Lecture 4 - Building Blocks and Features of Emotions Flashcards
Building blocks
versus features
- Emotions are adaptive functional states
that are at a level of complexity in
between reflexes and volitional control - Emotions are evolved packages of
functional adaptations that are more
constrained than volitional control but
more multidimensional and flexible
than reflexes. - We can define properties of emotions
including ‘building blocks’ and ‘features - Building blocks
- essential, basic properties of emotion
- shared by all or most specific emotions
- present in precursors to full emotion
states in simpler organisms - Features
- more elaborated and variable properties
of emotions - not shared by all emotions
- For example, in a car, wheels are
building blocks while air conditioning is
a feature - Emotion building block
Valence
All emotion states share a quality of
evaluating good or bad, pleasant or
unpleasant, approach or avoid. - Emotion feature
Social communication
Very prominent in mammals but likely
recently evolved and not present in all
animals
A provisional list of emotion properties
The division of building blocks and features is not black
and white. It is instructive, not absolute.
Emotion properties are the processing features that define
emotion states (i.e. the things that we look for in the brain
to discover an emotion states)
We can put together a provisional* list of operating
characteristics of emotion states that are essential to
carrying out the functional role of emotion to begin to
illustrate how we can investigate emotions in general
*(i.e. not complete, there could be others, some could be
removed, this is not ‘truth’)
Scalability
Valence
Persistence
Generalization
Global coordination
Automaticity
Social communication
Higher up scale = building blocks, lower = features
We can use these properties to differentiate
emotion states from reflexes
We can also use these properties to
characterize a specific emotion state and
differentiate it from another emotion state
We can think of emotions in dimensional space
Scalability
Scalability: An emotion state can scale in intensity. Importantly, parametric scaling can result in
discontinuous behaviours, such as the transition from hiding to fleeing during the approach of a
predator. Intensity is often conceptualized as arousal, although these two are not the same thing.
- Emotion states are commonly classified based on
valence (positive or negative) and intensity - Different emotions have different levels or
intensity e.g. sadness, rage - In psychological models, scalability is often
incorporated as arousal - Scalability differentiates from stimulus-response
reflexes which tend to be all or nothing
responses - It is not yet clear if the intensity of emotion is
inherent to the mechanism of a specific
emotion or if there might exist some kind of
general arousal system for emotion - Differences in intensity could be graded or
qualitatitive - A graded increase in intensity could be
observed by increasing vigor of the same
behaviour e.g running from threat - Gradations in emotional intensity can also have non-linear effects
on behaviour - For example, the threat imminence continuum of defensive
behaviour in rodents and octopi
Valence
Valence: Valence is thought by many psychological theories to be a necessary feature of emotion
experience (or ‘affect). It corresponds to the psychological dimensions of
pleasantness/unpleasantness, or the stimulus-response dimension of appetitive vs. aversive (but
again, these two are not the same thing).
- Good- bad, pleasant- unpleasant,
appetitive- aversive - Darwin talked about ‘antithesis’
Emotions come in pairs of opposites which are
expressed by physical opposite and
complementary behaviours.
This could be important for social
communication functions
Ex: dog hostile vs. dog affectionate. Position of body is the opposite.
Same with fruit flies
Persistence
Persistence: An emotion state outlasts its eliciting stimulus, unlike reflexes, and so can integrate
information over time, and can influence cognition and behaviour for some time. Different emotions
have different persistence. Emotions typically persist for seconds to minutes
- Emotions often outlast the stimulus that
elicited them - This is a key feature that distinguishes
emotions from stimulus response reflexes - For example, fear has a long-lasting effect
on behavior: heart rate, stress hormone
levels, breathing rate etc. remain elevated
for some time after encountering a threat - Persistence makes emotion states
flexible - A persistent emotion state allows for
integration of different kinds of
sensory information over time which
may be important for neural
computation and action selection - Persistence also allows emotion states
to interact with other other internal
states and powerfully influence
cognition and behaviour
In Drosophila (fruit flies), air puffs
cause a persistent state of
increased movement
In Drosophila, brief
optogenetic stimulation to
activate specific neurons
leads to courtship wing
extensions that lasts for
minutes
- Different emotion states tend to persist for
different amounts of time - E.g. surprise or joy vs sadness
- Emotion states that persist for hours, days
or longer are ususally classified as moods - Persistence seems to be independent from
memory and consciousness - Amnesic patients still experience
persistent sadness after watching a sad
film, even though they don’t remember
having seen the film (Feinstein, Duff &
Tranel, 2010)
Side Note: Moods
* Emotions generally do not persist for very
long after the situation that triggered the
emotion has been resolved
* Moods are emotion-like states that last
much longer than emotions (hours-years)
* Moods may be more prominent in humans
than animals
* Emotions function to cope with present,
acute situations
* Moods may function to cope with events in
the past or future
* May not just be emotions on a longer time
scale
* Moods often have no clear trigger
* Often involve effects on cognition more than on
behaviour
* Moods are similar to emotions in having
dimensions of persistence, scalability, valence,
generalization and automaticity
* Moods are distinct from emotions in not
serving a clear social communication function
Generalization
Generalization: Emotions can generalize over stimuli and behaviour, much of which depends on
learning. This creates something like a “fan-in”/ “fan-out” architecture: many different stimuli link to
one emotion state, which in turn causes many different behaviours, depending on context.
