Lecture 3 - A framework for studying emotions (from Adolphs and Anderson) Flashcards
What is emotion according to Joe LeDoux?
- We need to define terms
- There is a lot of disagreement and
confusion even in how scientists use
the term ‘emotion’ - e.g. Joe LeDoux
- Joe LeDoux & his team
published many influential
studies on fear learning and
fear circuits in rodents - More recently he stopped using
the word ‘fear’ or ‘emotion’ to
describe his findings with nonhuman animals - He now talks about survival
behaviors and survival circuits
What is emotion?
- We can start to answer the question of
‘what is emotion?’ by answering other
questions
What do emotions do?
What causes emotions?
What is caused by emotions?
i.e. a definition of emotion in terms of its
functions
Emotions are
functional states
- Emotions are states of the brain
- But if we define emotions literally by
their brain state, we need different
definitions for humans, flies, octopus
because they have completely
different nervous systems - A more abstract definition can be
more useful and broadly applicable - Emotions are states of the brain i.e.
they can be implemented in the brain - We also need a functional account of
emotions that defines emotions in
terms of the stimuli that cause the
brain state and the behaviours etc.
that are caused by the brain state - Functional definitions identify states by
their causal relations (i.e. what does it
do?) NOT by how they are constituted (i.e.
what is it made of?) - Functional definitions describe causal
effects in an abstract manner that is
independent of the physical way that that
the state is implemented - Functional definitions are generalizable
(e.g. to other species)
Ex:
What is a clock?
* A clock is a device that measures time
* By defining a clock by its function
(measuring time) we have a definition
that is broadly applicable to a category
Functional definition
of mental states
* We can define mental states in terms of
their causal relations to inputs and
outputs
* Emotions are functional states that are
caused by sensory inputs and cause
behavioural outputs
* Emotions can also be caused by and
cause changes in other mental states
(e.g. memory, perception, attention etc.)
‘Fan-in’ ‘fan-out’ architecture
Stimuli (+ context) - Central emotional state (+ volitional control) - Observed behaviour, subjective reports, psychophysiology, cognitive changes, somatic responses
Stimuli can be external and/or internal
- Observed behaviour, subjective reports, psychophysiology, cognitive changes, somatic responses can also be the stimuli, not just the output
Context exerts a moderating effect
Volitional control can inhibit emotion
Researching emotions as functional states
Ex: stimulus of spider vs. butterfly
A clock measures time but what does an
emotion ‘do’?
- We can also have a functional account in evolutionary terms that explains
why an emotion is adaptive and what purpose it served in the evolutionary
environment - Anything that evolved via natural selection is selected based on its
functional effect for an organism in its given environment - Unlike for the clock, the adaptive functions of emotions may not be
immediately clear - Ultimately we need a functional explanation of emotion at the level of the
entire organism and how it interacts with its environment
Recall….
Disgust served to
avoid poisonous or
contaminated food
- Functional definitions of emotion states
are useful because they CAN generalize
(e.g. to different animals, to robots) - However, whether emotions can be
embodied in all brains or even in systems
that are not brains is an empirical
question - It may turn out that there are limits on the
kinds of systems that can instantiate
emotion, even once emotion is more fully
understood
The problem of
‘type identity’
- When you experience an
instance of emotion e.g. joy, this
corresponds to a pattern of
brain activity and this is
presumably different from the
pattern of activity when you
experience another emotion
e.g. fear - Is the brain activity when you
feel fear today the same as
when you experienced fear last
week? - Is it the same as when your
friend experiences fear? - Is it similar to when your cat
experiences fear? - How similar at the
neurobiological level are
different instances of fear? - How can we measure this
similarity? - Even though any instance of
emotion is a brain state, it is
less clear if a particular type of
emotion (e.g. fear) corresponds
to a type of brain state - This is the problem of ‘type
identity’ i.e. how does emotion
state map to brain state - Coming back to the clocks, knowing their
function allows us to group them in the
category of ’clocks’ - Observing an organism’s behavior (in its
natural environment) reveals what emotions do
and so what type of emotion a brain
mechanism might implement - Behavioral readouts are necessary to interpret
brain states - How to assign physical brain states to
functional categories of specific
emotions is an empirical question - It may be that even a single human
brain can implement emotion in many
different ways (e.g. Lisa Feldman
Barret- more on this later) - Or there may be only one way to
implement emotion by a particular
circuit architecture - Likely the truth is somewhere in
between - It is plausible that evolution arrived at
several different mechanisms for
implementing an overlapping set of
computational functions that work
together to coordinate an emotion state - We can start to discover these through
experiments integrating neurobiology
and behavior - In this way we can start to answer
‘what is an emotion?’
