PSYC 365 Midterm 1

lesion studies

neuropsychology & patient studies

• lesion: abnormality or injury to any part of the brain

• causes:

  • born with it

  • epilepsy

  • stroke

  • injury (concussion)

  • disease (tumour)

• neuropsychology: uses laboratory tasks to measure behaviour and assess people’s capabitlies

• strongest way causality can be inferred

  • can reveal with brain regions are necessary for certain healthy cognitions and functions

  • infer causality between damage and behaviour

advantages

• can demonstrate if a region is necessary for particular functions

limitations

• patients in short supply

• damage not typically neatly limited to one region

• damaged connection to other regions could be responsible for deficit

neuromodulation (transcranial stimulation)

TMS: transcranial magnetic stimulation

• imposes magnetic pulses through a metal coil placed on the scalp

• TMS can stimulate neurons artificially, in order to demonstrate the map in the primary visual cortex

  • participants see flashes of light (phosphenes), and the location of the flashes is dependent on where the scalp is stimulated

TES: transcranial electric brain stimulation

• uses an electric coil/electrodes to ramp up or damp down neuronal activity

• can activate or inhibit specific regions of cortex

• tests behaviour and cognitive processes

  • if change is apparent after electrical stimulation, can conclude the region is necessary for a cognitive function

advantages

• non-invasive; creates ‘temporary’ lesions

• infer causation (specific region is necessary for a function)

limitations

• poor spatial resolution

manipulate brain, measure behaviour

• behaviour as the dependent measure

  • allows for inferring if a brain region (or broader network) is necessary for a certain function

lesions, neuromodulation (TMS/TES/tDCS)

transcranial direct current stimulation (tDCS)

form of neurostimulation that uses constant, direct current delivered via electrodes placed on the head

• if behaviour changes, can infer this region is necessary for certain cognitive processes

• stimulation of certain regions can also mitigate depression

EEG

electroencephalography

• electrodes place on the scalp to record the electrical activity of the brain

• EEG waves reflect the electrical output of columns of cortical neurons

• EEG measures the brain’s cumulative electrical activity

  • aka Event Related Potentials (ERPs)

• frequency domain analyses

advantages

• good temporal resolution

• inexpensive

• non-invasive

• direct measure of brain activity

• potentially portable

limitations

• poor spatial resolution

• correlational

  • can’t infer that activity in a region is necessary for a behaviour

methods for probing the brain

1. studying brain lesion patients (neuropsychology)

2. neuromodulation (transcranial brain stimulation)

3. EEG/MEG (electricity)

4. PET (neurotransmitters moving and interacting with brains)

5. MRI/DTI (info about structure, not function; DTI looks at white matter connections)

6. fMRI (measures brain activation and function)

broca’s area

damage to this area affects language production, but not comprehension

• necessary for successful speech production

• Broca’s research on aphasia was foundational for later studies on the regions responsible for language production and understanding

  • lesions allow us to infer if a region or network plays a causal role in a cognitive function

lesion network mapping

emerges from modern understanding that any one function requires the activity of various regions which work together synchronously

• locating lesions in multiple locations across patients, all of which were linked to problems in any given behaviour (eg, morality)

• concludes may need all parts of a network to work together for a cognitive function

  • necessary: lose this region, lose this function

  • sufficient: losing this region may impair function, but damage is not guaranteed lead to the loss of this function

lesions & criminal behaviour

criminal behaviour correlated with lesions in a moral decision making network

• lesions to brain regions which are structurally connected to each other

• damage to this network can lead to trouble understanding morality, making bed decisions, difficulty with empathy

• suggests intact function of the whole network is required for moral and appropriate behaviour

• we can infer causality at the level of the network, but not he individual brain regions

  • suggests disruption of a network, not single region, is key

event related potentials

• averaged EEG signal following a stimulus or response

• compared between groups and conditions

  • ERP components” linked to specific processes

• can tell you when but not where

frequency domain analysis

breaks down EEG signal into signals of different frequencies

• by breaking down EEG, we can study how populations of neurons in distant brain regions work together to produce cognition

• looking at oscillation in the brain to examine mechanisms through which communication within and between brain regions occur

  • synchronized oscillation may indicate distant population of neurons communicating with each other

