Jason Flashcards
(156 cards)
<p>What direct and indirect ways neural activity can be measured</p>
<p><u>Directly</u></p>
<ul> <li>AP (Single neruon)</li> <li>Local Field Potentials (Summed activity)</li></ul>
<p><u>Indirectly</u></p>
<ul> <li>Metabolic changes</li> <li>Blood flow <ul> <li>Cerebral blood volume</li> </ul> </li> <li>Blood volume <ul> <li>PET/fMRI</li> </ul> </li></ul>
<p>What is the function of fMRI</p>
<p>Localise hemodynamic changes from neural activity.</p>
<p>Why is fMRI popular, contrast with other methods</p>
<ul> <li>Non-Invasive <ul> <li>No needles, unlike PET</li> </ul> </li> <li>Enable human studies</li> <li>Focal <ul> <li>High precision, unlike EEG</li> </ul> </li></ul>
<p>What is the physics behind the basic MRI (fast)</p>
<ul> <li>Hydrogen has a single proton which preccesses around an axis</li> <li>RF pulse, aligns parallel/anti-parallel</li> <li>After swtich off, spins back</li></ul>
<p>Cognitive processes require energy. Where do we get energy and how does it relate to fMRI</p>
<ul> <li>Cognitive Processes = <ul> <li>ATP =</li> <li>Use oxygen from hemoglobin =</li> <li>Reverse ion influxes underlying synaptic potentials and action potentials</li> </ul> </li> <li>fMRI relies on difference inmagnetic responses between oxyhemoglobinand deoxyhemoglobinblood</li></ul>
<p>Oxygenated vs Deoxygenated Blood. Difference in MRI signal</p>
<p><u>Oxygenated</u></p>
<ul> <li>Weakly diamagnetic</li> <li>Does not distort magnetic field</li></ul>
<p><u>Deoxygenated</u></p>
<ul> <li>Paramagnetic</li> <li>Distorts magnetic field</li></ul>
<p>What is the standard practice in analysing BOLD?</p>
<ul> <li><u>Spatial smoothing </u>by 8mm <ul> <li>Allow for group averaging by correspondence across brain</li> </ul> </li> <li>Use g<u>eneral linear mode</u>l to quantify BOLD changes <ul> <li>Correlation between time course of the BOLD signal change in each voxel of the smoothed images with the measure of cognitive function.</li> </ul> </li></ul>
<p>What is the implication of the BOLD response (thus far)</p>
<p><u>Cortical Segregation/Modularity</u></p>
<ul> <li>Explains spatial structure of brain responses</li> <li>'Neo-phrenology'</li></ul>
<p>What is the amptitude of BOLD signal correlated with?</p>
<p><u>Amplitude of the BOLD signal associated with</u></p>
<ul> <li>Local field potential <ul> <li>Large no. of active neurons responsive together</li> </ul> </li> <li>Increases in gamma-band electrical <ul> <li>EEG</li> </ul> </li> <li>Quite often correlated with spike frequency <ul> <li>Animal Studies</li> </ul> </li> <li>Electrocorticographic (ECoG) at mm accuracy</li></ul>
<p>> Confidence that BOLD is associated with activity</p>
<p>What does magnetic suspectibility of blood depend on?</p>
<ul> <li>Blood oxygenation, but also depend on</li> <li>Regional cerebral blood volume (CBV) <ul> <li>Not independent</li> </ul> </li></ul>
<p>What are the main limitations of fMRI (together with elaborations)</p>
<p>1.) Mislocalisation of hemodynamics</p>
<ul> <li>Local changes in oxygen use and blood volume are carried downstream <ul> <li>Mislabel brain region</li> </ul> </li> <li>CBV is useful but most still use BOLD</li></ul>
<p>2.) Slow Changes in hemodynamics</p>
<ul> <li>Might not capture true response</li> <li>Precise neural coupling invisible to fMRI</li></ul>
<p>3.) Uncertainity in type of neurons involved</p>
<ul> <li>Positive BOLD signal could be excitation or inhibitory</li></ul>
<p>4.) Direction of causation is unclear</p>
<ul> <li>Separate region co-active, but does not say how it influence one another</li></ul>
<p>5.) Spatial Limitations (<em>localising)</em></p>
<ul> <li>Vague despite being much better than EEG.</li> <li>Cannot explain layer-dependent activity</li></ul>
<p>6.) Sparse encoding vs population encoding (resolving)</p>
<ul> <li>Spares Encoding: Poor</li> <li>Population Encoding: Good <ul> <li>A BOLD signal driven by a stimulus does not mean that the entire area is used to process that stimulus, or even that class of stimuli</li> </ul> </li></ul>
<p>What are the 3fundamental limitations of fMRI</p>
<ul> <li>Some nerual activity (Magnetic Field) are too small to be localized with fMRI</li> <li>MRI only shows vascular responses to neural activity</li> <li>Conclusions are ambiguous because it could reflect (blood velocity? volume? oxygen?)</li></ul>
<p>What are recent advances in fMRI</p>
<p><u>Multivoxel pattern analysis</u>(Statistical techniques)</p>
<ul> <li>Whole Brain View</li> <li>Does not require spatial smoothing</li></ul>
<p><u>Voxel encoding and population field mapping</u>(Statistical techniques)</p>
<ul> <li>Functional property of neurons</li> <li> <p>Not possible with group averaging</p> </li></ul>
<p><u>Hi Resolution (7T) Scanning</u></p>
<ul> <li>Isolate activity in single cortical column (sub-mm)</li></ul>
<p>What are recent advances in structural MRI</p>
