RLG100Y1 Lecture Notes - Lecture 4: Optic Chiasm, Receptive Field, Gestalt Psychology

4 views8 pages
20 Apr 2012
School
Department
Course
Professor

For unlimited access to Class Notes, a Class+ subscription is required.

Lecture 4
Slide 2 hemi means half, an means doesnt work/not, opia means vision… half not vision
-half of visual field is missing relative to your fovea,
-heteronymous hemianopia means that the hemianopia doesn’t have to be in both eyes…
heteronymous stands for in one eye having visual field defect in left visual field and in other eye
its visual field defect in right visual field.. which eye has what defect?... in left eye, left part is
missing and right eye its the right side thats missing, thats why its heteronymous (meaning its
different).. homonymous means both eyes the same side would be defected…
-the fact that its heteronynmous means its indicative of which part of the primary visual pathway
is disrupted… optic chiasm is whats disrupted (part where path of visual info is crossing over)…
what eyes see is left and right visual field…both eyes see both visual fields more or less…
-what happens after optic chiasm? Visual fields are being sorted, whatever is in left visual field
ends up in right hemisphere and vice versa… when this crossing is disrupted, it is the optic
chiasm that is missing/lesioned
-looks like an inverted U, the contrast sensitivity function, horizontal axis is spatial frequencies
e.g. cycles per degree…
-on vertical axis is the sensitivity contrast (not a physical measure) but rather the ability of a
person to perceive it…lower spatial frequencies (lower cpd) and lower sensitivity.. if SF gets too
high we cant resolve it because our photoreceptors dont have receptive fields small enough so
that’s why it goes down after a point…
-exam will not cover lecture 6
Slide 4 important distinction b/w recognition and perception which explain diff b/w middle
vision and high level vision… early level vision is about breaking objects apart…
Slide 6 point of this slide is object recognition, we have pixels on a screen and we assign the
arrangement in space of these pixels to a certain class of object (house)… matching up visual
input with memory…special kind of object recognition is object identification, panel C has same
house as panel A so can identify it to be same object as panel A even though its rotated view
-object naming = word you can assign to perceptual impression of what you see
-object perception = steps that lead to object recognition…needs to happen at some point, need to
put together the visual input so that an object becomes a coherent thing, not a given, at the
Unlock document

This preview shows pages 1-3 of the document.
Unlock all 8 pages and 3 million more documents.

Already have an account? Log in
moment we are at area v1 where things are atomistic, everything is broken down to individual
lines/edges, nothing coherent yet… object perception is process that will bring all these lines
together…
Slide 8 gestalt laws = visual rules that system adheres to help us see coherently…
-perceptual committees, gestalt laws have certain rules of extracting info from visual input,
certain committees sort out what info is correct
-ground perception goes back to gestalt laws as well, about identifying an object as something
that im looking at as something in the foreground vs the background…
-global images like forest and trees, joining visual info together to create the global images…
Slide 9 even computers cant simulate our visual system… can perform the algorithm for
luminance and different pixel contrasts… very simplistic, still pixels, not lines… at this point in
visual system , V1 does this for us… neurons respond to local patches of edginess and at least
locally, we can identify something of a line… the smooth transition on left pic, the gray level is
identical so you still have gaps b/c of the exact same shade, no contrasts for computer to detect
Slide 10 circles in panel b show neurons from v1 in visual field that analyze those orientation
sensitivities those circles are just 3 individual neurons but in reality there are way more, panel
b is to illustrate that the info from that house is very disjointed… panel c, v1 itself has no way of
telling the difference between the snowman and house… most of its receptive field shows a
vertical line and ignores the squiggle from the snowman… same case with the car… these are
problems that v1 faces… cant just join together lines in order to perceive… structuralist idea was
just bottom up, identify lines, and then join them and finally you have a house (does not work)
Slide 11 gestalts pointed this out, gestalt means whole, the gestalt psychology major statement
is that the whole is greater than sum of its parts… gestalt psychologist most imp contribution is
formulating set of gestalt laws that help middle vision to put things together…
Slide 12 panel a:N1 and N2 could be grouped together (high degree of co-linearity)… lines
after that become less and less collinear (less smooth)… panel b: you still see a loop but area 1
Unlock document

This preview shows pages 1-3 of the document.
Unlock all 8 pages and 3 million more documents.

Already have an account? Log in
cannot perceive this loop… need something else, gestalt law helps us join these lines
together…have some kind of neuron that has receptive field on one of these lines with a certain
orientation tuning… wont be as interested in a line with a totally different orientation… because
of the location of n1 and n2 receptive fields, if they respond together, we have two collinear line
segments.. neurons have on avg 10 000 connections with other neurons, so probably a
connection between n1 and n2 neurons… if this part of v1 is really good at gestalt law, there will
be a positive connection b/w these two neurons, if n1 sees a line with this kind of orientation it
should encourage n2 to see that kind of line as well.. n2 probably does the same thing and
activates n1… both encourage each other to look for the line.. mutually enhance each other…
they will also bully any other neuron responsive to lines that are not collinear e.g. n3… would
inhibit n3…have to have this balance of power b/w inhibition and activation.. balance of power
is brought into imbalance b/c of collinearity, so that tips balance of power in one way.. making
certain lines more salient
-so gestalt law can even be implemented in V1 through lateral connections…
Slide 13 what people perceived as being smooth versus not smooth is simply a breaking
point… slide b is the better continuation of the two lines, this way the lines are smoother... the
point you perceive smooth versus not smooth is what you find in outside world… lives in trees
curve around in a certain extent, if the curve is too disrupted were probably looking at two twigs
overlapping.. if you interpret C as the lines having two turning points, thats quite a coincidence,
visual system hates coincidences and tends to dismiss them..
Slide 14 gaps = no neuron for that part of visual field e.g. the blind spot.. so its important to fill
in those gaps, mechanisms in visual system do this job… kanizsa was Italian psychologist who
did a lot of work on this… arrow is clearly visible to each of us but there is actually no arrow
drawn there…
-illusory contours… the arrow looks even whiter than the background but we see it as
brighter…tend to perceive foreground as a little lighter, all these things made up visual system
b/c it occurs in normal life… visual interpretation of line end stoppings is that there must be an
occluding contour, cant be a coincidence, panel b is just line stoppings but you see something
circular… visual input is always ambiguous but this lack of information is compensated by the
fact that coincidences are very unlikely…panel C, one lines goes like a finger coming from
above and one from below, the different colours create this…colors overpower law of good
continuation and is an example of gestalt law of similarity…
Unlock document

This preview shows pages 1-3 of the document.
Unlock all 8 pages and 3 million more documents.

Already have an account? Log in

Get access

Grade+
$10 USD/m
Billed $120 USD annually
Homework Help
Class Notes
Textbook Notes
40 Verified Answers
Study Guides
1 Booster Class
Class+
$8 USD/m
Billed $96 USD annually
Homework Help
Class Notes
Textbook Notes
30 Verified Answers
Study Guides
1 Booster Class