Persistence and generalization underlie the flexibility of emotion states.
- Because of the property of
persistence, an emotion state induced
by one stimulus can generalize to a
different context and influence
responding to different stimuli - This is context generalization or
trans-situationality - Context generalization allows
emotions to bias cognition and
behaviour - Context generalization is another way
that emotion states differ from
reflexes - Applying this criterion can distinguish
between a behavioural response
mediated by a simple reflex and by a
persistent internal state that
generalizes to other context to
influence subsequent behaviour - Stressed honeybees show a negative
bias in a test of ambiguous odor cues
(Bateson et al 2011)
Bees that were shaken (stressed) avoid the ambiguous odour. Bees who weren’t shaken (stressed) don’t avoid the ambiguous odour. - This suggests that the stress
manipulation induced an internal state
in honeybees that influenced their
behaviour in other contexts - Stimulus generalization &
pleiotropy contribute to other
aspects of generalizability of
emotion states - Many stimuli ‘fan-in’ (stimulus
generalization) to cause an
emotion state which can then ‘fanout’ (pleiotropy) to cause many
effects - The same behavioral expression
can be triggered by many different
stimuli, including those for which
the behavior appears to serve no
useful purpose, if those stimuli
evoke the same internal emotion
state - This is stimulus generalization in
action - Darwin’s example : cats kneading
their paws on a blanket - This behaviour serves to stimulate
milk flow from a nursing mother but
serves no purpose in the blanket
example - Darwin argues that it became
associated with the same state
(“pleasure”) either through habit
(learning) or ‘inheritance - Emotion states are pleiotropic
i.e. they have multiple, parallel
effects on behaviour, body,
cognition - Simple reflex responses
generally don’t induce
multidimensional responses - For example fear induces
defensive behaviours (freezing,
fleeing) as well as endocrine
changes (stress hormone
release), autonomic changes
(heart rate, blood pressure,
sweating) and cognitive changes
(attention & memory)
Emotions &
Learning
* Stimulus generalization is
closely linked to learning
* Most stimuli that cause emotions
gain this property through
experience (i.e. associative
emotional learning)
* The best understood example of
this is Pavlovian conditioning
* In Pavlovian fear conditioning, through
presentation with a foot-shock, a
previously neutral cue comes to elicit the
same behavioural response (freezing) as
the shock itself
* All species studied show associative
emotional learning
* Some species also show learning through
observation without the need for direct
experience
* Humans can also learn from being told
about things
* Learning is a key mechanism that
increases stimulus generalization
- Emotions also have some ‘domain specificity’
- There is a restricted range of stimuli or
circumstances that can cause an emotion state
and some stimuli can be more readily learned
about than others - For example, tastes readily come to elicit
disgust but is much less likely that a tone will
come to elicit disgust - Domain specificity distinguishes emotions
from volitional control - Associative learning can also be used to test if a
stimulus induces an emotion state - Conditioned place preference pairs a neutral stimulus
(one half of a box) with something potentially
rewarding or aversive (e.g. an injection of a drug). In
a later test session, where the animal spends time is
interpreted an index of any internal state induced by
the stimulus - If the state is rewarding, animals will spend more
time where they encountered it - If the state is aversive, animals will avoid the location
where they encountered it
Global coordination
Global coordination: Related to the property of generalization is the broader feature that emotion
states orchestrate a very dense causal web of effects in the body and the brain: they engage the
whole organism. In this respect, they are once again differentiated from reflexes.
