Social behaviour and cognition

Question 1

Parts of the brain active in social tasks

Answer 1

Sallet et al., 2011 - grey matter volume of rostral PFC in macaques correlates with social network size. The homologous area in humans is associated with theory of mind.
Amodio and Frith, 2006 - medial frontal cortex active during social encounters, specifically when mentalising
Bickart et al 2010 - amygdala volume in humans correlates with social group size
Noonan et al 2014 - relative amydala volume in macaques correlated with social dominance

Question 2

Social brain hypothesis, evidence for

Answer 2

Brothers, 1990 - there are parts of the brain specialised to process information solely in the social domain. It’s a specialisation for processing the dispositions and intentions of others.
Dunbar and Shultz 2007 - neocortex ratio increases with group size, with a grade shift between monkeys and apes.
West, 2014 - Brain size in birds is correlated with social, not genetic, monogamy

Question 3

Alternative to social brain hypothesis

Answer 3

Social information is processed using the same tools and brain parts as other information
Izuma et al 2008 - The same voxels in the striatum were active when processing a monetary as a social reward

Social information is ‘mapped’ in the hippocampus
Tavares 2015 - subjects played a role-playing game where hippocampal activity correlated with movement through a 2D social space framed by power (sub/dom) and affiliation (sharing info or touch) The correlation was stronger in subjects reporting higher social skills.

Question 4

Face cells - discovery, evidence for

Answer 4

Gross et al 1972 - some neurons of visual association area TE responded better to complex coloured patterns e.g. faces
Rolls et al - in macaques, some neurons respond mainly to faces, and some respond differently to different faces.
Tsao et al 2006 - Face Patches include the temporal lobe, OFC, lateral PFC. 97% of cells in these patches are face-sensitive. If you activate one face patch, others will become active. You can read out a face’s identity from face cell activity with high accuracy, but only low for other types of stimuli.
Gothard et al 2007 - there are face identity cells and face expression cells, both exist in amygdala

Question 5

Face cells - coding property, advantages of

Answer 5

Sparse ensemble coding rather than gnostic ‘grandmother cell’ coding
high coding capacity, increases exponentially with no. of neurons (so efficient)
high accuracy and resistance to noise
easy pattern completion
decays gracefully (removing one neuron makes little difference)
correct coding strategy for pattern association learning
allows generalisation/invariance (can use a low contrast image)
Faster/easier readout by downstream cells

Question 6

Face cells and eyes

Answer 6

Mosher et al 2014 - In macaques, some face cells in amygdala were most active during mutual gaze.
Adolphs et al 2005 - Patient SM - Urbachs-Wiethe disease, spent less time looking at eyes of face photos, when told to look at eyes was better at detecting fearful expressions
Rutishauser et al 2013 - epilepsy patients (because they already had electrodes implanted) with autism spectrum disorder looked at eyes less, had less amygdala activity to faces, and amygdala activity to face parts correlated with behaviour (looking at those parts).

Question 7

Amygdala role in social cognition - for

Answer 7

Janak and Tye 2015 - Amygdala well conserved, but ratio of BLA to CeN size increases up the phylogenetic tree
Kluver-Bucy syndrome 1939 - lesions of amygdala induced tameness, hypersexuality, inappropriate eating, visual agnosia
Weiszkrantz - suggested that amygdala lesions impair identification of reinforcing stimuli, positive or negative
Toscano et al - Lesioning amygdala of female monkeys before or after birth reduced interest in others’ offspring. Lesioning hippocampus had a similar effect but only before birth
Moadab et al - Lesioning amygdala (but not hippocampus) reduced grooming and social contact

Question 8

Amygdala role in social cognition - against

Answer 8

Kluver-Bucy syndrome not apparent with more selective lesions
Machado et al - amygdala lesions increased contact initiation, affiliative behaviours, approach behaviours, aggression initiation
Amaral et al 2003 - amygdala lesions did not impair social behaviour, though did decrease inhibitions

Question 9

Potential reasons for heterogeneity of findings in monkey amygdala lesion studies

Answer 9

How large were the group sizes the monkeys were kept in?
How much handling by the researchers did they have? Reared by mothers, or by peers/humans?
Were they family groups or mixed? I.e. complexity of social environment
Lesioning by cutting or injections?

Question 10

Amygdala and threat detection - Amaral et al 2003

Answer 10

Lesioned adult male macaques with ibotenic acid.
They showed lower social inhibitions (approached new monkeys without the usual period of evaluation)
Those new monkeys did not shun them, but gave more affiliative behaviour towards them, and they received this

Also lesioned 2 week old macaques, with a protocol to allow normal socialisation and returning to mother.
Reduced intrinsic fear (e.g. of snakes), but increased social fear. What structure was mediating this, if the amygdala was all gone?

Caveats: the early lesions may have allowed brain reorganisation, so redundant systems may have taken over. Reversible lesions e.g. optogenetics would be useful.

