Greg Detre
29/10/01
pre-striate � higher than V1 (= extra-striate)
retina = part of the brain
parvo/magno distinction works up until the LGN (konio = hidden (Gk)???, dust???)
distinguish by cell body size in LGN
so we can label retinal ganglion cells M/P
M (parasol) � transient, achromatic, PA, larger RF
P � both fairly homologous
K � some have circular RFs, but antagonistic inputs from short and long wavelength RF, colour opponent
highly heterogenous � in cat, With cells with large RF and very sluggish � some K cells are a bit like that
Hubel + Livingstone � thought that P/M stayed highly segregated in V1 � argued that they remained separate - wrong
actualliy much convergence, even in V1
K � blob, with P colour cells
lots of parallelism within the two streams
ODLs + visual field
off-centre-onsurround plus opposite in LGN
comes � 2 bipolar cells, one to each
both looking for edges, but opposite sides
you need that duplication because otherwise they couldn't signal local reductions if they�re already firing at low frequency, hence their complements
needs to be linear(sic???) because of huge range of intensities
2 eyes, rods + cones
10% project to the midbrain
= a large role in insectivores and non-mammals � don't have a cortical visual system at all
midbrain vis systems seems to worry mainly about (stereo) eye movements
blindsight � believed to be midbrain feeding into visual system
comes from V1 lesion
from midbraint o perirhinal + dorsal (???)
IT system � V4
lesions � achromatopsia
Holms(???) � viewed V1 as overly central � ignored parallelism + processing like motion blindness and achromatopsia etc.
distinguish between colour and wavelength
monkey cortex has cells that respond to colour, but colour constancy differentiates colour + wavelength
averaging mechanism � sample object, compare with the rest of the scene, and subtract average environment ambient light
factors out the nature of illumination � Helmholz
some cells respond to shapes and their mirror images
some are independent of viewpoint, some only from certain viewpoints
mirror images might be easy to do
cells that respond to faes � Rolls + Perez � seems respectable
hand responsive cells in monkeys too
fMRI � fusiform gyrus active when looking at faces (this month�s �Brain�) � autistic kids don't become active here
as though they don't have the right face sensitive area
�Face processing outside the fusiform� - Pierce
combingin shapes and colours (Zeki) � Monet vs Picasso
false colour art does not activate the same areas
Brain 121(???) Zeki fMRI � three(???) stages of colour processing
normal colours activated (as well as V1 � V4) elsewhere
binding � central oscillations, correlated firings of neurons across wide areas of brain
indicating that they�re processing the same object/part of the visual field
is there a binding problem?
also, even the streams are cross-connected
mnemonist � is my memory degrading or actively erased?
S/M/L
oscillation binding vs temporal synchronicity
scotoma � blind area(???)
typing vs writing
when you write, the shapes are organised at a very high level and then interpreted at a lower level in terms of motor patterns
writing signature on a blackboard or with chalk � still looks the same independent of muscle groups
typing might be organised differently
writing = formulating an output pattern at a higher level
why?? because we learn to write earlier???
keyboard = position, writing = form