    • synchronized waves — indication two regions are communicating

    • unsynced — sign of decreased or no communication between regions

frequency

how wide and slow or short and fast a wave is

• low frequencies: slow, shallow waves

• high frequencies: fast, steeper wave

• bands: delta, theta, alpha, beta, gamma

fournier transform

any complex time series can be broken down into a series of superimposed sinusoid functions with differing frequencies

• discovered information which changes over time can e broken into waves of different width that move up and down

MEG

magnetoencephalography

• uses magnetic detectors surrounding the head; magnetic fluctuations produced by electrical activity of neurons

  • pick up activity for populations of neurons

• signal not distorted by the skull, so better spatial resolution than EEG

• portable - helps surgeons better target where to operate; can provide earlier diagnoses and treatments for people with epilepsy

advantages

• good temporal resolution

• non-invasive

• direct measure of brain activity

• portable

• better spatial estimation than EEG

limitations

• correlational

  • can’t infer that activity in a region is necessary for a behaviour

intracranial EEG

ECoG — electrocorticography

• records brain activity from grids of electrodes

• placed directly on cortical surface, but can be implanted deeper

stereo EEG

• insering single depth electrodes directly on/into brain

advantages

• directly measuring brain activity; know exactly where you’re measuring from

• good temporal and spatial accuracy

• useful for synchrony measures

  • measuring neuronal activity at the source, not estimated source

• good for locating at high frequencies

limitations

• very invasive and data is rare

• no control over where electrodes are placed

  • has to have a medical reason for doing intracranial EEG

• correlational

• typically done for epileptic brains, which differ from healthy brains

  • may activate in different patterns

PET

positron emission tomography

• early PET measured glucose metabolism

• radioactive tracers inserted into brain to tag neurotransmitters

• used increasingly in clinical studies

• measure neurochemical activity

advantages

• can measure molecular processes

• can measure anything as long as there is a radio tracer (eg, glucose levels)

• cool

limitations

• expensive

• invasive and slow

• poor spatial and temporal resolution

manipulate behaviour, measure brain

• EEG

• MEG

• intracranial EEG (ECoG/stereo EEG)

• PET

allow measure of oscillations as a mechanistic signal of communication

• can use multiple methods to triangulate on a process of interest

brain imaging limitations

brain imaging cannot inherently tell us why

• same pattern of activation found in people who are clinically depressed and people who are just instructed to think sad thoughts

shows the state of one’s brain when in a particular mental state

• does not inherently tell us why they are in those states

• have to account for environmental, historical, and current event contexts

brain imaging — activation for depressed thoughts

• different degree of activation largely found in amygdala and subgenual cingulate cortex

• depression not a defect in internal workings

  • could be due to inheritance of particular genes (internal) or life circumstances (external)

brain activation — erotic images

activation in hypothalamus for both gay men viewing erotic images of men and straight men viewing erotic images of women

• thus activation could be attributed to people viewing images of their gender preference

• not just gay men having abnormality in their hypothalamus

3 things brain imaging can address

1. account for human capacities

   1. do individual capacities affect for multiple facets of intelligence

2. human limitations

   1. is there a limit to our capacity to multitask

3. explanation for psychological effect of disorders in the nervous system

   1. explain phantom limbs

MRI

magnetic resonance imaging

• portons in nuclei of hydrogen atoms (from water in body) resonate at distinctive frequency when direction of magnetic field is suddenly changed

• different tissues have unique resonance, depending on water density

  • produce images of white matter, gray matter, and cerebrospinal fluid

    • when magnetic field is changed, vibrations of H protons respond differently in matters and fluid

advantages

• very good spatial resolution

• produces anatomical (structural) images

fMRI

functional magnetic resonance imaging

• measures differences in activity between experimental conditions

  • magnetic properties associated with changes in ratio of oxygenated:deoxygenated blood (BOLD response)

  • when neurons increase activity, increase of blood in that region

• does not directly measure neuronal activity, measures BOLD

• “cognition in action”

advantages

• good spatial precision

• non-invasive

limitations

• expensive

• poor temporal precision

• indirect measure of brain activity

  • correlational relationship between activation and cognition

BOLD response

blood oxygen level dependent

• oxygenated blood delivered to active regions will alter the brain’s magnetic signal