<ul> <li><u>CBV Changes</u> <ul> <li><u></u>Allows resolution of the cortical layer</li> </ul> </li> <li><u>dMRI tractograhy</u> <ul> <li>Connectivity between brain regions using density of fibres <ul> <li>Map of how different brain regions are associated and correlated with one another</li> </ul> </li> </ul> </li></ul>
<p>Movement away from modularity to connectivity</p>
<p>What are the methods for intracanial, extracranial (a)electrical recordings and (b) electrical stimulations</p>
<p><u>Intracranial Recording</u></p>
<ul> <li>Single cell animal studies</li> <li>ECoG</li></ul>
<p><u>Extracranial Recording</u></p>
<ul> <li>EEG</li> <li>ERP</li></ul>
<p><u>Intracranial Stimulation</u></p>
<ul> <li>DCES</li></ul>
<p><u>Extracranial Stimulation</u></p>
<ul> <li>tDCS</li></ul>
<p>Extracellular Recordings of Single Neurons: What did anesthetised and awake behaving studies on anmmal tell us? Can we study multiple neurons?</p>
<p><u>Anaethsized Studies</u></p>
<ul> <li>Sensory and Motor</li></ul>
<p><u>Awake behaviour Studies:</u></p>
<ul> <li>Higher level functions like attention</li></ul>
<p>Mulitple neurons can be studied with electrode arrays</p>
<p>What is Local Field Potential. What is it? What are thecons?</p>
<p><u>LFPs:</u></p>
<ul> <li>Not related to individual neurons <ul> <li>Measures neural activity up to 3mm from electrode</li> </ul> </li> <li>Use same electrocode as single unit recording</li></ul>
<p><u>Cons:</u></p>
<ul> <li>LFPs likely represents summed activity of large numbers of neurones with synchronous input</li> <li>More likely to reflect type cells with dendrites facing in the same direction away from cell body <ul> <li>e.g., pyramidal cells</li> <li>Same type of cells</li> </ul> </li></ul>
<p>ECoG overview. What is it used to clinically</p>
<ul> <li>Uses 2-256 electrocedes in an array placed directly on exposed surface</li> <li>Records LFP (Probably pyramid cells)</li> <li>Used to treat epilepsy by identifying region generating seizures</li> <li>DCES uses the same electrode</li></ul>
<p>ECoG Pros as a Research Tol</p>
<ul> <li>Understanding neural function <ul> <li>High spatial and temporal resolution</li> <li>Both single and multi-unit recording</li> </ul> </li> <li>Confirms electrophysiological recordings from animal models</li> <li>Understanding how indirect methods relate to direct neural responses <ul> <li>BOLD poor temporal</li> <li>EEG poor spatial</li> </ul> </li></ul>
<p>ECoG + BOLD Finger Flexion Results. What is the implication?</p>
<p>7T fMRI prior to ECoG in finger flexion</p>
<p>"<em>to what degree does localisation of neural activity from BOLD correspond to ECoG</em>"</p>
<p><u>Results</u></p>
<ul> <li>High frequency ECoG (65-95Hz) matches localised BOLD</li> <li>BOLD co-localises rapid neural changes at fine spatial scale (mm scale)</li></ul>
<p><u>Implications</u></p>
<ul> <li>Showed that 7T fMRIreliably captures important aspects of neural activity</li></ul>
<p>EEG Overview. What are the cons?</p>
<ul> <li>Electrical activity measured from large numberof synchronous, aligned neurons</li> <li>Usually recording pyramial neurons (sameas LFP)</li> <li>Best for Gyri,not sulci</li></ul>
<p>EEG Pros and Cons</p>
<p><u>Pros</u></p>
<ul> <li>Cheap</li> <li>Good Temporal Resolution</li></ul>
<p><u>Cons</u></p>
<ul> <li>Poor Spatial Resolution</li> <li>Not good for deep structures <ul> <li> <p>Voltage drops off rapidly with distance, so activity from deep sources is difficult to detect</p> </li> </ul> </li></ul>
<p></p>
<p>How does EEG move to ERP</p>
<p>x1000 trials + signal averaging</p>
<p>DCES Overview andCons</p>
<p><u>Overview</u></p>
<ul> <li>Stimulation of Single Neurons <ul> <li>Mostly on awake behaving non-human primates</li> </ul> </li> <li>UsingECoG electrodes to stimulate</li></ul>
<p><u>Cons</u></p>
<ul> <li>Clinical patients limit the basic research <ul> <li>Must have epilepsy</li> <li>No choice in electrode location <ul> <li>Gyri;Biased to seizures</li> </ul> </li> <li>Surgery <ul> <li>Expensive</li> </ul> </li> </ul> </li></ul>
tDCS Overview and Aim
Overview
- Passing a weak DC current between electrodes placed on the scalp
- Extra-cranial
Aim
- Primarily to improve mental function
tDCS vs other techniques
- Does not require medical intervention (Non-invasive)
- Uses DC to influence brain activity
- Uses weak current to influence brain activity
How does tDCS work
- Small current passed between two electrodes on the scalp
- Assume that current flows though the brain
- Neurons under the anode more easily activated than they otherwise would be
- Excitation: Anode
- Inhibition: Cathode
- Not generating action potentials, but changing response of neurons
- Neurons under the anode more easily activated than they otherwise would be
tDCS pros and cons
Pros
- Non invasive
- Cheap to purchase and use
- Easy to use
- Safe when using established protocols
- Straight forward ethics
Cons
- Precise mechanisms elusive
- Difficult to precisely and selectively stimulate a target brain region
Does scientific evidence suggest tDCS is effective? What is the criticism (of the scientific evidence)?
