Lecture 6 - Lisa Feldman Barrett's Theory of Constructed Emotion Flashcards
Theory of constructed emotion
- A relatively new theory that takes an entirely
different approach to understanding emotion - Proposed as an alternative to a ‘basic emotion’
approach which she argues is too simplistic
and doesn’t account for the complexity of
human emotions, which are shaped by
individual experiences, culture and context - Posits that emotions are constructed through
the interactions between brain, body and - Critiques basic emotion theories for not
adequately taking into account the role of
cognition and culture in creating emotions - Argues that emotions are not innate, but are
constructed form sensory information and
cognitive processes - Challenges assumptions that emotions are
based on specific, fixed neural states that
are triggered by certain stimuli - Argues that emotions are not specific to
particular brain regions or patterns of brain
activity (i.e. emotions do not have a specific
‘neural essence’) and that each instance of
emotion is uniquely constructed
Basic emotion theories propose that a number
of basic emotions are hardwired in the brain in
specific neural circuits that mediate the
experience of each basic emotion.
In this view, emotions are hardwired and
universal.
Barrett’s theory of constructed emotion
proposes that emotions are not hardwired in
the brain but rather constructed based on past
experiences and current context.
Emotions don’t have a specific ‘neural essence’
and are not universal.
In this view, emotions are constructed and
highly individualized.
- Challenges the existence of central emotion
states (‘black-box functionalism’) and of
emotion categories with coherent
responses
Start with the biology:
The brain is a
network of neurons
Neurons send and receive many
inputs making for a one: many &
many: one mapping that makes
all neurons ‘multipurpose
This means that a group of
neurons can create an enormous
number of different patterns of
activity
The brain achieves
complexity through
degeneracy
Degeneracy is the capacity for
dissimilar representations to
give rise to instances of the
same category in the same
context
i.e. the activity of distinct groups
of neurons can create the same
emotion (e.g. anger)
This is in contrast to a view of
hardwired functions
This allows for many-to-one
mappings of structure to function
Biological systems favour
degeneracy because it makes
them robust to damage
i.e. there is redundancy, a backup if one system fails, another
can take over
‘degeneracy means instances of
emotion are created by multiple
spatiotemporal patterns in
varying populations of neurons.
Therefore, it is unlikely that all
instances of an emotion category
share a set of core features (i.e.
a single facial expression,
autonomic patterns, or set of
neurons
Emotions are
biological categories
Biological categories such as
species, are conceptual and don’t
exist in a real sense
A biological category can be
conceptualized as a grouping of
highly variable instances
The ‘average’ doesn’t actually
exist in nature
Emotions are the same: what we
think of as an emotion, e.g. ‘fear’
is a category that groups a
collection of unique instances
We experience emotions as a
coherent, consistent category
because the brain constructs
these unique instances
effortlessly
What is the function
of a brain?
Brains evolved in concert with
bodies growing in size and
complexity
Neural processing is
metabolically costly
Metabolic cost acts as a
selection pressure to shape
brain function
The primary role of the brain is to
coordinate and regulate the body
in the most energy efficient way
possible so the animal can survive
& reproduce
The brain achieves this by
anticipating metabolic needs, not
simply reacting to needs because
this is more efficient.
Allostasis is the core
task of the brain
Allostasis is the maintenance
of functional physiological
systems through change.
This is a process whereby the
brain regulates the body
according to costs and benefits
The body anticipates & adjusts
its energy use according to
environmental demands.
This can be thought of as a
’body budget
Allostasis involves the brain
regulating autonomic nervous
system, immune system and
endocrine systems
Brain regions that implement
allostasis overlap with circuits
presumed to control emotion
These are ‘multipurpose’ brain
regions
Allostasis vs
Homeostasis
Homeostasis is the ability to
maintain stable conditions
through a constant setpoint
Allostasis achieves stability
through change to adapt to
challenge
Homeostasis corrects,
allostasis predicts
How does a brain
perform allostasis?
* The brain is working to achieve
allostasis i.e. maintaining stability by
anticipating and adapting to changing
conditions by adjusting its internal
physiological state
* The brain only has access to sensory
data from the internal and external
world and must use this data to
determine how best to respond
* This is achieved by running an ’internal
model’ (or simulation) of that body in
the world
* The internal model represents both
external and internal environments
What is an
internal model?
