Cognitive Science > Action and Cognition > Questions/3

Jump to: navigation, search

Contents

[edit] Pathways of the visual system

Explain the major division of the visual system into a ventral and a dorsal pathway. What is the experimental evidence? Name at least two areas of each pathway.

[edit] Major division after V1

  • Dorsal stream
    dorsal: at the back
    • Main input is magnocellular pathway from LGN to V1
    • Not color-sensitive
    • Not detailed
    • Quick
    • Sensitive to movement
    • Processes "where" information
  • Ventral stream
    ventral: front
    • Main input is parvocellular pathway
    • Color information
    • Detailed
    • Object recognition
    • Processes "what" information"

[edit] Experimental evidence

  • Perception of movement is independent of perception of color
  • Shape discrimination task: Recognize new object to find raisin
  • Landmark task: Recognize object closer to cylinder to find raisin
  • Monkey with lesion in temporal lobe ("what" lesion) cannot do shape discrimination task but is able to perform landmark task
  • For monkey with lesion in parietal lobe ("where" lesion) it is the other way round
  • Double dissociation of object recognition and position memory

[edit] Areas belonging to pathways

  • Dorsal stream: Layer 4b of V1, thick stripe regions of V2, ...
  • Ventral stream: Blob + interblob regions of V1, thin stripes and interstripes of V2, ...

[edit] Latency of a brain area

How is a "latency" characteristic of an area measured?

  • Measure response time from stimulus onset for several neurons of area
  • Cumulative distribution function gives percentage of neurons that are active after a certain time

[edit] Response latency in V1

What is the order of magnitude of response latency in primary visual cortex?

  • 50% of neurons have a latency of 62ms or shorter
  • Fastest: ~50ms
  • Slowest: ~70ms

[edit] Latency vs position in diagrams of visual areas

Compare the arrangement of cortical areas in the typical van Essen/Felleman diagram and the size of the latencies of the respective areas.

  • General principle: Areas higher in hierarchy are those that receive input mainly from output fibers of other (lower) areas
  • For areas at the bottom of the diagram, the latencies follow each other sequentially.
    Regular sequence: retina -> LGN -> V1
  • Beyond V1, most areas are concurrently active, no clear correlation between latency sequence and hierarchy.
  • Diagram of anatomy, not topography
  • Some areas have multiple connections (ill-defined position in hierarchy)

[edit] Cytochrome oxidase staining of V2

Describe the structures revealed by a cytochrome oxidase staining of secondary visual cortex in monkeys.

[edit] Thick stripes

  • Sensitive to orientation, direction and binocular disparity
  • Code depth and motion
  • High CO concentration

[edit] Thin stripes

  • Sensitive to wavelength (color, illumination)
  • High CO concentration

[edit] Interstripe regions

  • Low CO concentration
  • Similar to V1 interblobs
  • Sensitive to orientation, not to wavelength

[edit] Connectivity of V1 and V2

Give a sketch of the connectivity of layer 4 of V1 and its subdivisions with V2.

  • Layers 3-5 of LGN -> layer 4Cbeta of V1 -> blobs in V1 -> thin stripes in V2
  • Layers 3-5 of LGN -> layer 4Cbeta of V1 -> interblobs of V1 -> interstripe regions in V2, V4
  • Layers 1 and 2 of LGN -> Layer 4Calpha of V1 -> Layer 4B in V1 -> Thick stripes in V2, V3, MT

[edit] Non-classical receptive fields

Give an example of a non-classical receptive field effect.

  • Classical: Activity is fully determined by absolute position of stimulus
  • Example for nonclassical: Response of neurons in V2 influenced by additional properties
  • Sensitivity to border-ownership, close-far-occlusion (stronger response to stimuli closer than occluded object)
  • Response tuned to real-world setups

[edit] Reverse correlation

Explain the reverse correlation technique.

  • Technique to determine response properties of receptive fields.
  • Find out which properties of stimulus elicit response without precise knowledge of which stimulus to test.
  • 1. Show a wide range of different stimuli (e.g. using white noise) to subject.
  • 2. Group all those stimuli that yield the same neural response.
  • 3. Average over history of stimuli with same response (look at stimulus a few ms before response) -> average stimulus that created a certain response
  • Assumption: Linearity of neural response
Source: http://wiki.stuhlmueller.info/Cognitive_Science/Action_and_Cognition/Questions/3