Meta-analysis found no reliable effect.
Criticism of meta-analysis
- Not enough studies
- Hetereogneity of poor designs (gold-rush)
What are the difficulties in establishing whether tDCS is effective?
- High prevalence of “adverse” events = strong placebo
- No active sham control
- Participants can tell whether they're in sham or active
Has the rapid increase in studies contributed to the tDCS confusion?
- High rates of “new” findings biases against verification
- Gold rush mentality (citations, funding, no replication)
Has the way we do science contributed to the confusion tDCS. What are the phenomenas?
1.) File drawer phenomenon
- Publish positive results
- Ignore negative or non confirmatory results
2.) Forking path phenomenon
- Lack of specific predictions in the absence of a good understanding of how tDCS works
3.) Increase in importance of science communication
- Expectation > Truth with single result
- Single result can define field if widely promoted
What are paradoxical image effects?
- Tiny image difference may change emotion and identity
- Big image difference have no effect on identity
What are some models of face-processing
(3 questions we can ask when we process faces)
- Figural
- Face / non-face
- Semantic
- General (Gender)
- Specific (Familiar)
- Learnt/Innate
What is viewpoint dependency
Recognition drops with face inversion
What is image volatility?
Recognition drops with reversed contrast
What is Identity stability
Caricatured faces are often more identifiable than veridical photographs
Evidence that face recognition is consistent across visual arrangements
Recognition
- Occurs in extreme deformation
- Depend on external features
- (e.g. prosopagnosics)
Behavioural evidence for a specialised face pathway
1.) Face inversion effect
2.) Holistic processing
- The composite effect
- The whole-part effect
3.) Neuropsychological evidence
- Prosopagnosia
- Visual object agnostic with intact face-processing: CK
Behavioural evidence for face-inversion effect. Upright vs invered
- Configural processing for upright faces
- Featural processing for inverted faces
Behavioural evidence for holistic processing. Composite effect
Composite
- Slow to identify half of a chimeric face aligned with an inconsistent other half-face
-
Interference from the other parts of the face
-
-
Easier to identify the top half-face when it's misaligned with the bottom one than when the two halves are fitted smoothly together
- Suggest mandatory processing of whole face
Behavioural evidence for holistic processing. Part-whole. What does it not occur for?
- Better at distinguishing two face parts in the context of a whole face than in isolation
- Does not occur for controls
- inverted
- scrambled
- house
Evidences for expertise in face-inversion
Diamond and Carey (1986)
- Inversion for houses
- Inversion for landscapes
- Not as much as faces, but the statment that "only faces show inversion effect" is not true
- Comparative inversion for dog experts (Not novices)
Rossion and Curran (2010)
- Greater inversion effect correlates with self-declared car-expertise
Why are greebles good controls?
Face-like properties.
- Small number of parts in common configuration
- Hard to identify based on single feature
- Identification is best by using all features and relationships between them
Gauthier and Tarr (1997). Results. What does it suggest.
Results
- Experts - Defined as someone who could recognise a Greeble’s “gender”, “family”, "name"
- Faster
- Accurate
- More sensitive to configural changes (Transformed)
- RT to upright Greebles slower in the Transformed Configuration relative to the Studied Configuration condition
Argued for qualitative change in recognition - Understanding the rules of greebles
What did Farah (1990) argue in terms of cases of visual agnosia
Argued for two independent recognition systems
- Structural/Part-Based mechanisms
- Associated with “normal” object recognition
- Holistic mechanisms
- Associated with face recognition
Is there evidence of a double dissociation for Farah (1990)
Separate modules for face and object recognition
(a) Prosopagnosia
- Normal object with poor face recognition
- Usually damage to fusiform gyrus
- Pure prosopagnosia is rare
(b) Visual Object Agnosia
- Poor Object with normal face recognition
- Only CK
How do we measure facial recognition
- Before They Were Famous
- Cambridge Face Memory Test
- Cambridge Face Perception Test
BTWF Test on Facial Recognition. What does correct identification require? Flaw?
- 59 pictures of celebrities (as children)
- Correct identification requires generalization across substantial change in the appearance of the face
- Flaw
- Does depend somewhat on prior exposure
CFMT on Facial Recognition. Flaw?
- 6 male faces
- 3 trained view
- Different perpsectives
- 3 alt forced choice
- Which of this faces have you seen before
- Recognise picture from non-trained views
- 4 difficulty levels
- 3 trained view
- Flaw
- Might be reliant on memory
CFPT on Facial Recognition.
- Test images at ¾ view
- 6 frontal non-target faces morphed with target (different %)
- Can do for upright and inverted faces
- Rank from most to least similar
Greeble learning in a prosopagnosic
- Edward
- Poor face inversion, no face-inversion effect
- Normal Greeble recognition performance
- Suggests face deficits do not involve brain processes used to acquire Greeble expertise
What are some properties of congenital or developmental prosopagnosia. What are 2 notions on face recognition ability.
