PSY270H1 Chapter Notes -Alarm Clock, Occipital Lobe, Parietal Lobe
Perception and Pattern Recognition
Today we will be looking at information surrounding the following question:
"How does the process of perception depend on incoming stimulation and
First, remember the difference between perception and sensation. Sensation
is what the senses do and perception is the process of recognizing,
organizing and interpreting information.
Human beings posses something called perceptual intelligence, which is
the knowledge of perception (such as previous experiences) and the
knowledge we bring to a situation (i.e., when you wake up to your alarm
clock and hit the snooze button, knowing you can still sleep for 10 minutes
and make it to class).
The central problem to perception is the question of how we attach meaning
to raw material we receive from the environment. The primary goal of
perception is to figure out what is out there and where it is.
How do is the "what"
and "where" processed
in the brain?
The "where" pathway:
spatial properties. Goes
from occipital lobe to
The "what" pathway: object
properties. Goes from
occipital lobe to temporal
How do we perceive with our eyes?
The retina receives an image of a stimulus. This activates the receptor cells
in the retina (called rods and cones) that convert light energy into neural
impulses. This information is meaningfully interpreted (it becomes the
precept). Precepts are not the same thing as the information as it is out
there and perception itself involves something more than the formation of
images on our retina (size constancy).
Bottom-Up Processing versus Top-Down Processing
(1) Bottom-Up happens through the stimulation of receptors. If these
receptors are not stimulated nothing will be perceived. Template, Prototype
and Structural theories of perception fit into this type of processing.
(2) Top-Down Processing is when we bring in our knowledge to identify and
These two processes usually collaborate in order to create perception.
Basically, incoming data from the environment can be affected by our
knowledge, whether that knowledge be new, old or just realized.
What is the rat man demonstration?
There is a rat image and a mouse image in one. At first it may look like a rat
or a mouse but as soon as someone mentions the other possibility, you begin
to see both options. This shows that recently acquired knowledge can
influence perception and that knowledge of context provided by a scene can
influence perception (consider Stephen Palmers experiment).
What is Template Matching?
This is how early cognitive psychologists studied perception. It eventually
lead to the idea of perception as based on features. Basically, the concept is
that when a perception fits a template, you recognize it as such (template for
triangle, box, cat, etc). This works well with simple situations (like cheques).
The problem however is that there are too many objects out there in order to
have a template for all. So, from there developed the theory of perception
based on features.
What is the Interactive Activation Model/Prototype
This model purports that there are feature detectors that respond to oriented
lines. This idea led to one that said that perception is constructed from
simple features of and from the environment. Which ultimately solved the
problem with template matching (requiring too many templates). But,
although this theory allows for more flexibility than template matching it is
not a complete theory. It does not explain how prototypes are formed or how
the best match is identified and selected and is not flexible enough (since
there still needs to be a prototype for each pattern).
Posner and Keele did a study in 1968 that required participants to
classify images presented. The images presented were shapes made of dots.
Each shape had a series of shapes that belonged in the same category but
that were distorted. Participants were not shown the prototype (the original
shape) and were asked to categorize the distorted shapes they saw. The
results concluded that people are naturally good at creating prototypes and
almost always got the distorted shape classified correctly.
The prototype theories got their influence from Lombroso's idea that there
is a prototypical face of crime. This lead to experiments on face prototypes.
Where researchers mixed a set amount of faces together (4-10-20-34) and
then measured how many times each face was measured more attractive.
They found that the more mixed the face was (made more average) the
more attractive it was said to be.
The Model of Letter Recognition
Created by McClelland and Donald Rumelhart. They proposed that activation
of letters is sent through three levels:
(1)Feature level: possesses feature units like straight and curved lines
(2) Letter level: has letter units like letters of the alphabet
(3) Word level: has word unites like your vocabulary.
So, according to this theory, how do we recognize single letters?
First we see the stimulus. This activates feature units then the letter units,
which filter out those letters that don't match. Then out of those letter units
that do match, the one that is activated the most is the one we perceive.
But how do we recognize a word?
It's similar to the process explained above. However, in the letter processing
stage, there is a letter unit for each letters position in a word (C1 A2 T3).
Then, there is activation of the word units that contain those positioning. The
word activated the most is the one we perceive.
What is the word superiority effect?
This concept purports that letters are easier to recognize in a word and not
alone or in a non-word. G.M. Reicher (1969) showed this effect in word
reading. It was therefore concluded that letters are processed and affected
by surrounding letters.
How do we recognize a letter in a word?
Through feedback activation which is activation that goes from word units
back to letter units (for that word). There would be an enhanced activation
for the letter you are trying to focus on (A2, for example). Which kind of
explains the word superiority effect.