Lecture 4 - 01-28-12
From last week....
The Visual System
periphery of retina
There is a blind spot. Our brain fills in the blank perceptually.
A series of neurons communicates information from the retina to the cortex
in the eye:
Ganglion cells and optic nerve (goes from eyeball, where the blind spot is, and carries
information to different parts of the brain)
V1, primary visual cortex There's some neural activity in the individual system in that visual system diagram. One's a
bright stimulus and one is dull. It can excite or not excite. If we think of two images coming in,
cell B has information coming in from both cell A and C where cell C is only getting information
from cell B, cell D would not have information coming in. This is what elicits one of our optical
illusions i.e. Mach bands, lateral inhibition between neighbouring cells and the retina. Because
the two pieces of information are coming into your visual stream, there's some inhibition going
on in between your cells, it is trying to find the edges to make the perception of the white band
look darker than it actually is. The other example, the checkerboard with the cylinder, square B
is the same level of darkness as A. There's also a shadow coming from the cylinder that
compensates for the darkness as well.
the frequency of the action potential are recorded as different stimuli presented to a subject.
Researchers can manipulate what's being shown and record how they fire. What makes them
fire? There are studies involving animals, there are certain cells firing in the brain. There's this receptive field. The main thing is that in the RF, there's a center and surround.
When something is stimulated in the centre, the neural firing frequency increases from normal.
When something is in the surround, it actually inhibits/decreases from normal firing. How is this
The receptive field of the V1, they are related to line orientation. These kind of cells are edge
detectors. Flat, no stimulus, when it's an actual angle, there's more activity, when it's upright,
there's lot of activity.
Different neurons in V1 are specialized, resulting in parallel processing, not serial processing.
Parallel processing in the visual pathway:
P (parvocellular) Cells
M (magnocellular) cells
The what system:
Identification of objects
The where system:
locations of objects and guiding our responses
Problems with reaching for seen objects
Binding problem is how do we recognize this as one single unit when you have so much
information coming in from many areas. While things are firing about one particular object, it'll
also be firing about the movement. Our brain interprets that synchrony as coming from one
object, this is the neural synchrony. The systems themselves, the spatial position, the visual
areas that are processing these features - like shape and colour - know that spatial position of
the object. You can think of the coordinates on your GPS. Attention is also critical for the binding
of visual features. When attention is overloaded, people will make conjunction errors. E.g.
recognizing the colour and letter while concentrating on other tasks, you may mix up the
colours, that's the idea of overloading the visual stream (attention is overloaded), this is the idea
of conjunction errors. ***
different objects, different recognition systems
top down influences on object recognition
The words are spelled wrong but we are able to recognize what the words are.
why is this important?
If we couldn't recognize objects, we are incapable of moving on in our daily lives, recognize the
toothbrush, keys, etc. Sometimes it's hard to imagine what object recognition is because it
comes natural to us. But if we think about when we first saw something that was brand new or
something that was outside your domain of expertise; i.e. going to Lowe's/Home Depot, you
have no clue what it does; if you used the tool, that would look to you just like a table or
whatever, you got used to seeing those objects. It's a valuable part of our daily lives and is
crucial for learning. This is especially important for recognizing people; friend or foe, family.
Beyond the information given
form perception: shape and size
object recognition: ID
The whole is greater than its parts:
Knowledge about something gives us more information about an object more than the simple
visual features that we see. E.g. Necker Cube
If we look at the cube at the top, you can either see cube A/B. Only one can be seen at a time,
and you can pop back and forth
E.g. the vase/faces
You either see two faces looking at each other or a white vase. Your knowledge helps interpret
which you see.
What you see is going beyond the information because it is only lines.
E.g. skull/lady in mirror
People resolve ambiguity in everyday situations. In the fruit basket image, there's an apple
behind a banana.
Your ability to interpret scenes that are ambiguous are governed by a few principles
there's red dots/greyish pink dots, we tend to view this as dots in a column because the colours
rather than 15 dots on a screen, people will say that there are three groups of dots. We connect
the dots closest in proximity as being in one group
3) good continuation
if something seems like it's being interrupted by another object, we think of simple terms of
we look at the triangle with big gap in the left and right side. We can look at it and it feels like a
triangle to you, it's one image instead of two.
We tend to think about it as two pieces of wood placed on top of the other, we take the
simplest explanation and we perceive it without any effort or thought The Mona Lisa example
We ID the head as being something different than the outer object. Alternatively, the actual
pattern is influencing us in other ways.
The circle/loop example
there's actually 5 separate circles. Because of the shape and pattern, we're taking it as one
whole spiral and it takes effort to see the separate circles
The X/V example
we take the simplest explanation possible. We avoid interpretations like coincidences or bringing
in extra theories. We tend to see it as an X or two lines crossing. It's a coincidence that there is
precision for it at the junctions to be two Vs
The LIFT/black background example
People usually see the BG as the foreground, they don't see the white necessarily, as soon as
you ID as the foreground, you see that there are letters there. Until you had the basic
information feed in, you have some knowledge and it changes the way you see what's on the
The OPTICAL ILLUSION/field example
You can't go back once you seen it; as you begin to apply knowledge to it, top down processing,
it's always available to you. The next time you see it, you're not interpreting from bottom up
anymore, it becomes automatic.
PERCEPTION word with shadows
It doesn't really have information to inform you of the letter structure. We interpret the dark
lines around it as being a shadow which we perceive as an edge. We consider them as one
object. It gives us the sense of a closure as well.
Brain areas for basic visual feature brain areas for large scale form
form perception, the process through which the basic shape and size of an object are seen
object recognition, the process through which the object is identified E.g. The "Dolphin" example
top down information is influencing what we see. We've probably been exposed to situations
where a loving couple embraces from arts, real life, school, TV., media, etc. If you show this to
children, they see nothing except dolphins. The idea is that this is a good visual information to
be shown that once you have a certain external information, it's hard not to see that.
E.g. THE CAT (example from the textbook)
H and the A are the exact same shape and we say that in a context of what's there, we don't
have an issue of seeing the first on