- Poor facial recognition
- Absence of brain damage or other cognitive deficits
- note: prosopagnosic is usually FFA damage
- 2%–2.5% population
1.) Healthy/Pathological
2.) Broad (normal) distribution of face recognition ability, with developmental prosopagnosia on lower tail and superrecognisor on upper tail
How do superrecognisors display the face-inversion effect. What does it suggest?
- Perform well on facial recognition task (CFMT and CFPT with upright)
- Larger face inversion effect (CFPT with inverted)
- Supports normative idea that inversion effect is not qualitative different processing compared to normals.
Evidence for face neurons from human adaptation
- Faces show adaptation
- Like Neurons
- Bistable perception in semi-upright
- See one face then the other
- Suggest neurons adapted to see one face than the otehr
What are the 3 studies of neural mechanism of face-processing in non-human primates
- Single Cell
- fMRI
- Microstimulation
Non-Human Primate Study (1): Single Cell Study in face-processing in non-human primates
- Non-human primate has face neurons
- Face cells in IT (fusiform gyrus) responded to an intact face
- Not selective for individual features presented in isolation
Non-Human Primate Study (2) :fMRI study in face-processing in non-human primates (READING)
- Identified "Face Area" using fMRI
- In temporal lobe
- Recorded 400 cells in "Face Area"
- 97% of visual cells responded exclusively to faces.
- Apple and clock showed some response (roundness? property of faces?)
Non-Human Primate Study (3): Micro-stimulation in face-processing in non-human primates. Methods and Results
Is IT (Fusiform Gyrus) a face perception area?
- Stimulated neurons in IT and influence face/flower perception
Results
1.) Stiimulation
- Especially 50-100ms
- Higher likelihood to see faces at all levels of noise
No stimulation
- Equal probability to see face/flower
2.) Stimulation effect greatest for face-sensitive cells within IT
What are the human physiological evidences for specialised facial pathways
1.) MEG
2.) fMRI
3.) ECoG
4.) Stimulation
Human Study (1): MEG Study for face-perception in humans neural.
MEG (Cross of EEG and fMRI)
- Temporal responses for faces consistently higher M170 compared to cars and shoes
- No difference between novices and experts
- Suggesting signal for faces
Human Study (2): fMRI Study for face-perception in humans neural.
Manipulaed parts and configuration of faces and houses.
FFA
- Faces have bigger respones than houses, hands, two-tone-faces
- Does not depend on changing spaces or parts
- Sensitivity to faces
LOC (Lateral Occipital)
- Bigger reponse to changes in Parts
- No difference for faces or house
- Insensitive to identity
- Senstivity to whether 2 images are the same
Human Study (2.5) fMRI Study for face-perception in humans neural. What happens to bistable stimuli and FFA
For bistable stimuli, the FFA responds more strongly when subjects perceive a face than when they do not
- Suggests FFA is activated specifically by whole faces, not by low-level stimulus features that comprise faces
What are neural evidence (not greebles) to suggest FFA is related to level of expertise
- Car expert no centre of right FFA (subset of FFA) for bird object
- Bird expert no centre of right FFA (subset of FFA) for car object
Why are cars, birds, dogs criticized for using as expert for face-perception? (What are face-like properties)
They do not have face-like properties
- Similar features arranged in similar configurations
- Recognition at subordinate level
- Stimuli for which participants are experts
What happens when greebles become experts in FFA vs novices
- FFA fMRI response change
- Part whole effect (for upright)
- Greeble expert recognition behaviour and physiology consistent with that found for faces
- Configural effects not present in novice data
Comparative fMRI: Human v Macaque
Activate the same areas
Face Perception: 8 Epilepsy Faces Single Units
- Single units responded not only to faces, but familiar faces such as Halle Barry
- Not only face information, but person-identity units
Human Study (3) Face Perception: fMRI + ECoG
- Category-selective response in ventral temporal cortex (VTC)
- High Frequency Broadband (30-160HZ) [only high]
- Strong positive correlation between fMRI and ECoG signal in HFB
- Spatial coupling was tighter for face-selectivity than house selectivity
Note: HFB
Human Study (4) What happens with FFA is electrically stimulated. ECoG + HFB
- Left and right FG contained face-selective high-selective HFB
- EBS of right FG: Distort face-selective distortions
- EBS of left FG: Distort non-specific vsual changes (e.g. color)
- Stimulation of sites that caused face-related changes (right FG) were more face-selective
Facial Emotional Recognition: What is the hypothesis to suggest innateness. Evidence from animals (1 evidence) and humans (3 evidences)
Universal facial expression hypothesis
- Evolved to recognise emotions for survival
Evidences: Animals
- Similar expressions in closely related species in animals
Evidences: Humans
- Expression evident in deaf and blind people
- But blind are less proficient at posing
- Babies
- Innate or learned early
- Though different emotions have different timings in learning
- Cross-cultural studies
- High cross-cultural agreement in judgments of emotions in faces
- Both literate and preliterate cultures
What is the role of the amygdala in facial processing
Guide attention to emotionally relevant information
Anger Superiority Effect.
Study
Innate mechanism for detecting anger
Finding the face in the crowd
Target Absent
- Longer RT than Target Present
In both Target Present/Absent
- Angry faces identified consistently faster (Processed first)
Anger Superiority Effect Study's Caveat.
