Cognitive Neuroscience Midterm 3

pathways of perception

• dorsal “where” pathway from occipital to parietal cortex

  • neurons have eccentric non-foveal receptive fields

• ventral “what” pathway from occipital to temporal cortex

  • neurons have foveal receptive fields for fixation

pathways of perception, Mishkin et al., 1983, monkey lesion study

• what are the monkeys doing? object/landmark task: selecting a foodwell based on either the identity of the object or the location of a landmark

• IVs? 1) task, 2 conditions: object or landmark and 2) lesion location, two conditions: dorsal or ventral

• DV? accuracy

• results: monkeys with a dorsal legion performed well on the object task but poorly on the landmark task and vice versa for monkeys with a ventral legion

  • double dissociation proves independence and lesions prove necessity

dorsal path damage

• optic ataxia: an inability to accurately reach for still objects, moving helps some

  • contralateral parietal cortex damage

ventral path damage

• card slot test: participants must orient their hand so the card they are holding matches the orientation of the slot

  • controls without damage are successful, but patients with damage are effectively guessing because they cannot recognize the shape but are able to physically put the card into the slot when it is oriented for them because they can recognize where objects are in space

pathway to object recognition

object → early vision (sensation and perception) → shape encoding (perception) → object matching (recognition) → “apple”

shape encoding, Kanwisher et al., 1997, human PET study

• what are the participants doing?

  • viewing stimuli

• IVs? type of stimulus, three conditions: scrambled (unrecognizable), novel (unnameable so separates shape from recognition), and familiar (nameable)

  • 1 2 and 3 = feature extraction, 2 and 3 = shape encoding, 3 = memory matching

  • for shape encoding, compare familiar and novel to scrambled (cognitive subtraction)

• DV? PET signal

• results: specific cortex regions including lateral occipital cortex (LOC) had increased activity for shapes over parts and thus support shape encoding

deficits in the object recognition pathway

• visual agnosia: a failure of visual recognition, 2 types:

  • apperceptive agnosia: difficulty recognizing shapes and objects by sight

    • cannot name or draw objects but can identify them using hearing and touch

    • unusual views test: are these 2 objects the same? (same object from different angles/positions)

    • shadows test: manipulate luminance and ask if these 3 objects are the same

    • what is impaired? shape encoding leading to issues with object constancy

    • what is damaged? posterior temporal lobe, LOC

  • associative agnosia: inability to recognize objects by sight despite intact perception

    • cannot name objects but can draw them and identify them using hearing and touch

    • can see distinct shapes but cannot name them

    • matching-by-functions task: which objects match in terms of what they do even though their shapes are different?

      • apperceptive agnosiacs also fail this task

    • what is impaired? object matching, the ability to associate percepts with meaning

    • what is damaged? anterior temporal lobe

how are objects represented in the brain?

• ensemble coding: voxels code for different information that all together forms an object

• grandmother-cell coding: 1 cell or voxel forms 1 object

• object representations, Haxby et al., 2001, human fMRI study

  • hypothesis: object categories are represented by patterns of activity in VTC

  • what are participants doing? viewing stimuli

  • IVs? stimulus category, multiple conditions: scissors, shoes, chairs, cats, etc.

  • DV? fMRI BOLD signal in VTC

  • multivariate pattern analysis: show participants many examples of an object category and see if voxel patterns in VTC are predictable → algorithm accuracy > 90% supporting hypothesis and ensemble coding

behavioral evidence for face-specific processing

• face inversion effect: it is harder to recognize inverted faces related to upright faces

  • this effect does not occur for other objects

  • holistic processing account

    • study 1

      • hypothesis: faces are processed as a ‘whole’ rather than in terms of parts

      • IVs? 1) stimulus category: face or house and 2) test type: whole or parts

      • DV? accuracy

      • prediction: test accuracy will be higher for faces tested in the “whole” vs. “part” condition, but there should be no different for houses → correct!

    • study 2

      • hypothesis: faces are processed as a ‘whole’ rather than in terms of parts

      • IVs? 1) two study conditions: whole or parts and 2) two test conditions: upright or inverted

      • DV? accuracy

      • prediction: test accuracy will be higher for faces learned as a whole vs. parts, but only when the test is on upright faces

face processing in the brain, Kanwisher et al., 1997, human fMRI study

• what are participants doing? viewing stimuli

• IVs? type of stimulus, two conditions: faces or objects

• DV? fMRI BOLD signal

• results: activity in specific brain regions for either faces or objects, including the fusiform face area (FFA) in the temporal lobe that responds more to faces than objects

face processing in the brain- convergent approach, Tsao et al., 2006, fMRI and SURs in monkeys

1) fMRI to locate FFA monkey homologue ‘face patch’ → two monkeys view faces and objects

2) electrodes placed in ‘face patch’ regions → monkeys view faces, objects, and scrambled images (preserve low-level visual features)

• results: 97% of FFA neurons responded only to faces, not objects or low-level features → stimulus selective

link between behavioral and neural face processing, Yovel and Kanwisher, 2005, human fMRI study

• what are the participants doing? face inversion task

• DV? fMRI BOLD signal

• does the FFA respond more to upright than inverted faces (matching the FIE)? yes!

• does the LOC respond more to upright than inverted faces? no → specific to FFA

• are behavioral (better recognition of upright vs. inverted faces) and neural (more FFA activity for upright vs. inverted faces) FIE related? yes → positive correlation between them: more FFA activity = better recognition of upright faces and vice versa

FFA and specific face identities, Yovel and Kanwisher, 2005, human fMRI study

• repetition suppression: if the exact same stimulus is presented twice (repetition), then neuronal activity will decrease (suppression)

  • commonly measured with fMRI

  • theory: neuronal responses become more efficient, so less neurons respond the 2nd time

  • should be observed in brain regions sensitive to the type of stimulus

• hypothesis: if FFA represents face identities, it should show repetition suppression only for repeated faces (only upright not inverted faces) correct!

neuropsychological data and face processing

• prosopagnosia: a selective deficit in recognizing faces

  • produced by damage to the FFA

  • perform FIE task with prosopagnosia patients: perform the same as controls for objects and inverted faces but perform worse for upright faces → demonstrates necessity

  • perform FIE task with both prosopagnosia and agnosia patients: prosopagnosiacs recognize objects but not faces, while agnosic patients recognize faces but not objects → demonstrates independence