• dependent variable for fMRI studies

• not a direct measure of neurons firing

• signal processing done on fMRI signal, and those statistics produce blobs

  • map: red patterns are more active at a certain threshold

  • colours indicate different BOLD thresholds, not brain or neural activity

fMRI encoding

standard analysis/approach

• brain mapping: measures activity during a stimulus or experimental task

  • looking for areas with higher activation in one condition over another

• compare/contrast approach; looking at activation of individual voxels across conditions

• shows what regions prefer one experimental condition over another

  • can infer these regions are involved in the cognitive process the experiment has manipulated

• whatever region is more active in a specific condition, we can say that part of the brain encodes the action

goal: functional brain mapping

fMRI decoding

representational analysis/approach

• looking at brain activity to predict (or decode) what stimulus or cognitive process could be causing it

  • what does brain activity tell us someone is doing

• data-driven approach: brain activity is a code to be interpreted

• can examine pattern of BOLD activity to see how it represents specific stimuli (eg, looking at animals)

• if we know a certain brain region encodes xyz, we can then use activity in that region to decode possible stimuli

• representational approach: coding depends on activation across voxels rather than with individual voxels

  • high activity throughout a task is not inherent indication of this brain region begin important for the task

goal: ‘mind reading’

voxel

• MRI and fMRI images composed of units called voxels

  • BOLD data collected in slices and volumes, then chopped further into cubes (voxels)

    • “volumetric pixel”

    • fMRI is measuring BOLD activation for every voxel

• brain divided into grids of cubes, no relation to anatomical divisions

• voxels can vary in size

• 150k voxels in the whole brain

high resolution MRI

• most scans at a strength of 3T (Tesla)

• now, more scans being done at 7T or beyond

• higher resolution scanners can be used to capture activation in thin cortical layers

standard approaches

encoding

• creating a basic functional map of the brain

• assumption: an area is coding for a stimulus if that area is more active that other regions in a condition

• understanding coding could depend on patterns of high and low activation

  where in the brain do clusters of individual voxels light up more for animate vs inanimate objects?

representational approaches

decoding

• pattern of activation across voxels holds information

  • we can decode information about psychological states from these patterns

multi-voxel pattern analysis

identifying representative patterns for categories, visual objects, cognitions, etc. within a brain region

• decoding individual psychological states. based on MVPA, are you looking at a bird or furniture?

• looking at brain activity to predict what the stimulus or cognitive process producing it is

this information can be used in:

• classification analysis

  • "can the brain region dissociate this information?

    • searchlights, ROIs, RFE

• representational similarity analysis (RSA)

representational similarity analyses

comparison of voxels at different points in the task

• evaluate how much the representations correlate with each other

  • when the voxel pattern evoked by two stimuli is correlated, this is a similar representation

    • how do patterns of brain activity evoked by different stimuli relate to each other?

    • how ar mental representation instantiated in the brain?

    • does multi-voxel partners tell us our brain represents a robin close to a parrot or a chair?

structural connectivity

how water moves in the brain can tell us where axon tracts are

• measured by the presence of axonal connections

  • measures axonal sheaths using Diffusion Tensor Imaging (DTI)

    • measures direction of diffusion of water which flows along white matter

• also measures myelin sheath: myelin imaging

functional connectivity

correlation in BOLD activation in different brain regions over time

• data gathered while participants doing an experimental ask or lying “at rest”

  • structural connection between regions is unimportant

• brain at rest is still working —> emergence of intrinsic networks

• fMRI allows us to see very low frequency functions

intrinsic networks

the networks of the brain which are correlated with each other while at rest (working together while at rest)

• patterns of brain activity that arise naturally and spontaneously

• highly preserved across people and cross-species

canonical intrinsic networks: networks observed by researchers and functions are known

• default

• control

• somato-motor

• dorsal attention

• visual

• salience

default network

network engage in self-related thoughts or memories; active when the mind is left to its own devices

control network

network used when exerting effort to override a habituated response (eg, placing laundry in hamper instead of leaving it on a chair)

somato-motor network

network active when moves or are touched

dorsal attention network

network engaged when deliberately directing your attention to something (eg, looking at your laptop, not the TV screen)

visual network

network used when looking at something

salience network

network used when attending to something relevant

• anterior cingulate cortex and anterior part of the insula

• decreases in salience network as anaesthesia takes hold (as people lose awareness)

