Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
64 Cards in this Set
- Front
- Back
|
Young Hay Ellis 85
|
22 ppl
diaries errors miss recog fam only couldn't place couldn't name |
|
|
Bruce and Young 86 model of face recognition
|
see
|
|
|
Types of encoding
|
pic
structural visually derived ID specific expression name facial speech |
|
|
young 86
|
face vs name for naming fam occuapt
face= fame fastest, name slowest name= name fastest, occupation slowest |
|
|
Brennan 90
|
TOTT state :
rpt quest who last US president show pic give initials - most resolved Supports Bruce and Young prediction |
|
|
Hanley 98
|
face and voice
fam only unfam named fam and occup (not name) |
|
|
voice recog
|
more fam only
VRU harder to activate? Give fave and voice = more recog More routes to PIN activation |
|
|
Hanley & Turner 2000
|
Faces recog easier or pref access PIN/NPU/SIU
blurred faces = voices faces easier to recog not pref access |
|
|
Moore & valentine 98
|
Tinme to name famous face= affected by fam and A0A and frequency (not D or name length
(poor method) rated fam so may be biased by currency of fame |
|
|
Lewis 99
|
soap faces
Ctrld freq AoA & Currency Current, f and A0A early = faster did soap from AOA + freq = familiarity |
|
|
Burton Bruce and Johnson 96
|
IAC
|
|
|
Burton, Johnston & ellis
|
IAC accounts for prosopagnosia
|
|
|
FRU
|
match stored reps to fam faces
|
|
|
PIN
|
detect familiarity
|
|
|
SIU
|
semantic units containing info
|
|
|
Links
|
can be inhibited or excitatory
|
|
|
Threshold
|
If threshold reached= activation
|
|
|
Prosopagnosia
|
deficit= partial disconnection of FRU and PINS
|
|
|
Prosopagnosia
|
face neither familiar or has semantic info
|
|
|
Morrison 2001
|
POR patient : provoked overt recognition
|
|
|
POR
|
patient given faces sharing an occupation, can ID occupation and name
|
|
|
Morrison 2001
|
prosopagnosics, 8 faces, common occupation, stimulation= below threshold, above pins activation
|
|
|
Pins to SIU
|
multiple pins (8 faces); increased threshold for SIUs to activate
|
|
|
Morrison 2001
|
attentional mechanism needed to retrieve names
|
|
|
attention
|
do prosop have attention inhib at SIUs?
|
|
|
SIU-PINS
|
if enough stim in SIUS from individual pins, - reaction should increase the threshold of the PINS (reactivation) therefore SIUs increase threshold and name retrieved
|
|
|
Stone & Valentine 2003
|
Morrison 2003 - is person recog not face recog!
|
|
|
attention
|
not ususally specified in IAC model
|
|
|
POR
|
must be an increase in activation in PIN; results in right face being recognised
|
|
|
Sergent & Signoret 92
|
subjective recognitoin requires not just the activation of related sets of info but connections between sets
|
|
|
Valentine and stone
|
high activation to lots of faces not just one face ; pin must be linked to one face
|
|
|
Diamond 94
|
prosopagnosics = above chance for force choice test selecting names for faces - couldn't do if face given only
|
|
|
Sergent & Signoret 92
|
matching task faces to name not successful; therefore not clear how stimulation fx recog- can't recall all 8 names
|
|
|
Valentine and stone
|
attention only in context of POR - they find it incomprehendable; impact on other components not considered
|
|
|
Morrison 2001
|
attention prompted by expter; not spontaeous - must be due to attention!
|
|
|
Stone & Valentine 2003
|
attention not neccesary
|
|
|
Stone & Valentine 2003
|
if SIU activated proso would act spontaneously state name
|
|
|
dehann 91
|
study-spontaneous w prosopag and face recog
|
|
|
Diamond 94
|
patientnot allowed to be spontaneous - problem w method
|
|
|
Valentine & Stone 2003
|
If Threshold reached and feedback to pins ; pins inhibit within PINS -not activated before - presentation of each face - one by one increased activation of pins; therefore PINS activated by one face
|
|
|
Prosos able to face recog
|
generating shared occupation provides sufficient activation to raise the SIUs above threshold
|
|
|
Learning new faces
|
Jets and sharks
|
|
|
McClelland and Rumelhart 98
|
IAC-I (interactive)
|
|
|
NOU
|
name output units
|
|
|
NOU explains
|
time taken for familiarity, occupation, name, decisions, tot, covert recognition, semantic repetition priming
|
|
|
Associative and repetition priming in IAC
|
charles and diana
|
|
|
SIU linked to Diana's PIN
|
Therefore quicker to recog w associative priming (latent activation
|
|
|
Bruce & Valentine
|
Face primes for face, name primes for name, no cross priming with later recog
|
|
|
Burton 90
|
repetition priming like synapitic plasticity
|
|
|
Hebbian IAC
|
small activation at ends of links - link only strengthened abit
|
|
|
Hebbian IAC
|
Strengthening increased until at max excitation level
|
|
|
Hebbian IAC
|
if one end not excited- hebbian- stops activation therefore false recognition
|
|
|
Hebbian IAC
|
Inhibition proportionate to activation - eventually reaches full inhibitioin -1
|
|
|
Stevenage and lewis 2002
|
IACAPA- computer to use FRU, PINS and NOUS when represented w new face
|
|
|
Stevenage and lewis 2002
|
recognition task acquisition data over time of new faces - twins lizzie or rosie - compared w computerised IAC ; was able to aquire twin names sigmoidal curve
|
|
|
Bredart 95
|
names and semantic in sep pools
|
|
|
Burton 94
|
indicated that hebian synapse should be used for learning
|
|
|
see pic
|
|
|
|
Ellis, Burton, Young & Flude 97
|
Strengthening at FIU and FRU and FRU - PINS = help face recognition priming
|
|
|
Present all features
|
prime whole face
|
|
|
Rumelhart Hilton Williams 86
|
Golbal modification of weights in network therefore unale to recog error
|
|
|
Lewis 2002
|
golbal modification of weights in network difficult to apply to ICA
|
|
|
Sergent & Signoret 92
|
in these studies prosos may have may have recog individual faces as occupation links were made;
|
|
|
Sergent & Poncet 90
|
didn't name occupation unprompted - delay btw recall & time to recall , but shared occupation activations; may have increased SIU above threshold=recog
|