- Emotion states causally interact with
other internal states to a large extent - Emotions influence behaviour &
cognition as well as endocrine and
autonomic responses (pleiotropy) - Emotions evolved to deal with
challenges that required a whole body
response - Outputs of emotion states need to be
cohesive and to achieve this they need
to be coordinated - Co-ordination is a global feature of
emotions and a property to look for in
the brain - Also differentiates emotions from
reflexes - There are different ways this could
occur
e.g. anatomical projections to different
downstream targets - E.g. Projections from the central
nucleus of the amygdala to brain
stem and hypothalamic nuclei
mediate different components of
the fear response - This is likely too simple to be a full
explanation of co-ordination
because emotion states are
almost certainly more distributed
than one single brain region - Also, different subsets of
responses are seen on different
occasions and/or on different time
scales - Synchronized oscillations across
networks of brain regions could be
another potential mechanism of coordination - For example, freezing associates with
a brain state of synchronized 4Hz
oscillations in prefrontal cortex and
amygdala (Karalis et al., 2016) - There are multiple system architectures
that could achieve co-ordination - Distributed systems could control
individual components with
synergistic/antagonistic interactions
between components - Centralized systems with a single
command neuron could execute a range of
responses - Likely the brain uses multiple solutions to
co-ordinate responses to emotion states - The co-ordinated control exerted by
emotion states is distributed in time and
space - Emotion states often involve a large
time range of sensorimotor processing - E.g. shrinking back from an attacking
bear vs planning how to escape from a
bear that is still some distance away - This is extraordinarily complex
because it is distributed in space and
time but also the many components of
the response interact with each other - Much remains to be understood about
how this occurs
Automaticity
Automaticity: Emotions have greater priority over behavioural control than does volitional
deliberation, and it requires effort to regulate them (a property that appears disproportionate, or even
unique, in humans).
- Somewhat like reflexes, emotion states
exhibit automaticity over behaviour i.e.
no effort is required to elicit the
behaviour - It is generally effortful to inhibit the
behavioural response - Emotions could be thought of as an
‘interrupt’ mechanism for prioritizing
urgent/important needs - Control of emotions most commonly observed
in adult humans - In young children and animals, emotion seems
to exert a larger control on behaviour - Emotion regulation through conscious control
may be largely unique to adult humans
Emotion Regulation
* The ability to have some degree of control over your emotion state
* Regulation could occur at multiple levels:
at the point of inducing the emotion state
E.g. choosing circumstances and environments that will influence if and how an emotion is induced (e.g.
avoiding taking a class that has an oral presentation to avoid experiencing fear of public speaking)
reappraising the stimulus
E.g. Internally reinterpreting a situation that could induce an emotion (e.g. A friend ignores you when you
say hello to them in the hall, you convince yourself they didn’t hear you.)
directly trying to control the experience or expression of the emotion state
e.g. telling yourself to stop feeling sad
* Disrupted emotion regulation is
implicated in a range of
psychiatric disorders
* For example, PTSD, phobias,
depression
* Cognitive-behavioral therapies
develop strategies to re-establish
cognitive control over one’s
emotions
* Emotion regulation involves the prefrontal
cortex
* The prefrontal cortex is one of the last brain
regions to develop and plays a major role in
emotion regulation
* There are also substantial species
differences in prefrontal cortex and it
is largest and most elaborated in
humans. This could be relevant to
understanding differences in emotion
regulation.
Social communication
Social communication: In good part as a consequence of their priority over behavioural control,
emotion states are pre-adapted to serve as social communicative signals. They can function as
honest signals that predict another animal’s behavior, a property taken advantage of not only by
conspecifics, but also predators and prey.
- Because emotional behaviours
are difficult to control, they can
serve as authentic social signals
about an individual’s emotion
state - Emotional behaviours are poised
to be co-opted as social
communication signals - We can infer something about
another person or animal’s
emotion state from their
behaviour - But are we always right?
- Volitional control over emotional
expressions leads to the possibility of
deception and manipulation - In humans, facial expression may
have evolved from emotion
behaviours to also serve as social
communication signals - Facial muscles are controlled by a mix of
volitional control and automatic controls - We have more volitional control over the
lower half of our faces - We can see this in the difference between a
smile elicited by a genuine emotion and a
‘fake’ smile - Facial expressions are used in social
communication but they can be very complex
and ambiguous to interpret - E.g. People smile in different circumstances &
for different reasons (anxious, happy,
submissive) - Cultural differences in the meaning of facial
expressions and when it is appropriate to
display them - It is challenging to reliably link facial
expressions to a specific human emotion
A dimensional space for emotions
‘Core affect’ is commonly represented in 2
dimensional space with axes defined by
arousal and valence (e.g. Russel, 1980)
Arousal: not aroused, aroused
Valence: unpleasant, pleasant
We can locate any emotion within this 2D
space
The proximity/distance between two
emotions is interpreted as an indication of
similarity/distance
We can think of this as a ‘similarity
structure’ for emotions
This 2D representation is common to many
psychological theories
The two dimensions of valence and arousal
correspond well to human ratings of
emotions
This dimensional representation is based on
human subjective emotional experience
rather than behavioural or neural data.