- Representing and using information
about the internal environment is
called ‘interoception’ - Interoception is central to the brain’s
internal model and results from
allostasis - Neurons are not quietly waiting to be
stimulated to respond but rather are
constantly processing and shaping
incoming sensory information - This ongoing neural activity is the brain
modeling the world from the
perspective of the body’s physiological
needs
The internal model
is predictive
* The brain uses incomplete incoming
sensory information to probabilistically
predict the current state of the internal
and external world & the actions
required to continue meeting
physiological needs
* Your brain is working to assign
meaning to sensory experience and to
do this, the brain uses past experience
to create simulations (predictions) of
incoming sensory events
* E.g. the image of the bee on a flower
How you perceive the image is influenced
by having seen the full image
- Prediction signals are the brain
changing neural firing to plan
visceromotor actions to regulate the
body (i.e. allostasis) & then receive
incoming interoceptive (sensory)
signals that are the consequence of
these actions - Sensory signals can confirm those
predictions or change them if there is
something unexpected - When there is an error in the
prediction, this changes the future
prediction (i.e. learning occurs) - The goal is to predict better & predict
more efficiently next time - Predictions are brain simulations to
continuously anticipate the changing
sensory environment - This kind of predictive processing is
fundamental to how the brain operates
and underlies all perceptual and
cognitive processing as well as
emotion - The brain is constructing experience
- That is, for a given event, perception (&
emotion) follows action
Affect arises from
interoception
- Interoceptive sensation leads to the
experience of affect (e.g. valence,
arousal) - Affect is NOT specific to emotion but is
a basic feature of consciousness
Interoception is the origin of feeling but
it is NOT the same as feeling and these
sensations do not have discrete qualia - Emotional information emerges from
the relationship between the incoming
sensory signals and the prediction
signals constructed in the brain - The brain generates low dimensional
summaries of your ‘budgetary state’ &
these give rise to affective properties
e.g. valence - Affect can be thought of as the brain’s
best guess about the current
‘budgetary state’ of the body
Emotion is the brain
giving meaning to
sensory signals
- The brain implements the ‘internal
model’ as ‘concepts’ to ‘categorize’
sensations - A concept is a population of predictions
i.e. the group of patterns of neural
activity - Incoming sensory information helps to
select from the predictions to find the
best fit - As a result, the incoming sensory
events are categorized as similar to
some set of past experiences - In this way, the brain uses emotion
concepts to categorize sensations to
construct an instance of emotion - The goal is to guide action to fit the
situation - The result is the perception of an
emotion
The neuroanatomical basis of the internal
model
- Maps of brain regions activated by 5
emotion states converge on similar
functional brain networks - These are the ‘default mode network’ and
‘salience networks’ - Hypothesis that these networks underlie
the brain’s internal model - Simulations are initiated in the default
mode network - The default mode network contains a low
dimensional summary that initiates a
sequence of predictions throughout
distributed cortical areas - All the resulting neural activity is the
‘concept - The salience network tunes the internal
model by predicting which sensory
information to pay attention to i.e. altering
the gain on neurons that compute
prediction error - In this way, the salience network
functions to adjust the internal model to
changing conditions
Summary
- Emotion is constructed using the same fundamental conceptual system that the brain
uses for all cognition and perception - Allostasis (predictively regulating the internal milieu) and interoception (representing the
internal milieu) are the core nervous system function - To make sense of the world, the brain is constantly running an internal model that is
determined by past experiences that are implemented as ‘concepts’ - Concepts are collections of whole brain representations that predict upcoming sensory
events, the best actions to deal with these impending events and how these events will
impact allostasis. - The brain compares concepts with incoming sensory information to identify the most
probable causal explanations for sensory inputs. - Emotion is a brain state that arises in making sense of the sensory information
Some clarifications
- Emotions are not an illusion, they are real – they just don’t have distinct neural ‘essences’ (i.e. not
hardwired) - There are no emotion-specific neurons but neurons do have some specificity
- Understanding the neural bases of emotion will require focusing on ensembles of neurons, rather than
individual neurons - Not everything is cortical- Instances of emotion engage subcortical ‘pattern generators’ to elicit relevant
actions - Animals may have emotions but we cannot know this because emotion depends on the perceiver
Some implications
- Every instance of emotion is unique and although ‘patterns’ of e.g. neural activity
associated with emotions can be detected, this is not a brain state of an emotion, just a
statistical summary of a set of highly variable set of instances - Behaviours are not specific to emotions. Suggests that circuit mapping studies that link
emotion to specific circuits are just revealing mechanisms of behavioural control. - ‘You are not at the mercy of your emotions’ : emotional experience can be changed
through experience
Take home message
‘Emotions are constructions of the world, not
reactions to it.