Question 11

Amygdala and threat detection - other evidence for

Answer 11

Patient SM required much less personal space, e.g. when allowed to approach to whatever felt most comfortable would go to 34cm away. She understood the concept of personal space, and that others required more than her. She was also comfortable with full eye contact, and did not recognise negative social cues.
Kennedy et al 2009 - amygdala of humans is more active in fMRI when the experimenter is standing close by
Tillfors et al 2001 - increased blood flow to amygdala in social phobics when they think about giving a presentation

Question 12

Mirror neurons in monkeys

Answer 12

Originally recorded in area F5 of premotor cortex
Fire when doing an action, watching an action, or doing an action in the dark
Umilta et al - Firing dependant on the goal of the activity, not the specific motion necessary (pliers and reverse pliers to grasp a peanut).
Kohler et al - Firing is multisensory - happened when breaking a peanut, watching a peanut being broken, and hearing a peanut being broken, but different pattern for ring grasp task
Firing is not a result of covert muscle contraction - electrodes on muscles in arms

Question 13

Mirror neurons in humans

Answer 13

fMRI data shows overlap between observation and execution of a task, in same areas as monkeys, but fMRI is low resolution so commonly activated areas does not mean commonly activated neurons.
Mukamel et al 2010 - used epilepsy patients with intracranial electrodes to get single cell recordings. Found mirror neurons in entorhinal and supplementary motor cortex

Question 14

Why/how do mirror neurons exist

Answer 14

Evolutionary adaptation for ‘action understanding’ - but what exactly is this? No evidence to correlate mriror neuron activity to any behavioural competence

Produce of associative learning - Cook et al said ‘sensorimotor learning changes mirror neurons’, but his only data was from fMRIs, no single-neuron stuff.

Empathy

Question 15

There is empathy if:

Answer 15

De Vignemont and Singer
1 - one is in an affective state
2 - that state is isomorphic to another’s affective state
3 - state is elicited by observation/imagination of that other person’s state
4 - person is aware that the source of own state is the other person.

Question 16

Neural areas associated with empathy (3)

Answer 16

Wicker et al 2003 - insula responds to both experience and observation of odour-induced disgust
Singer et al 2004 - insula and anterior cingulate activity to own and loved one’s thermal pain. More activity in more emphathic subjects
Singer et al 2004 - In prisoner’s dilemma, the degree of empathy-related activity was reduced when viewing a player you perceive to be unfair. Reduced much more in men than women!
De Vignemont and Singer 2006 - activity in insula and anterior cingulate when viewing different types of pain

Question 17

Potential mechanisms of empathy in associated areas

Answer 17

Insula has been linked with interoception, pain sensation, converting bodily sensations into social emotions

ACC activity reflects various subjective feelings
Rudebeck et al 2004 - Lesioning ACC reduces social interest in macaques (they’re less distracted from food by conspecifics)

Allman et al 2011 - in animals we think of as social (humans, dolphins, elephants, dogs), spindle cells are found specifically in insula and ACC

Question 18

Prosocial behaviour, neural correlates

Answer 18

Hare et al 2010 - subjects gave freely determined or forced donations to charities of their choice. VMPFC activity signalled subjective donation values - same region as value of food rewards. [BUT this area is also crucial in decision making, as shown by Bechara et al, and in SCR to aversive stimuli.] ACC and insula activity correlated with VMPFC activity.
Chang et al 2013 - a monkey will always give a reward to themself rather than the other, but will give reward to other rather than neither of them. Single neurons in the ACC were active when the ‘actor’ monkey gave the reward to the other, at the choice and reward stages
Báez-Mendoza, Harris and Shultz - Monkeys observed each others’ choices and actions. A subset of striatal neurons were activated when either own or conspecific’s action resulted in own reward (and more activity for the latter). This activity was reduced when the conspecific was replaced with a computer.

Question 19

Predicting others’ behaviour - monkeys

Answer 19

Grabenhorst et al - monkeys learn faster from observation than from own experience.
-amygdala neurons signal reward value (shared code for own reward and conspecific’s reward)
-amygdala neurons signal whether self or conspecific is making the choice (i.e. will only have a different signal for A vs B when self is choosing, or when other is choosing).
-amygdala neurons signal which object should be chosen
-amygdala neurons signal reward value first, then which object should be chosen, all for either self or for conspecific
So single amygdala neurons learn reward value from observation and experience, and use this to simulate decisions of others

Haroush and Williams 2015 - monkeys in prisoner’s dilemma were more likely to cooperate if their partner recently or frequently cooperated. Cingulate neurons predicted the other’s unknown choices, and stimulation of cingulate disrupted cooperation.
So single ACC neurons predict partner’s decisions

Question 20

Predicting others’ behaviour - humans

Answer 20

Hampton et al 2008 - humans in MRI scanner played a competitive game, as either employee or employer, with choice to work/not, and inspect/not. Three potential strategies were hypothesised - Reinforcement (‘what did I do last time?’), Fictitious (‘what did my opponent do last time?’) and Influence (‘how does my choice affect my opponent’s choice?’). Influence was the best fit to the data.
Activity in medial frontal cortex was higher in subjects that used the ‘influence’ strategy more.

So human medial frontal cortex signals high-level strategy use to predict opponent’s behaviour

Question 21

Humans learning from others

Answer 21

posterior Superior Temporal Sulcus - distinguishes biological motion (i.e. an agent) from non-biological motion

Gaze-following - babies and adults do it, and many other animals. Adults do it persistently, even when the person whose gaze they’re following persistently looks the wrong way. They’re unaware of the contingency, but they rate that person ‘less trustworthy’.

Imitation - babies will imitate both instrumental and arbitrary actions of an adult. Mongoose pups did this too, learning to break into modified Kinder eggs.
When we’re being covertly copied, we tend to like the copier more, and be more helpful to others generally. When we’re being overtly copied, we feel uncomfortable and the prosocial effects don’t occur.

Learning trust - pre-verbal babies prefer characters who help others to those who hinder others.

Bayliss et al 2006 - people prefer objects that others are looking at, over those that receive no attention.
Laland 2004 - as long as our own knowledge is sufficient for success, we don’t copy. When it becomes unreliable, we copy. When that becomes unreliable, we innovate.