  • decreases in thalamus, which relays sensory info to the cortex

  • auditory cortex then unresponsive to words and somatosensory cortex unresponsive to painful stimuli

naturalistic viewing

control for differences in brain activity when “at rest”/mind wandering

• show participants movies and measure the neural connectivity while participants watch

• allows for researchers to observe patterns of functional connectivity

retina-geniculate-striate system

comprised of:

1. retina

2. lateral geniculate nucleus of the thalamus (LGN)

3. striate cortex (aka primary visual cortex in occipital lobe)

• information (axons of retinal ganglion cells) leaves the retina via the optic nerve

• information in the nasal hemiretina (half of retina closest to the nose) projects to the contralateral LGN of the thalamaus

  • crosses to other side of the brain via the optic chasm

    • travels contralateral

• information in the temporal hemiretina (half of retina closest to peripheral) of each eye travels ipsilaterally

  • information is projected to the isiplateral lateral geniculate nucleus of the thalamus

  • primary visual cortex (V1), where visual information first enters the cortex

    • image from retina enters the V1 upside down and bulging (due to light fraction)

ipsilateral

same side

contralateral

opposite side

blind spot

spot where the ganglion cell axons leave the retina is called the blind spot

• no photo receptors; only axons departing from eye

optic chiasm

optic nerves come together to form the optic chiasm: x-shaped

• how information from one eye is communicated to the other

if optic chiasm is severed, would be blind in the nasal hemiretinas (half of retina closest to nose)

primary visual cortex

necessary for visual awareness, but does not tell us if that awareness is sufficient

• visual information can still reach the cortex through routes that bypass the primary visual cortex

• visual information enters the cortex via the V1

• retinotopically organized (neurons with receptive fields close together in visual space will have cell bodies close together

• V1 cells respond to specific types of information

  • allow us to detect basic essential features: contrast (light v dark), edges, and motion direction (moving or stationary)

    • no info about colour

• relays information from various visual areas (V1 - V4) to the temporal and parietal lobes via the dorsal and ventral streams

includes 24 secondary visual cortices

• most input received from the V1

• located in the prostrate and inferotemporal cortices; also visual association cortex

V1 mapping

• image enters the V1 flipped due to light refraction through the convex lens

  • each eye receives a 2D image map on the retina

  • passes signals via the LGN of the thalamus to the primary visual cortex (V1)

• V1 is retinotopic

  • functions as topographic map of what’s on the retina

  • more real estate is given to the centre of vision rather that peripheral vision; more light receptors located in the fovea (centre of the eye)

    • we take in more information when we fixate our gaze on something

V1 retinotopic maps

• each neuron only sees one point

  • neural firing in response to little pieces of information

• only captures basic info: edges, contrast, motion

  • but difficult to differentiate the person from the mug they’re holding

limitations

• retinotopic map has no more information than a photograph

  • the problem the brain must solve is the same as a computer vision algorithm

feedforward processing

information moves through more advanced systems, from the back to the front of the brain

• information from the V1 travels from the back of the brain to the front through the occipital cortex

  • eg, V4: sensitive to colour

  • Middle Temporal (MT): activates in response to motion

• dorsal and ventral streams move information to the parietal and temporal loves

• downstream areas (such as the Lateral Occipital Lobe and Inferotemporal Cortex) receive the small bits of information and integrates them into whole shapes and scenes

• early stages: process elements of stimuli

• later stages: integrates the elements into a whole picture

visual streams

visual information is transmitted from the primary visual cortex to the visual association cortex via two pathways

• dorsal stream (where pathway)

• ventral stream (what pathway)

ventral stream

what pathway (visual cortex to temporal cortex)

• result of lesion to inferotemporal cortex, causes impairment in discriminating objects (naming things)

  • vision for object identification

  • vision which mediates conscious perception of objects

• damage can result in visual agnosia

  • apperceptive

  • associative

dorsal stream

where pathway (visual cortex to parietal lobe)

