10/14/13

paraderollAI and Robotics

Nov 17, 2013 (3 years and 8 months ago)

125 views

This Week


Email materials to me by Tuesday, 10/15


Excel spreadsheet (plus 4 practice trials) and stimuli


Regular class on Wednesday, 10/16


Thursday, 10/17


Meet in Mac computer lab (HUM 304)


Test out your experiment


Collect some data


Friday, 10/18


Meet in Mac computer lab (HUM 304)


Collect more data


Test 2, Wednesday, 10/23


Material since Test 1


Similar to Test 1


Multiple choice; fill
-
in
-
the
-
blank; short answer;
long answer


Important concepts, Wednesday, 10/16


Long answer questions, Friday, 10/18


Face Recognition


Faces are special


Face
-
inversion effect


Fusiform face area


Prosopagnosia

Face Recognition


Faces are not special; effects are due to
expertise

Gauthier and Tarr (1997)


Used non
-
face stimuli (Greebles)

Greebles

Ribu

Zoti


Gauthier and Tarr (1997)


Used non
-
face stimuli (Greebles)


Subjects trained to recognize Greebles


Novices


Experts


Do subjects exhibit a “Greeble”
-
inversion
effect?

30
40
50
60
70
80
90
Novices
Experts
Upright
Inverted
Subject Group

Percent Correct

Gauthier and Tarr (1997)

Gauthier and Tarr (1997)


Used non
-
face stimuli (Greebles)


Subjects trained to recognize Greebles


Novices


Experts


Do subjects exhibit a “Greeble”
-
inversion
effect?


Experts do


Novices do not

Face Recognition


Faces are not special; effects are due to
expertise


“Greeble”
-
inversion effect in experts

Gauthier et al. (1999)


fMRI study


Subjects trained to recognize Greebles


Tested on upright and inverted Greebles at
different stages of training


Does activity in fusiform face area change
with training?

5
10
15
20
25
1-2
3-4
5-6
Training Session

Upright
-

Inverted

Gauthier et al. (1999)

Gauthier et al. (1999)


fMRI study


Subjects trained to recognize Greebles


Tested on upright and inverted Greebles at
different stages of training


Does activity in fusiform face area change
with training?


Becomes more sensitive to inversion with
increased expertise

Face Recognition


Faces are not special; effects are due to
expertise


“Greeble”
-
inversion effect in experts


Fusiform face area becomes sensitive to
inversion of Greebles with increased expertise

Gauthier et al. (2000)


fMRI study


Bird experts and car experts


Judged whether birds were from same
species


Judged whether cars were different years of
the same model


Examined activity in fusiform face area

0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Bird Experts
Car Experts
Birds
Cars
Subject Group

BOLD Signal Change (%)

Gauthier et al. (2000)

Gauthier et al. (2000)


fMRI study


Bird experts and car experts


Judged whether birds were from same
species


Judged whether cars were different years of
the same model


Examined activity in fusiform face area


Greater in bird experts for birds


Greater in car experts for cars

Face Recognition


Faces are not special; effects are due to
expertise


“Greeble”
-
inversion effect in experts


Fusiform face area becomes sensitive to
inversion of Greebles with increased expertise


Experts rely on fusiform face area for expert
domain

Prosopagnosia


Bird
expert
no longer able to identify
birds at
expert level
(Bornstein, 1963)


Car
expert
no longer able to identify cars
at
expert level (
Lhermitte

et al., 1972)

Face Recognition


Faces are not special; effects are due to
expertise


“Greeble”
-
inversion effect in experts


Fusiform face area becomes sensitive to
inversion of Greebles with increased expertise


Experts rely on fusiform face area for expert
domain


Experts with prosopagnosia lose expertise

Busigny

et al. (2010)


Patient P.S.


Closed
-
head injury


Extensive lesions throughout occipital and posterior
temporal lobes


Prosopagnosia


Difficulty recognizing faces


No other object recognition deficits


Two alternative forced
-
choice sequential
matching


Busigny

et al. (2010)


Busigny

et al. (2010)


Busigny

et al. (2010)


Busigny

et al. (2010)


Busigny

et al. (2010)


Patient P.S.


Closed
-
head injury


Extensive lesions throughout occipital and posterior
temporal lobes


Prosopagnosia


Difficulty recognizing faces


No other object recognition deficits


Two alternative forced
-
choice sequential
matching


Normal performance with cars


Special or Not Special???


“We learn to distinguish faces to a degree not seen with
other categories because facial recognition is from the
very earliest age and throughout life such an essential
and determining aspect of daily living. Thus, it can be
argued that we might have acquired the same
perceptual skill in relation to the configuration of one tree
relative to the next, if tree configuration were as major a
determinant of behavior as is faces. […] Facial
recognition becomes by far the most complex and
frequently encountered example of relatively pure visual
discrimination learning that occurs in everyday life.
Considered in this way it becomes less surprising that it
may be disturbed in relative isolation.”
(Meadows, 1974)

Mur et al. (2012)


Does fusiform face area exhibit categorical
activity?


In general, activity is greater in the fusiform for
faces compared to non
-
faces


Does fusiform face area exhibit graded
activity?


Certain faces and even some non
-
faces evoke
more activity than other faces

Mur et al. (2012)


fMRI study


Present 96 images


Faces (human, non
-
human)


Bodies (human, non
-
human)


Objects (fruits, vegetables, tools)


Places (landscapes, architecture)


Rank order images based on BOLD signal
change


Fusiform face area

Fusiform Face Area

Face

Body

Object

Place

BOLD signal change (%)

Fusiform Face Area

Mur et al. (2012)


Does fusiform face area exhibit categorical
activity?


In general, activity is greater in the fusiform for
faces compared to non
-
faces


Yes


Does fusiform face area exhibit graded
activity?


Certain faces and even some non
-
faces evoke
more activity than other
faces


Yes


Same vs. Other Race Faces


Greater activity in fusiform face area for same
compared to different race faces (
Golby

et
al
., 2001
;
Feng

et al
., 2011)





Golby

et al. (2001)

Mur et al. (2012)


fMRI study


Present 96 images


Faces (human, non
-
human)


Bodies (human, non
-
human)


Objects (fruits, vegetables, tools)


Places (landscapes, architecture)


Rank order images based on BOLD signal
change


Fusiform face area


Parahippocampal

place area

Parahippocampus

Parahippocampus

Parahippocampus

Fusiform

Parahippocampus

Fusiform Cortex

Fusiform

Parahippocampal

Place Area

Face

Body

Object

Place

BOLD signal change (%)

Parahippocampal

Place Area

Mur et al. (2012)


fMRI study


Present 96 images


Faces (human, non
-
human)


Bodies (human, non
-
human)


Objects (fruits, vegetables, tools)


Places (landscapes, architecture)


Rank order images based on BOLD signal
change


Fusiform face area


Parahippocampal

place area


Early visual cortex

Early Visual Cortex

Face

Body

Object

Place

BOLD signal change (%)

Early Visual Cortex