L4 module 1 Flashcards
what does the Dorsal and Ventral Visual Stream process
processes information about object identity
processes information about object location
is responsible for executing movements under visual control
Dorsal = location of an object relative to your body position
Ventral = what an object is
describe the ventral stream
Ventral = processing that is going on in the parietal cortex (posterior). It is information that travels from primary visual cortex into the inferotemporal cortex
what does the medial aspect of the temporal lobe contain (what structures)
The temporal lobe (medial surface) contains the hippocampal cortex perianal cortex and enterinal cortex
what would you see in a Single neuron recording from inferior temporal cortex
When you look at cells in the inferior temporal cortex you get cells that response to specific things
When looking at the side view of a face the firing rate is highest (eg looking at an ear)
In these recording sessions they found other cells that responded to other things (other face orientations)
what did the jennifer aniston cells show
The found that cells responded to very specific images
These cells responded to jen, however all of these images are quite different
Therefore the cell is not responding to a template (eg the distance between the eyes) but to recognition of the person. these cells only responded to images of jen therefore are very specific
what did halle berry cells show
These cells even react to drawings of the person. She also played cat women. These cells responded even when her face was covered my the mask but the word cat woman did not produce a response but the work halle berry did produce a response
what is emotoions role in jennifer aniston cells
Her with other people produced less of a response
Therefore it is possible that an emotoional response that is modulation it
what causes cells in the medial temporal lobe (eg jennifer aniston cells) to respond
People that you havent seen before it will be hard to find a cell that responds to that image
Family and people that you see in the context of the experiment will produce a greater response
Therefore there is some sort of familiarity effect
what is invariance
when the cell will respond even when there are changes in the image being presented
technical definition - Encoding a representation so that it is identified regardless of size,
orientation, colour etc
what is the feature detection theory
this has to do with grandmother/gnostic cells
There is a hierarchy of processing that starts in the PVC (and before) were there are cells that respond to edges and then you have cells that respond to edges in different orientations, then there is a cell at the top what puts all the signals (edges) together to say this is a table. This is a gnostic cells (nostic = knowing) also called grandmother cells
what do grandmother cells mechanism allow for
excellent discrimination between objects
what are the problems with the grandmother hypothesis
1) requires huge number of gnostic units
- You would need to have millions of neurons to represent all of the different things that you are able to recognise
2) susceptible to damage
- If that one cell was damaged then you might be able to recognise one table but not another
3) how to you perceive novel objects?
- You would have to have cells available for when you learn something new which would be metabolically very costly
4) In the Quiroga et al study what is the probability of
finding the one ‘grandmother’ (Aniston) cell out of the millions available with just 100 or so images?
All of this suggests that in the extreme version of the grandmother hypothesis is very unlikely (one cell is responsible for one thing)
what is the Ensemble, population or distributed coding theory
there is NO one “grandmother cell”
this is the hypothesis that there is a group of cells firing to make up an image eg glasses and another wrinkles. then the pattern in which the fire and the network make up the representation of your grandmother
This network could be unique for your grandmother but different to another person’s grandmother
This is more likely to finding a cell that responded to your grandmother
what is the most important feature for recognition in the Ensemble, population or distributed coding hypothesis
The pattern of firing across the whole ensemble is
important for recognising the difference between your grandmother and someone else grandmother
Grandmother cell and Distributed encoding representaions
are at each end of a continuum of ‘sparseness’
what does this mean
the range of activation when responding to an image
with Distributed encoding the representation in the network of neurons will be fully distributed
whereas in the grandmother hypothesis it would be sparse
therefore they are at opposite ends of a continuum