• vision for guided action

• vision which directs behavioural interaction with objects

direct connections with areas which control movement, in the frontal lobe

• carries information about the movement of object

  • connection between MT (activated when seeing movement) and parietal cortex

  • parietal cortex uses the info about movement to guide hands and eyes

apperceptive agnosia

result of damage to ventral stream, early in the pathway

• issues with identifying whole picture, due to inability to identify pieces

  • identify car from side view, but no bird’s eye

  • loss of ability to recognize degraded stimuli

  • better with local vs. global aspects

    • a letter H made of S: could identify the individual letter S, but not the overall H

  • difficulty copying objects or shapes (reCAPTCHA; drawing a clock; copying writing)

associative agnosia

result of damage to ventral stream, later in the pathway

• problem with knowing

• able to recognize parts of an image; difficulty identifying the whole image

  • if shown an object, could not name it

    • able to describe objects

  • could copy an image accurately

  • shown a clarinet: knows what to do with hands (mimes playing it) but struggles with the size and cannot explicitly say “clarinet”

DF & agnosia

bilateral ventral stream lesions; damage in the LOC

• DF has no signal in the LOC when viewing objects

• visual form agnosia: trouble identifying visual forms

  • knows how to use an object (e.g., insert card into a slot as the orientation changes)

    • but can’t do a purely visual task (turn card to match the slot orientation, but not insert it)

• DF is able to do visually-guided actions, but not purely visual tasks

  • damage to the ‘what’ (ventral) pathway

feature detection model

bottom-up process of vision (simple → complex)

• ventral stream: V1 (primary visual cortex) → V2-MT (associated visual cortices) → Lateral Occipital Cortex → Inferotemporal Cortex

• whole object thought to be pieced together from elements or features (colour, orientation, motion, contrast)

mapping the steps from fragment to whole

• detect features of an object and glue them together into a whole

• edges may be detected first

• features put together at increasing levels of abstraction, until the object can be grouped together and matched to a mental template

• info passed from V1 in feedforward sweep along the ventral stream

lateral occipital cortex

important for object recognition

• relies on shape

• located along the ventral visual stream, outside the visual cortex

• key in detecting whole objects

gill-spector fMRI study

didn’t know what happened between the beginning and end stages of the ventral stream, which allowed for feature-based forward sweep

• tested hemodynamic response to different degrees of scrambling

  • measured BOLD activity in the ventral stream

• compared HRF plateaus in the LOC

results

• V1: prefers very scrambled

• V4: prefers partial objects (shapes; simpler, localized features)

• LOC: prefers whole objects (though some voxels behave like V4 aka responds to partial objects)

conclusion

• as information progresses along the central stream (from V1 to LOC), regions process bits to simple shapes to whole objects

  • processes features to whole objects and object categories, consistent with increasingly large receptive fields and stage-hierarchical scene of object processing

• encoding approach: more view and size invariant processing further along the ventral stream

what we knew about visual processing in 1998

• retinotopic mapping was 2D and upside down

• beginning of ventral stream: sensitive to part of an object

• end of ventral stream: sensitive to whole object

• receptive field size increases as move forward in the ventral stream

encoding approach to vision

vision builds representation of complex objects from different elements

• responds to object parts, then whole objects as they move along the occipital cortex

  • visual stream gets smarter about recognizing objects in various perspectives as information progresses down the ventral stream

• more view and size invariant processing further along the ventral stream

challenges to object recognition

• objects can be occluded

• size variance

• viewpoint variance

size invariance

ability to know an object is the same when it as a different size on the retina

• e.g., seeing someone from a distance and they look small, but knowing their actual height is larger than presented

viewpoint invariance

ability to recognize objects from different viewpoints

• childhood experiences teach us how objects look from different angles

• earlier V1 neurons: see objects from one view

  • higher order regionsL collect info from other neurons and can match that info to mental representations of objects from early experiences

testing size invariance using adaptive suppression

if voxel’s activity is size invariant: should treat different sizes of an object as the same

• suppression in hemodynamic responses

if voxel’s activity is not size invariant: should treat different sizes of an object as different objects

• no suppression in hemodynamic response

results

• early visual areas are not size invariant: same hemodynamic response for different sizes of the same object

  • treats images as different

• anterior LOC is size invariant: reduced hemodynamic response when shown different sizes of the same object

