Visual Imagery

Psy315h - Class 25

(This page last updated 25 November, 2005.)
  

How can we study imagery scientifically?

• Introspection?
  • Can provide useful insights and suggest ideas about what to study, but we need something we can measure objectively.
• Need to find circumstances in which imagery affects behavior in measurable ways.

Two Ways to Study Imagery

• Type 1: Ask subjects to manipulate their images in different ways, and try to find something objective that varies that we can measure.
  • Stephen Kosslyn
  • The longer the "distance" scanned in the image, the longer the time before the scan is completed.
• Type 2: Find tasks for which people use imagery spontaneously. Try to determine what sorts of tasks are best done with images and how they are done.
  • Roger Shepard
  • RT increases with the amount of rotation from the normal orientation.

Stimuli and Data from Experiment on Mental Rotation with Letters

One of the main questions driving research in mental imagery has been "How are mental images represented?"

Propositional Representation	Analog Representation
(proposition: something with a truth value) propositional = descriptive	Something about form of representation that matches form of thing it represents. analog = depictive
Example of Propositional Representation: cab on left window at top of cab door handle on cab below window trailer from back of cab to far right wheel at lower left of cab wheel at lower left of trailer wheel at lower right of trailer stripe across bottom of cab stripe across bottom of trailer • • •	Example of Analog Representation:

Which format is used in mental images? Propositional or Analog?

• Experiencing a mental image is like seeing a picture. Therefore, the image must be represented as a picture?
• But when you see a real picture, your visual system almost certainly builds both analog and propositional representations. Thus the subjective similarity of perception and imagery tells you little about the underlying representations.
• Kosslyn claims that imagery relies on an analog image buffer.
  • Special-purpose mechanism just for doing imagery. The picture-like properties of images are built into the representation.
  • Imagery representation like large pictorial array. When you "scan" image, you are shifting contents of this array.
  • You can only move the image across the array in small steps. That is why long scans take more time than short scans.
• Zenon Pylyshyn Disagrees
  • Pylyshyn believes that when you tell subjects to scan an image, they use tacit knowledge about the world to simulate the experience of scanning.
  • But what about mental rotation? Subjects are not told to imagine anything. Harder to see how tacit knowledge could explain mental rotation.
• Who is viewing the image buffer? Is there a homunculus?
  • Homunculus: "little man in the head"
    • Placing a homunculus in a psychological theory is a way to avoid giving a real explanation.
  • There has to be a mechanism for analyzing visual representations in normal perception.
  • Image representations can be fed into this same system.

next class: Imagery in the Brain
 
Psych 315H: Cognitive Psychology Kyle Cave Psychology Dept. U. Mass.