Some people find an advantage for happy faces
Caveat
- May depend on stimulus set
- Larger set size take longer
- May depend on feature strength
- Whiteness of teeth
What are evidences to support non-holistic processing of facial emotions
Non-holistic processing
1.) Facial Action Coding (FAC)
- Code emotion by looking at the muscle activity that gives rise to these changes in the features
- Recording movement of muscles and decode
2.) Emotion perception of morphed faces reveals categorical perception
3.) Calder et al. (2000)
- Top half of the face: Anger, fear and sadness
- Bottom half of face: Happiness and disgust
- Equally: Surprise
What are the evidence suggesting facial expressions are processed holistically?
Holistic Evidences:
Composite effects
- Mismatching emotions aligned and misaligned. Aligned poorer.
- (1) Upper Expression
- Aligned vs. misaligned (happy faces): slower
- Aligned vs. misaligned (angry face): faster
- Effects disappear with inversion
- (2) Expression judgements for composities unaffected by identity, vice versa.
- Evidence for holistic processing in recognition and identificaiton but some independence between emotion and identity
Can developmental prosopagnosics decode expressions of emotion?
Mixed
Yes
- Can label basic facial expression
- Can decode difficult to categorise emotions
No
- Deficits in facial expression recognition
- Suggest using individual features to decode
Evidence for holistic processing in emotions in prosopagnosics
Study and Results
Composite Task for Prosopagnosics: Weaker holistic processing in Emotions
Match identity and emotions
Results
- Controls shows drop in performance in aligned compared to unaligned
- Increased RT in aligned
- Prosopagosics smaller drop in aligned compared to unaligned
- Less increase in RT in aligned
- Bigger effect for control suggest prosopagnosics have impaired holistic processing
- For both identity and expression
What do models of face recognition suggest for expression and identity
Independence - Expression and Identity are separate
Bruce and Young:
- Dedicated route for Expression
Haxby
- Superior Temporal Sulcus (STS) = Expression
- Ventral Temporal Route (include FFA) = Identity
Do we have different locations for different emotions. Compare emotion and identity
Yes (unlike faces, emotions have a more dynamic network)
Behavioural Evidence to suggest different locations for different emotions
- Dynamic changes in muscle activity over time for different facial expressions suggest decoding over time
- Disgust, anger, fear move more quickly.
Physiological Evidence to suggest different locations for different emotions. Amydala Routes
Activation of Amygdala
- 120ms - Fast (low-dirty road)
- 170ms - Detailed perception
- 300ms - Conceptual knowledge of emotions
- Unlike FFA, processing emotions involve a much larger network
Physiological Evidence to suggest different locations for different emotions. MEG Speed study
MEG - Response to happy/fear/neutral in identity/emotion task
- 90ms
- Orbito-frontal response to emotion without attention
- 170ms
- Right-insula response to emotion with attention
- 220ms
- Identity processing
Thus, emotions are processed subconsciously before identity or even attention
To what degree does is amygdala fear responses specific for faces?
Faces vs non-face (gun) threatening stimuli in matching task
- Preferential right amygdala response to faces
- Implying amygdala playing a role in decoding fearful information in faces
What else does the Amygdala mediate other than fear
- Anger, Disgust, Sadness, Happiness
- May be responsive to all emotional-relevant information, not just fear
- Always activated in tasks requiring emotional decision-making
Are all facial features important for encoding fear? Amygdala
- Amygdala responsive to large eye whites in fear (and surprise) expressions
- Black eye no difference, implicate whiteness of eyes are important
- Responds more to whole faces
- i.e. responds to eyes, but holistic processing leads to larger responses
Who is SM. What can he do/ cannot
Bilateral amygdala lesion. Loss of visual information for danger.
Can
- Perceive fearful tones in voice
- Perceive body positions
- Normal startle
- Neural pathway independent of the amygdala
Cannot
- Recognize facial emotion in others (whiteness eyes). Look at mouth instead
- Lack of natural fear to snakes and spiders
- Lack of loss aversion (gambling)
Eye tracking in SM. What happens when he's told to look
- Absence of fixations on the eyes across emotions
- When instructed
- SM decoding moves up to normal
- Suggests that the amygdala is important for initiating eye movements for emotional decision-making (recognizing fear)
What are some disorders associated with disrupted facial emotion processing
- PTSD
- Phobias
- Depression
- Schizophrenia
- Autism
Why are numbers important to study?
- Central to who we are (16,000 numbers a day)
- Basis of Civilization
- Technological advancement depends on number
- Number deficits affect people’s opportunities in society
What does sophisticated use of number depend on?
Sophisticated use of number may depend on the development of language
- Relationship is not clearly causal
- Perhaps language just generally enriches/focuses learning
What are evidences suggesting numbers are different from language. Broad Evidences
- Neuropsychological (Dyslexia vs Dyscalculia)
- Animals
- Preverbal Children
- Cultures with limited language for numbers
How do we describe numbers and language
Numbers: Foundation of Knowledge
Language: Basis of communication
Animal Studies. Why are numbers important for animals
Evolutionary Significance
What are the 6 animals discussed in W3L2
- Clever Hans (Horse)
- Jakob (Raven)
- Alex (Parrot)
- Desert Ants
- Lions
- Chimps (Ayumi)
Animals in Numbers: Clever Hans Horse
- Initially thought to be able to do math by hoof stamping (including square roots)
- But later found he was reading owner's face
- Faces, not arithmetic
Animals in Numbers: Jakob Raven
- Could select a pot with a specific number of dots on the lid (1-7)
- Trained to open boxes and eat the seeds contained in them until a precise number of seeds had been eaten
Animals in Numbers: Alex Parrot
- Numerical ability in context of langauge
- Could squawk "1-6" how many specified colour blocks there are
-
Enumerate total number of objects
- Number of “blue” blocks plate surrounded by blocks of other colour
- Specify colour of largest or smallest object
- Suggesting he had sense of greater or less than as well.
