980) experiment
fixation
dot
separation
Angle of separation
Experiment measures field of resolution: the angle of separation into the visual periphery where you cannot distinguish dots any longer

12. Fields of resolution are similar in perception and imagery
Horizontal and vertical fields of resolution in perception and imagery as a function of dot separation and vividness of imagery. Data from Finke and Kosslyn (1980).

13. Evidence from brain imaging (fMRI) for involvement of visual processing areas during visual imagery
(Le Bihan et al., 1993)

14. Mental Rotation
Can mental images be transformed in a way analagous to physical objects? How could we tell?
Mental rotation task: look at the time it takes to rotate two shapes into correspondence
Demo experiment:

15. Example Trials same different different different same different

16. Results
linear relationship between rotation and reaction time in object comparison
The mental process seems to be analogous to the physical process of rotation.

17. Mental rotation in monkeys
Stimulus: light comes on M
Movement: to light 90o CCW S M S
Direction indicated by motor cortex
Time (10 msec intervals)
Georgopoulos et al. (1989). Science, 243 (4888).,

18. Are visual images also subject to visual illusions? Ponzo Illusion
Imagine an inverted “V” over these lines. Which horizontal line is longer?
Which horizontal line is longer?

19. Differences between pictures and images
Images cannot be reinterpreted
Images = perception + interpretation
Strong influence of conceptual knowledge on imagery
Simplification of object models
Distortions in cognitive maps

20. Imagery and Ambiguous Figures
What would this object look like when rotated 90 degrees?

21. Imagery and Ambiguous Figures
If you see one interpretation, it is very difficult to then imagine the other interpretation (unless you are trained in this task)
One difference between imagery and visual perception: visual images, unlike mental images, can be easily reinterpreted
People usually cannot respond “seahorse” if using memorized image
People usually can respond “seahorse” if asked to draw image (from memory), physically rotate paper, and then reinterpret drawing
Information is present in image, but hard to reinterpret flexibly

22. Mental Images might miss important aspects of object being imagined
Imagine you have a cube between your thumb and index finger. One corner of the cube touches your thumb, and the diagonally opposite corner touches your index finger. Now, point to the locations of the rest of the corners in space.
Many people point (incorrectly) to four points on the same plane half way between the top and bottom corners.
Correct Solution:
Most people imagine that there are four remaining corners of the cube
and that all four corners are aligned on a horizontal plane parallel to the ground
BUT 6 corners remain and only 2 corners are in the horizontal plane

23. Mental distortions in Cognitive Maps

24. Which is further west: the atlantic or the pacific entrance to the panama canal?

25. Which is further south: Philadelphia or Rome?
Which is further east: Florida or Chile?
Which is further east: Reno or San Diego?

26. Cognitive maps are affected by conceptual knowledge
Relative locations of small regions is determined by a conceptualization of larger regions.
Line of reasoning:
Nevada is east of California
Reno is in Nevada, San Diego in California,
Therefore, Reno must be east of San Diego

27. Experimental evidence for hierarchical organization in cognitive maps
Ss. study maps. Later, from memory, they judge relative position of locations x and y
Performance was better when superordinate information was congruent with question
Congruent
Incongruent
(Stevens and Coupe, 1978)

28. Summary
Imagined information is processed in similar ways to perceptual information
Neuroscience evidence (fMRI)
Neuropsychological evidence
Behavioral evidence:
Kosslyn studies/ Scanning studies
Mental rotation
Some visual illusions
But there are also differences:
Images are difficult to reinterpret
Cognitive distortions in mental maps