Malfunction and
psychiatric disorder
- Understanding the proper function of
emotion also leads to the study of when
emotions ‘malfunction’ and are not serving
an evolved adaptive function - We can conceptualize disorders as an
inappropriate application of emotion - For example, think of all the ways that a
clock could be used other than for
measuring time - Adaptation is relative to context. Emotions
may not be adaptive in these contexts - For example, fear in post traumatic stress
disorder (PTSD) - Symptoms of PTSD could be adaptive in the
context of immediate threat (e.g. a warzone)
but become maladaptive in another context
(e.g. in a safe home environment)
Emotions &
Consciousness
- The conscious experience of
emotions i,e, feelings are often
considered the defining feature
of emotion - Feelings are just one aspect of
emotion - Explaining feelings requires
explaining consciousness - This is really hard!
- Also limits the ability to study
emotions in animals - Focusing on conscious emotion
influences how we study
emotions across people from
different cultures who may
describe their feelings differently - While cognitive scientists are solving
consciousness, affective neuroscientists
can make a start on understanding
emotion by side-stepping feelings - Focus on the functional state i.e. what do
emotions do - This provides an objective starting point
that can be used across areas of
research - Functional definitions can eventually also
be applied to study feelings - But feelings and the words we use to
describe emotion are unlikely to provide
a reliable starting point
Our conscious
experience is not
veridical
* Feelings are a kind of conscious
experience
* Our brains construct our conscious
experience. This conscious ‘reality’ is
not infallible and selectively presents
elements of some external (or internal)
reality
Vision and beliefs
* Conscious visual experience is also
constructed.
* What we experience as vision is an
incomplete representation of the
external visual world
* We have the sense that we are
consciously aware of everything in our
visual field yet experiments and day to
day experiences disprove this
* e.g. failing to see an oncoming car
when crossing an intersection
* Our conscious visual experience is not
a complete and faithful reproduction of
an objective reality but we experience
it to be this
* Our conscious experience is not the
best guide to understanding what is
happening in our brain during vision
Other problems with
feelings
* Equating emotions with feelings raises
the problem of attributing conscious
experience to animals
* This is part of Joe LeDoux’s concerns
about talking about emotion in nonhuman animals
* If an animal has feelings then they
have consciousness.
* How can we know about the conscious
experience of emotions in non-human
animals?
* How can we know about conscious
experience of emotions in other
humans?
Separating the study of
emotions from the study of
conscious experience
- We can define ‘emotion’ in a scientific
usage as something distinct from
conscious experience of emotion - For example, vision scientists study
vision using behavioral measures
without studying visual consciousness - Conscious experience can be
separated from behavior and other
functional measures - Blindsight demonstrates this in the
case of vision and visual
consciousness - We can apply a similar principle to the
study of emotion - Use behavior, and neurobiology as a
foundation, not self-reported feelings - There is still room for studying
feelings as one aspect of emotion - The starting point should operationalize
emotion in a way that allows for a science of
emotion across different kinds of data and
scientific disciplines e.g. ethologists studying
animal behaviour, psychologists studying the
mind, and neurobiologists recording the
activity of the brain - None of these pursuits require an account of
consciousness
Recap
- We need a functional approach to
understanding emotions: define emotions by
what they do, not how they are physically
implemented - A functional approach to emotions also
implies the possibility of ‘malfunctions’
when emotional behavior is not adaptive in
the context - Feelings are not the same as emotions. We
can study emotion without tackling
consciousness
Reading: A framework for Studying Emotions across Species by Anderson and Adolphs
“The paradox of emotions
is that, on the one hand,
they seem self-evident
and obvious when
examined introspectively;
on the other hand, they
have been extremely
difficult to define in
objective scientific
terms.”