  • treats images as the same

testing viewpoint invariance using adaptive suppression

early visual areas and posterior LOC: not viewpoint invariant

• no reduced HDR, treats different perspectives of the same object as different objects

  • only recognize one viewpoint

anterior LOC: viewpoint invariant

• reduced HDR, treats different perspectives of the same object as the same object

  • recognizes objects from any angle

convolutional neural networks

computer programs used to combine bits and features into whole objects

• modelled after neurons, built in layers

  • convolution layers create maps similar to the LOC

• filters break down an image into different elements, and each successive layer is more abstract and complex than the previous one

• AI learns just by looking at data

  • learns and refines purely by looking at data; figures it out, is not told or coded for features to look for

• models determined edges, textures, and object parts are most useful to the human brain

the thatcher effect

identified in early 1980s - an inversion effect

• when face viewed upside down, not clear if someone is smiling or frowning

  • used as evidence that faces are viewed holistically

    • process faces as whole objects, rather than looking at them one feature or piece at a time

  • faces are uniquely processed; this is supported by evidence from neuropsychology and brain imaging studies

prosopagnosia

can recognize objects, but not faces

• some people born with prosopagnosia, some have it due to damage to the right temporal lobe

  • suggests faces are processed separate from objects; have their own dedicated system in the ventral visual stream

Bowers: Deep Neural Networks

• Deep Neural Networks are good at classifying objects, but they are not similar to human vision

  • psych findings indicate “core properties” of human vision

• DNNs can predict behavioural or brain responses, but they can’t tell us how or why

advantages of DNNs in prediction-based experiments (predicting of averaged experimental results)

• accurate at classifying images

• accurate at predicting human errors

• accurate at predicting brain responses

limitations: inconsistencies between humans & machines

• DNNs use texture to categorize images (humans use shape)

• DNNs use local shape (humans use global shape)

• DNNs are bad at identifying degraded or deformed images

• DNNs can’t distinguish boundaries from surfaces

  • human vision: boundaries and surfaces processed separately and combined early in visual processing stream

    • V1 neurons code for line orientations independent of colour and contrast; other neurons code for colour but are not dependent on orientations

DNN & RSA

• series of stimuli are used as input to two different systems (e.g., human brain and DNN)

• neural activity is recorded; distances in activations calculate to get a representational geometry of each system

  • representational dissimilarity matrix (RDM): how the representational geometry is expressed

• RSA score determined by computing the correlations between the two RDM

  • correlation ≠ causation; similarity could be driven by confounds

neural circuitry for face recognition

fusiform face area (FFA)

• key role in facial recognition

• region of the fusiform gyrus

• part of the inferotemporal cortex (IT) in the ventral visual stream

• not inherently evidence for innate face recognition machinery

  • suggest facial recognition reflects expertise; from birth onwards we gain lots of experience with faces

face-specialized regions in the LOC

• often right-lateralized

  • people with damage to the right LOC often have difficulty with facial recognition

monkey studies indicate the FFA is one one part of a face-selective network

• early lesion and imaging studies argue brain system respond more to faces than objects because we adapted to look and recognize faces

face-selective neurons

face-selective neurons: more likely to fire in response to a face than objects

• ECoG studies show widely distributed patches of face-preferring neurons

• FFA has the biggest cluster of face-selective neurons

  • not only cluster

• face-preferring neurons also found along LOC and IT

• fMRI indicates specialized locations in the brain for face-selective neurons

face representations & time

MEG study: participants looked at familiar and unfamiliar faces; applied RSA

• measured how quickly info about facial characteristics are extracted

• coarse to fine progression: general information is extracted before specific information

  • age and gender assessed before identity

    • age quickest

    • identity = how long it took the brain to understand it was looking at an object of some kind

      • identity processed more strongly for familiar faces (eg, racial in-group)

• some face information extract at different stages, others processed concurrently

  • suggests interdependence of their processing

race and vision

• perceptual vs conceptual influences of race & culture

• humans extract social information at a glance

  • race is a salient social category: extracted within fractions of a. second

  • implicates range of cognitive processes: attentional allocation to memory

other-race effect

• typically harder for us to tell picture of people of other races apart

  • babies can discriminate between pictures of people of all races equally well

  • by a year, develops more expertise in discriminating the type of faces they see frequently

