Papers

Question 1

Attention: Moore and Armstrong

Answer 1

Moore & Armstrong
Selective gating of visual signals by microstimulation of frontal cortex

• Visual signals enhances with attention
• We don’t know the mechanism that causes this enhancement
• Suggestion: eye movement commands cause enhancement/bias
• Studied interaction between saccade preparation and visual coding
  
  • By electrically stimulating sites in the frontal eye field (FEF) and measuring its effect on activity in neurons in the extrastriate cortex
  • Show that visual responses in V4 could be enhanced after FEF stimulation to evoke saccade
    —> Visual signals are modulated by strength of spatially corresponding eye movement commands
• FEF stimulation evokes saccade movement

1. Find receptive field for V4 neuron with electrode.
2. With second electrode, locate FEF site where correlated saccade will be evoked.
3. Mapped point in space that monkey’s gaze has shifted, which could fall within or outside of V4 receptive field.
4. Also stimulated after appearance of visual stimuli to see the effects of FEF stimulation for visually-driven activity (amplify/null/interfere)

• Sub-threshold stimulation does not evoke saccades, but does bias the selection of eye movements and can improve a monkey’s ability to covertly filter visual stimuli
• FEF stimulation caused a significant enhancement of v4 firing rate when there was a preferred visual stimulus in its RF
• FEF stimulation had no effect when cell was not driven by RF stimulus
• FEF stimulation caused moderate enhancement of the v4 firing rate when stimulus presented at non-preferred orientation
• The largest response enhancement occurs when the preferred stimulus is presented in V4 receptive field with a distractor outside
• V4 suppression occurred when preferred stimulus was present in receptive field but distractor was present and microsimulation was targeted outside V4 receptive field
• Enhancement depended on whether preferred or non-preferred stimulus appeared in the RF (preferred yielding greater enhancement)
• —> Enhancement grew with increasing visual drive
• Biggest effect: when preferred stimulus was presented to RF with distractor stimulus
• RESULTS:
• microstimulation of FEF sites appears to have activated a network that controls the gain of visually driven signals
• FEF microstimulation activates a network that controls visually-driven signals; stimulation biases eye movement selection and strength of visual cortical signals, revealing a common network for visual and oculomotor selection

Question 2

Object Recognition: Haxby

Answer 2

Why is our brain capable of generating a near unlimited number of distinct representations of objects and faces?

Are representations of different object categories distributed throughout the ventral temporal lobe or localized in specific regions?

1. Ventral temporal cortex contains a limited number of areas that are specialized for representing specific categories of stimuli (FFA, PPA)
2. Different areas in ventral temporal cortex are specialized for different types of perceptual processes (FFA is specialized for visual representation of individual exemplars from any object category, not just faces)
3. Face and object recognition is widely distributed and overlapping  “object form topography” 

[single cell recordings, fMRI, hypercolumns]

Logic: If distinct representations for different categories of objects are widely distributed in ventral temporal cortex, then activity across this region to objects within the same category should be more highly correlated than activity for objects between different categories.

Measured with fMRI — patients viewed faces, cats, man-made objects (house, chair, scissors, shoe, bottle), and nonsense image
- Voxels that responded differentially to categories

• Two sets of data: even and odd runs
• Indices of similarity: Difference between pattern of A odd run, A even run (within category)…compared to difference between A odd run, B even run (between category)

Are WITHIN category correlations greater than BETWEEN category correlations? Yes=Accurate

Distinct patterns of neural response for multiple categories of objects.

Are within-category correlations driven by specific regions that are maximally responsive?
Sol: Take out maximally responsive regions

Are low-level features driving results? Concept/representation driving, not shapes! [Done by correlations between photographs and line drawings]

IN SUMMARY:

• Arguing that in ventral temporal cortex, there are distinct and distinguished patterns for visual representations of different categorizes of objects
• Not being driven by FFA and PPA, even within FFA and PPA they represent same activity for other objects, also not driven by low-level features

IMPLICATIONS: Object representations may be based on
continuous dimensions (e.g., configurations of
features) instead of structural primitives

Question 3

Perception: Tong

Answer 3

Tong Paper

• Used fMRI to monitor stimulus-selective responses of FFA and PPA during binocular rivalry
• Face and house stimulus presented to different eyes
• Constant retinal stimulation, perceived changes correlated to increased PPA/FFA activation
• Responses during rivalry were equal in magnitude to those evoked by nonrivalrous stimulation
• Suggests that PPA/FFA activity reflects perceived stimulus, not retinal stimulus
• Rivalry resolved by stages of visual processing

3 fMRI Scans

1. Localizer scans — nonrivalrous images showed to locate PPA and FFA
2. Rivalry — face to one eye, house to another, while subject maintained central fixation
3. Non-rival using same temporal sequence reported

IN SUMMARY:

• Both stimuli present in both situations (both faces and places)
• But consciously aware of only one at a time
• When aware of face, increased FFA activation; when aware of house, PPA activation

Question 4

PPA/FFA

Answer 4

FFA = Midfusiform gyrus
PPA = Parahippocampus gyrus