” ..an ‘emotion’ constitutes an internal, central
(as in central nervous system) state, which is
triggered by specific stimuli (extrinsic or
intrinsic to the organism). This state is encoded
by the activity of particular neural circuits that
give rise, in a causal sense, to externally
observable behaviors, as well as to associated
cognitive, somatic and physiological responses”
Conventional view: emotions
incorporate many components
that need to be coordinated and
often synchronized for the
experience of emotion
Anderson & Adolphs: emotions involve all the same components but these are not part of
the emotion, rather they are caused by the emotion state
Charles Darwin considered emotional expression
from a functional and evolutionary standpoint.
He asserted that emotion homologous to our own
is easily recognizable in humans and other animals
and can even be observed in insects.
Darwin did not provide objective criteria for
identifying emotion.
Anderson & Adolphs agree with Darwin:
even invertebrates have primitive emotion states
But they differ from Darwin in that they argue that
these states are not necessarily homologous to
specific categories that have been used to define
human emotions (e.g. fear, anger, happiness).
They assert that emotion states in all animals share
certain fundamental properties: ‘emotion primitives’
Emotion primitives are the evolutionary building
blocks of emotion that are shared across species.
Species-typical behaviours that arise from these
emotion primitives are not necessarily shared
The key question is not if animals share a specific
emotion (e.g. fear) but if animals have central states
that share features of emotions in general.
We can investigate general features of emotion in
animal models without using anthropocentric labels
like ‘fear’, ’anger’, or ‘sadness’.
For this to work, we need operational criteria for
emotion across species. Their goal is to propose a
way of thinking about emotion.
They discuss 4 points
1) The causal relationship between emotions and
observable behaviour
2) The relationship between emotion states and
subjective feelings
3) The characteristic features of emotion states
shared by specific emotions
4) If there are uniquely human features of emotion
Causal relationship between emotions and behaviour:
There is not yet universal agreement about the direction of
causality between emotion and behaviour.
Part of this disagreement comes
from lack of data.
Purely observational approaches
studying the link between
emotion and behavior cannot test
the directionality of causality.
Studying the neural basis of
emotion states in model animals
allows us to directly and
rigorously test causality by
manipulating neural activity with
modern neuroscience tools.
The relationship between central emotion
states and subjective feelings
Behaviourist view:
Emotional stimuli directly evoke behaviour in animals.
In humans, the conscious awareness of behavioural and
somatic responses evoked by the emotional stimulus
gives rise to subjective feelings
Anderson & Adolphs
Emotional stimuli evoke central emotion states.
In animals and humans these states give rise to behavioural and other responses,
including subjective feelings (in humans)
Feelings can only be assessed by verbal report so
we can only study them in humans.
If we equate emotions with subjective feelings, we
cannot study emotion in animals.
There is no reason a priori to conclude that
animals do not have central emotion states.
Ex: Aside: Koko the gorilla who learned sign
language
Depressed when kitten died.
“When we told Koko, she acted like she didn’t hear us
for about 10 minutes,” Cohn said. “Then she started
whimpering–a distinct hooting sound that gorillas
make when they are sad. We all started crying
together.”
‘Koko the “talking” gorilla whimpered with grief and
“discussed” the death of her pet kitten, struck and
killed by a car, for several days after getting the bad
news, her teachers say.’
Affective neuroscience should look for evidence in
animals of central emotion states with certain
fundamental properties that are causal in
responding to certain stimuli with specific
behaviours and the corresponding neural
mechanisms.
What are these fundamental properties?