    • typically people of their own race

    • expertise develops at the cost of being able to discriminate between other races and species as well; at the cost of bias

Eberhardt (2019)

• participants: 20 white Stanford students

• stimuli: separate sets of black and white male faces

  • 2 faces from the same race morphed into each other (0-100)

• measure: collected fMRI data while participants observed faces

  • adaptive suppression to measure other-race effect

results

• early face-selective cortex in the ventral stream showed white participants were more likely to perceive two white faces as different and the two black faces as the same

  • thus: less likely to be adaption in early ventral stream when perceiving white faces (treats them as different people)

• behavioural judgements showed that black faces had to be much more dissimilar to white faces to be perceived as different

  • reduced neural sensitivity to variability among other-race faces in early ventral stream

    • LOC confuses two different people as one person

    • have to be more dissimilar to NOT see adaptive effects

• racial disparities in discriminating individual identities occurs in early stages of facial perception

  • emerge in the face-selective cortex and mirror behavioural differences in memory and perception

conclusion

• other-race faces elicited less adaption in the ventral stream compared to own-race faces

  • error of perception

• people are more sensitive to physical variation among own-race faces and have broader tuning to other-race faces

  • habituated to repeated instances of seeing other-race people as part of the same social category rather than as distinct individuals

    • if more likely to see someone as part of a broad category, then more likely to associate them with traits associated with that category (eg, criminality)

influences on emotion perception

culture

• better at perceiving emotions from our own cultural group

• aware of the cultural norms of what emotions are most appropriate to show in certain situations

  • tunes us to certain emotional signals over others

race

• the more we identify with our own racial group, the more we empathize with the suffering of our group members

  • brain systems involved in empathetic responses show more activation

self-constural

• collectivist vs. individualist culture may affect how emotion is perceived; can influence the degree to which one would empathize with another’s emotions

  • individualistic values: perceive themselves as stable entitle and autonomous from other people and their environment

  • collectivist values: dynamic entities, defined by their social context and relationships

amygdala & FFA

the amygdala is structurally connected to every section of the ventral visual stream

• strongly connected to the fusiform face area

amygdala lights up with the FFA when looking at emotionally relevant faces

• FFA typically follows the lead of the amygdala

  • processes in-depth only what the amygdala says is important

general

• both show more activation for emotionally/motivationally relevant stimuli

• different activation patterns for racial out-groups vs in-groups

  • sometimes more sensitive to in-group than race

    • eg, more activation in the FFA and amygdala for members of one’s sport team, regardless of race

cultural in-group advantage hypothesis

suggests cultural group members show an advantage in the perception and recognition of signals of social communication from other group members

• better detection of identity and mental state in one’s own racial and cultural group

• Eberhardt (2019): in-group advantage is perceptual, not just conceptual

SAQ A: What was the big-picture question that motivated the study?

Do cultural factors influence how we perceive facial emotion in other cultural and racial groups?

SAQ B1: Describe one thing that previous research, described in the Introduction of the paper, has taught us about the influence of culture on responses to facial emotion. Describe one set of findings in the paper that support this knowledge.

Previous research has indicated that cultural group membership affects how emotional expressions are recognized; people more accurately recognize emotions expressed by members of their own cultural group.

Harada (2020) supports this knowledge through their findings of Caucasian-American and Japanese participants showing higher amygdala activity to negative faces of their own cultural-racial group, and the lowest activity to fans from other cultural-racial groups. This highlights an in-group bias to emotional processing.

SAQ B2: What two self-construal styles have been identified by cultural psychologists and what are the main characteristics of each?

1. Individualism: People with individualists values see themselves as stable entities and autonomous, separate from others and their environment.

2. Collectivism: people with collectivistic values view themselves as dynamic, defined by their social context and relationships.

SAQ C1: What was the main research question?

What is the influence of culture on amygdala responses to negative emotional expressions of racial in-group and out-group faces?

SAQ C2: What were two hypotheses put forward by the authors?

1. Bicultural (Japanese-American) participants will show greater amygdala responses to negative facial emotion of racial in-group members.

2. Collectivistic tendencies would be related to neural responses of inter-group negative facial expression.

SAQ D1: How did they define group membership?