Animals in Numbers: Desert Ants
- Judge distance by counting steps
- Legs are clipped they undershoot the journey
- Given stilts, the go too far when returning from foraging
Animals in Numbers: Lions
Lions decide to attack based on the ratio of their numbers compared to the number of voices in an “intruder” pride
Animals in Numbers: Chimps Ayumi
- Number tasks that exceed the ability of humans
- When length of visible number was reduced, Ayumni could still perform while the kids performance dropped
- Superior performance on this task may reflect additional processes in addition to number
- VSTM required exceeds the 4-5 items limit of humans
Evidences that preverbal children could distingush numbers. Three studies
Numeriosity:
Preferential looking task on dot displays that differ. Reflects the novelty
- Babies looked longer at displays that had change in number 2-2 (1.9s) vs 2-3 (2.5s)
- 6mo babies can distinguish between 8 and 16 and 16 and 32 (1:2 ratio)
Arithemetic
- Babies understood 1 + 1 = 2. Stared longer when 3 dolls appeared instead of 2
What is the evidence for concept of number change over time?
Method, Results, Conclsusion
Number Line Task
Method
- Place a vertical line at the location where the above number would appear between the given number range
Result
- Children often overcompensate for the position of the given number
- Non-linear (Logarithmtic) to linear change
- Early representation represents ratio differences rather than linear separation
Conclusion
- Movement from logarithmic to linear representation over time may reflect formal education
Evidence that cultures with limited capacity for language had a sense of number
Similar number line task: Pica (2004)
- Munduruku have words that go up to 5
- Beyond 5 reflects the approximate number
- Smaller numbers are precise, larger numbers are less precise
- Progressively smaller intervals
- Logarithemtic
- Did not use numerals in a counting sequence
- Did not use numerals to refer to precise quantitites
Evidence that words are not necessary to understand exactness from cultures with limited capacity for language
Method
- Taps on wood up to seven times and compared with counters were placed on a mat
- Sometimes the number of taps matched the number of counters, sometimes not
Results
- Children had no words for the numbers four, five, six and seven, yet were perfectly able to hold those amounts in their heads
- Abstract enough to represent both auditory and visual enumeration.
What does numerical competence require one to do?
- Identify
- Order
- Compare
Numerical quantities
What are two tasks examining basic cognitive processes of number
- Enumeration
- Verbalize precise number
- Non-Symbolic
- Number comparison
- Magnitude comparison
- Larger or smaller
- Does not necessarily require verbalization
- Non-symbolic/symbolic/cross-modal
- Magnitude comparison
Object Enumeration. What are the performance measures and steps
Performance Measures
- IV: Random spots/dots
- DV: Accuracy and RT
Steps
- Symbolic encoding of visual information
- Accessing that symbolic representation for combining into a total or sum
Object Enumeration. What are the results?
Evidence for 2 counting mechanims for dot enumeration
- "Sibsitising" Number Range
- Set Size <4
- Rapid and Accurate
- "Approximate Number System"
- Set Size >4
- Slow And Approximate
RT curve shows "elbow" at the 4ish number
Number comparison: What are the types
- Symbolic (8+3)
- Non-Symbolic (Dots)
- Cross-Modal (8 + Dots)
Number comparison: Does it require a verbal response
Not necessarily
Number comparison: What are performance measures and steps
Performance Measures
IV: Non-Symbolic/Symbolic/Cross-Modal
DV: Accuracy and RT
Steps
Compare the magnitude of two regions of visual information, not necessarily verbalized
- Symbolic encoding of visual information
- Symbolic mapping onto numerical information
- Accessing that symbolic representation for combining into a total or sum
Number comparison: What are the results of the symbolic and non-symbolic task
Symbolic
- Number distance effect
Non-Symbolic
- Weber's Ratio
Number comparison: What are the results. Symbolic Elaboration
Number Distance Effect
Slower RT and Less Accurate for numbers closer in numerical distance
- Suggests that neural mechanisms are ordered in a functional way
- Supports mental number line
Number comparison: What are the results. Non-Symbolic Elaboration
Weber's Ratio
- Errors depend on the ratio of the magnitudes
- Smaller Weber ratio implies Higher sensitivity to ratio differences
- Weber ratios for numerical similarity are linear on a log scale
- Imprecise in subsitising range
What do Weber's Ratio on a log scale look like. Just like...?
- Linear (width are the same)
- Similar to log number spacing for children or aboriginal cultures on the number line task
Is Weber Ratio related to arithmetic competence. What does it additionally suggest?
Yes, it is correlated with school arithmetic competence.
- Smaller weber's ratio predictive of mathematical ability
-
Children with dyscalculia are less accurate in comparison of two sets of dots compared with age-matched controls
These studies suggest that number comparison involves access to numerical magnitude representations that form the basis of arithmetic
What is the problem with Weber's ratio and correlation with arithmetic performances
- Large individual differences in the data
- Relationship may be complicated
- Easy items may require formal calculation skills
- Studies of over-practiced indices of maths competence (adding two single digits) result in ceiling performance for youngest children
- Influence of experience would blur the real relationship
Cross model number comparison: What is the task
Cross modal of dots and tones.
Cross model number comparison. Results? and Conclusion?