Other theories propose different dimensions
e.g. Edmund Rolls proposed that emotions
can be defined as states elicited by
administration or withholding of reward or
punishment
Intensity increases away from the centre
Emotions are associated with different
reinforcement contingencies
- Presenting a positive reinforcer (S+)
- Presenting a negative reinforcer (S-)
- Omitting/terminating a positive reinforcer (S+/ S+!)
- Omitting/terminating a negative reinforcer (S-/S-!)
Different reinforcement contingencies will
produce different emotional states
Although the axes are different in these 2
representations, the dimensionality remains low in both
For example, fear is a high-arousal, negative-valence
state or a state caused by administration of a negative
reinforcer (i.e. the anticipation of something bad).
Any two dimensions are unlikely to be sufficient to
capture all the variance in emotions
Here, specific emotions have been given labels (e.g.
anger, fear, disgust) based on English words for
emotion categories and located in 2D space
A dimensional approach could also be used to classify
emotions without needing to classify under specific
labels
Emotion states could be categorized based
on their location within multidimensional
space
Here only 3 are illustrated (intensity,
valence, persistence) but more can be
added
We can then observe how emotion states
associated with similar or different
behaviours cluster (or don’t) in this space
Are some emotion
features uniquely human?
- Some emotion states are likely unique
to humans or primates e.g. pride,
embarrassment, awe - Are there general emotion features
that are unique to humans? - Volitional control i.e. emotion
regulation could be thought of as an
’add-on’ that is specific to humans - Subjective report i.e. the ability to
verbally report on our emotional
experience can be conceived of as a
human-specific emotion behavior
caused by an emotion state - Stimulus decoupling- in humans an
emotion state can be induced just by
thinking about stimuli - This could be thought of as an
extreme example of stimulus
generalization
Recognizing emotional
expression in mammals,
model organisms and
martians
- To study emotion states and their
neural mechanisms we need to
identify observable behaviors
that can be used as a ‘readout’ of
experimental manipulations - In mammals, this can often be
identified through similarity to
human behaviours e.g. freezing
to threat in rodents - In model organisms e.g.
Drosophila in which behavioural
repertoires are more primitive this
is more challenging - How can we identify (primitive)
emotion behaviours that are not
similar to our own? - How would you know if a martian
has emotions? - This requires taking a more
ethological approach - Observe range of behaviours in
the species and look for those that
exhibit the properties of emotions
that we outlined - Then, we can investigate how
these behaviours are controlled by
brain states
Why do Drosophila
mate?
* Do Drosophila have emotion states or is
all behaviour controlled by chains of
stimulus-response reflexes?
* Stimulus-response (S-R) view: specific
sensory cues trigger reflexive
behavioural responses that in sequence
produce mating behaviour.
Drosophila mate because they are genetically
programmed to respond to specific signals
(e.g. odour cues) emitted by a potential mate
* Emotion view: behaviours are organized
by a central emotion state
Drosophila mate (at least in part) because it it
is associated with a state of reward
Do Drosophila like
sex?
* Drosophila males will spend more time
close to an odour that they encountered
during mating (Shohat-Ophir et al.,
2012).
(Recall, stimulus generalization and
associative learning)
* Brief activation of ’courtship neurons’
leads to persistent courtship behaviour
* Together these observations suggest
that S-R accounts are at best
incomplete
Experimental
investigation of central
emotion states
- Emotions are a type of central neural state
that are caused by stimuli and that, in turn,
control a wide range of behavioural, cognitive
and bodily changes - These central states have defining properties
that are shared across different specific
emotions within a species and across
different species - How do we search for these central states?
- How do we know if we have found one?
- Understanding how any functional state is
implemented in the brain is a key challenge
that modern neuroscience is grappling with - Research on circadian oscillators provides an
example of how we might distinguish between
‘central states’ and the outputs of central
states
Circadian clocks
* Like emotions, circadian oscillators
control a ‘central state’
* This state is rhythmic changes in
system wide biological processes that
follow a 24h day-night cycle
* Circadian rhythms are evident across
brain regions (& in behaviour,
physiology etc.)
* Circadian research has identified a
central circadian oscillator in the
suprachiasmatic nucleus (SCN) that is
a master controller of circadian
rhythms
* Disrupting the central circadian
oscillator in the SCN disrupts all
circadian rhythms
* Manipulating a single output of the
clock only changes circadian rhythms
in that specific output
* This confirms that there is a central
state regulating circadian oscillations
Take home point
* The defining feature of a
central state is that
experimental manipulations
of that state should affect
multiple outputs of that state
* To determine this, it is
necessary to be able to
manipulate components of the
state (e.g. genes, brain cells)
* For this, we need model
organisms and modern
neuroscience techniques