Group membership was defined by cultural background and racial identity. There were three groups:

1. JP - Native Japanese (living in Japan)

2. CA - Caucasian-American

3. JA - Japanese-American (living in the US)

SAQ D2: What was the purpose of the shape stimuli?

The shape stimuli were a control for lower-level cognitive processes such as picture matching and button response. These controls were used to isolate the neural activity specific to processing facial expressions by subtracting the activity associated with the shape-matching task.

SAQ E1: What was the main pattern of findings for the amygdala? What patterns of brain activity reflected the “cultural in-group effect?”

The amygdala had the highest activation for negative faces of one’s own cultural/racial group, indicating a cultural in-group effect. The cultural in-group effect is reflected by JP and CA groups having greater amygdala activation when viewing faces of their own racial group.

SAQ E2: What pattern of brain activity did they find in one particular group of participants when they looked at the whole brain?

JA participants showed greater activation in the midline cortical regions (ventral medial prefrontal cortex and posterior cingulate cortex) when viewing negative facial expressions of Japanese individuals.

SAQ F1: How do they interpret the midline cortical findings observed in one particular group?

Harada et al. interpret these findings in the JA group as reflecting greater engagement of self-related processing when viewing negative expression from their racial in-group. The ventral medial prefrontal cortex and posterior cingulate cortex are associated with self-related processing, such as autobiographical memory.

SAQ F2: Was their hypothesis about individual differences supported? Why or why not?

The hypothesis about individual differences was partially supported. Though collectivistic tendencies were related to greater neural responses in the amygdala, Japanese-Americans with higher collectivistic tendencies had the highest neural responses when processing negative facial expressions of Japanese people. Overall, the JP and JA groups had greater activation to their in-group than CA had to their own group.

SAQ G1: How do they describe the overall meaning and importance of the results?

The study demonstrated that neural responses during the processing of emotional faces could be modulated by social factors such as cultural norms and racial identity. In-group biases are reflected in amygdala activity during the processing negative emotional faces.

This is especially meaningful for bicultural individuals, such as Japanese-Americans, as the findings indicate bicultural individuals may show a unique pattern of brain activation when processing faces from both their cultural groups.

SAQ G2: List TWO limitations of the study and explain why they are limitations.

1. No face control condition; only used angry and fearful faces. This is a limitation because the amygdala responds to all faces, not just threats and specific emotions. Research into activation towards positive and neutral emotions should be further investigated.

2. Insufficient sample size for correlations/low power for group comparisons. When making brain behaviour correlations, the limited sample size can lead to unreliable or misleading results. This calls into question the support for their second hypothesis.

SAQ H1: What are the larger cultural or societal implications of these findings? Do you agree with the authors’ conclusion? Why or why not? What do you think is important that they didn’t think of?

The authors conclude emotional responses are shaped by culture and group membership in complex ways, which have implications for facilitating social harmony (e.g., understanding and addressing biases and conflict). I agree with their conclusion regarding the implications for social connections between different groups, but I believe historical context is also an important consideration. For this study in particular, America and Japan have a contentious history in the wake of WWII, and it is highly possible that these historic biases act as a confounding variable in neural response of negative emotions.

why do some amputees continue to feel their arm, even though it’s absent?

representation of forearm and hand in somatosensory cortex

• stimulation of stump can evoke cortical activity

  • Cortical activity then perceived as touch to a nonexistent forearm

  • Stimulation of visual cortex evokes phosphenes 

why do some people see colours when they read or hear words?

aka synaesthesia

• Info about the senses is integrated via connections within and between sensory systems

  • Connections between areas for analysing shape and colour are abnormally strong for people with synaesthesia

• Illusion of seeing colour related to activation in V4 visual area

  • connection with parietal cortex for those who hear words and see colours

  • strong connection with inferior temporal cortex for those who saw colour while reading

• Activity in early sensory area + activity in insula and anterior cingulate cortex is sufficient for phenomenal awareness

do we need to recognize an object to know how to handle it?

Depends on ventral visual system; recognition and classification of objects achieved through hierarchical analysis 

• Identifying object is separate from dorsal visual system (responsible for knowing how to act with the object)

  • Even if one system damaged, the other can still operate

    • Demonstrates localization of function in the brain

dorsal

top surface of the brain

ventral

bottom surface of the brain