Results
- Cross modal number as accurate as unimodal
- Little/ No accuracy cost for comparing numerosities across stimulus format or modality
- But cross-modal is slower than unimodal in general
Conclusion
- Number comparison is a general brain mechanism, not a visual brain mechanism
- Judgements of approximate number is a due to a representation of number
Enumeration vs. number comparison. Do they involve same mechanism?
Enumeration
- Counting
- Eye Movement
- Attention shift
- Increase in RT may be due to shifting attention rather than eyes although people might shift eyes to different regions of subatisable dots
Comparision
- Can be rapid
- Precedes eye movement and attention shift
- However, subsitising limit is close to the point weber's ratio makes discrimination difficult
Brain areas implicated from primate studies for number processing? Include the "other areas"
- Intraparietal sulcus (IPS)
- Lateral prefrontal cortex (PFC)
"Other Areas"
- Superior parietal lobule (SPL)
- Ventral intraparietal area (VIP)
The areas associated with acquired acalculia
- Parieto-Occipital junction
- Frontal lobes
3 types of number neural coding found in single neurons
- Numerical Quantity
- Useful to estimate number
- Compute Exact Number
- Proportional Representations
- Judgement of relative size
- Not necessarily symbolic
- Symbolic Numbers
- Communication
- Language
What task do we usually do in primates
Delayed match-to-sample task
Delayed match-to-sample task. What are the steps involved in primates
- Fixation (By food)
- "Sample" spots flashed
- Delay
- 50% match/ 50% no match
- If match, release lever for food
What are some properties of numeriosity tuned neruons found for delayed match-to-sample task.
1.) Where are they found
2.) How are they distributed
3.) What is the latency
4.) Specificty
- Numerosity tuned neurons abundant in the lateral PFC and IPS
- Number selective cells are distributed, not clustered
- Average latency of IPS neurons shorter than PFC
- Responses in IPS precedes PFC
- Suggest numbers are processed in IPS (Attention) before PFC (control)
- Neurons are not specific to a single number, but they have a preferred stimulus tuning
Number tuning responses of neurons in the intraparietal sulcus
Clear evidence of stimulus dependent tuning
- Preferred number tuning
Evident at different times, for different neurons
- Some for sample
- Some for delay
Responses depend on numerisoity than features
Number tuning responses of neurons in the prefrontal cortex
Clear evidence of stimulus dependent tuning
- Preferred number tuning
Transient or Sustained
Responses depend on numerisoity than features
Average responses in PFC consistent with behavioural performance.
1.) What kind of distance effects do neurons show
2.) What is the average tuning depending on task
3.) What is the number tuning qualitatively similar to
4.) Can we predict behavioural from neuronal pattern?
- Average number tuning shows distance effect
- Weber's Fraction
- Number tuning is not task dependent
- Average tuning is the same for sample and delayed periods
- Number tuning of neurons is qualitatively similar to behavioural performance
- i.e., poor tuning = poor performance
- Bandwidth of behavioural > neuron
- At a neuronal level, single neurons perform better than a monkey performs behaviourally
Are neural responses are based on stimulus number rather than other stimulus features
Yes.
- Despite changing feature properties (area, size, density), the number tuning drop off is similar
-
Suggests these neurons were not tuned to properties apart from numerosity
Did the presentation of dots in simultaneous or sequential (basic arithemetic) positions made a different in neural response?
- Many neurons showed response in either
- Tuning for simultanenous but not sequential
- Tuning for sequential but not simultaneous
Non-symbolic number tuning in neurons: How does it mirror Weber's Law and the Log Number Line?
Weber's law
- Numerosity not coded by neurons in an exact manner
- Imprecision increase in proportion to magnitude
Logarithmic number line
- Symmetric number tuning functions are only obtained after log transformation of the number scale
- Coconsistent with Fechner’s law, which states that the perceived magnitude is a logarithmic function of stimulus intensity
- Found up to 30 items for non-human primates
Do the IPS and PFC have specific subregions that are exclusively for number coding? What does it suggest?
No single, isolated cortical region
- Within IPS and PFC, activation do not cluster.
- Neurons with object size and other properties intermingled with number tuned neurons in the same vicinity
- Suggest that number neurons are part of a more generalised analysis
Do neurons code for proportions/continuous magnitude? What is the study and results?
Task: Compared number tuning and line length tuning
- Some (20%) of anatomicaly mingled IPS neurons encode either line length ratio, numeriosity, or both.
- Interminged
- No one region where it is "line-selective" or "numeriosity-selective"
- Not clustered together
- Proportionals might have evolutionary benefits
Do non-human primate neurons show symbolic mapping onto numerosity? Where?
- Relatively large proportion of PFC neurons respond to symbolic form of number (dots into number)
- PFC encodes dots and visual sign as numerical values
- PFC for abstract associations
What is symbolic tuning similar to (neurons)
Symbolic number distance effect
- Drop-off with increasing numerical distance from preferred numerical value
- Trial by trial activity of neurons correlated with the monkeys’ performance
-
Neural responses were reduced when monkeys failed to match the correct number of dots to the learned signs
-
-
Suggests that neurons responded to abstract numerical value rather than visual shape or dot pattern
Implications of symbolic tuning for neurons in PFC and IPS
Both IPS and PFC non-symbolic
But only PFC symbolic of the IPS-PFC framework is engaged in semantic shape-number associations in symbol-training monkeys
- IPS
- Non-Symbolic
- Does not seem to be associated with symbols (Only 2%)
-
Quality of neuronal association in the IPS was weak and occurred much later during the trial (could have been feedback).
-
- PFC
- Semantic mapping between signs and categories
What are the brian areas implicated in number processing in humans.
Is humans found first or animals?
- Parietal and Frontal Areas
- Maybe ITG
- Humans predated Animals. Human studies before animal studies
What brain areas does symbolic stimuli activate.
What do symbols include
Brain Area:
- IPS
Symbols
- Culturally learned symbolic notations such as Arabic numerals or spelled-out or spoken number words
- Modality does not matter
-
Parietal activation occurs for an abstract, amodal representation of numbers
What brain areas does non-symbolic stimuli activate. What do symbols include. What are some conditions for activation.
- Brain Area
- IPS
- Non-symbolic stimuli
- Dots and Tones
- Conditions
- Attending to dot stimuli: Obligatory processing
- Passively viewing: IPS adaptation to number
What brain areas does both symbolic and non-symbolic stimuli activate?
Combined symbolic and non-symbolic presentations
IPS and PFC suggest common representation
Adaptation to Number Symbols: Study Overview and Results
Naache et al. (2001) fMRI adaptation study
- Rapid presentation of repeated/numerically distant numbers after adapting to a number
Results
- IPS of both hemsiphere
- Repeated number
- Lesser distance effect, BOLD Response reduced
- Distant number
- Greater distance effect, BOLD Response increased
- Repeated number
[Symbols = Number Distance Effect]
Adaptation to Dot Numeriosity: Study Overview and Result.
Piazza et al. (2004) fMRI adaptation of dots
- Rapid presentation of repeated/distant dot after adapting to a dot
-
Brief tests at a range of ratios relative to the adaptation numerosity
Results
- Symmetric Gaussian tuning on ratio scale - just like single cells in monkey studies
- Precision of coding consistent with Weber Fraction
- [Dots = Weber]
Cross-notation adaptation: Study overview and result
Piazza et al. (2007): fMRI adaption of dots and numbers:
- Adaptation: 17,18,19 dots with arabic numerial 20 or 50
Results
- Distance effect
- fMRI recovery if 19 vs 50
- no fMRI recovery if 19 vs 20
- Both IPS and PFC
How early does the IPS become specialised for number?
1st Studies Overview and results
Cantlon et al. (2006) fMRI adaptation study on 4 yo
- Changed
- object numeriosity or object identity
Results
- Numeriosity
- Right Parietal Cortex
- Identity
- Occipito-Temporal Cortex
How early does the IPS become specialised for number ? 2nd study overview
IIzard et al. (2008) Event-related potentials (EEG) from 3 mo
- Presented with a continuous stream of sets of objects
Results
- Numeriosity
- Right Parietal Cortex
- Identity
- Left Occipito-Temporal Cortex
Adapation to Proportions (Line). Study overview and results
Jacob and Nieder (2009)
- Habituated to a given line length proportion
- BOLD response larger the more different the ratio
- Both IPS and PFC
Adapation to Proportions (Fractions). Study overview and results
Jacob and Nieder (2009)
- Habituated to a given fraction
Results
- BOLD response larger to more different the fraction
- Fractions are the symbolic version of ratio
- Symbolic Number Distance Effect
- IPS
What is the role of the frontal lobes in numbers from childhood to adulthood?
- Frontal to parietal shift from childhood to adulthood as symbolic processing becomes automatic
- Prefrontal regions
- Essential for nuumbers calculation
- Related to task difficulty (Formal operations leverages on frontal areas)
- Superior frontal gyrus involved in complex calculations
What did ECOG study revealed in number. Another area? What is the implication
Inferior Temporal Gyrus
ECoG high frequency band (65-150Hz):
- Revelead neurons in ITG with a preferential response to visual numerals (not lines, curves, angles)
- Additional area for symbols
- We might have developed a special region for symbols
- Not evident in BOLD fMRI due to signal dropout
- Functional link with lateral parietal cortex (LPC) while doing arithmetic
What is Dyscalculia. What are some properties. Prevalence. Consequence. Co-morbaility. Persistence. Heritability
- What:
- Specific and Severe disability in learning arithmetic
- Normal intelligence
- Normal working memory
- Prevalence
- 5-7% (developmental dyscalculia = developmental dyslexia)
- Consequence
- More severe than dyslexia
- Co-morbility
- Occurs with other developmental disorders like ADHD and Reading
- Persistence:
- Persists into adulthood (Unclear whether it's a delay in ability or deficit)
- Heritability
- Mathematical abilities have high specific heritability
What difficulties do dyscalculics have?
- Simple arithmetic
- Deficit for even the most basic representation of numerosities
- Enumeration and Number Comparison
What is the neural activation behind dyscalculias
- Reduced grey matter in left IPS adoloscent
- Reduced grey matter in right IPS 9yo
- Reduced probability of connections from right fusiform gyrus to other parts of the brain, including the parietal lobes
- Evidence is inconsistent
fMRI behind developmental dyscalculia
- Huge individual differences
- Could be over/under activation
- Unclear
Dyscalculia treatment:Does neuroplasticity/teaching allow brain changes that improve maths?
Teaching
- 1:1 cognitive tutoring improved performance for children with mathematical learning disabilities
- Tutoring reduced overactivation in areas, normalising brain activity
Neuroplasticity
- Recent studies via. direct stimulaton suggest usefulness
