Predictive LeaRning Group
Group Head: Vincent Laurent
We study how the brain extracts predictive information from the environment and uses this information to control our behaviours and decisions.
Animals including humans learn about environmental stimuli that signal the arrival or the omission of important events. This predictive learning allows us to anticipate and prepare for future events by regulating our behaviours and decisions in the present. Our group uses rodents to understand how predictive learning is instantiated in the brain and how it influences our behaviours and decisions.
Interactions between aversive and appetitive predictions
Predictive learning can take multiple forms. It is excitatory when a stimulus predicts the arrival of an important event. It is inhibitory when the stimulus predicts the omission of the important event. Further, the motivational quality of the event dictates whether the predictions are appetitive or aversive. Appetitive predictions involve pleasant events such as foods whereas aversive predictions entail unpleasant events such as loud noises. Our group is interested in understanding the psychological and neural mechanisms regulating how these various forms of predictions interact with each other to control how we learn about rules in our environment and how we behave as a consequence of these rules.
Updating our predictions
We live in a constantly changing environment and a stimulus that used to reliably predict an important event may suddenly cease to do so. Successful adaptation requires detecting this sudden change and updating our predictions accordingly. These processes are commonly studied through a paradigm known as extinction. It involves breaking the predictive relationship between a stimulus and an important event by repeatedly presenting the stimulus alone. Our group investigates the neural architecture underlying extinction. Although we study appetitive extinction, most our effort focus on aversive extinction due to its relevance to the development a new strategies to treat anxiety disorders such as post-traumatic stress. Indeed, these disorders are characterised by a failure to extinguish fears towards environmental stimuli that no longer predict any threatening events.
Cholinergic regulation of aversive predictions
The amygdala and the medial prefrontal cortex are critical for learning and extinguishing aversive predictions. These two structures receive dense projections from cholinergic neurons residing in the basal forebrain. Recent work from our groups indicates that cholinergic projections to the amygdala regulate the strength and durability of the memory supporting an aversive prediction. Further, the same projections also appear to protect the memory from erasure when the aversive prediction is later extinguished. Current work aims to confirm these findings and to determine whether cholinergic innervation of the medial prefrontal cortex serves similar functions.
Appetitive predictions and action performance
A stimulus that predicts food energises performance on an action delivering food. This general Pavlovian-instrumental transfer (PIT) effect is exacerbated in obese individuals. Obesity has been associated with brain insulin resistance and our group is interested in examining whether this resistance distorts activity in the neural circuitry supporting general PIT. Specifically, we focus on the role played by insulin receptors expressed on cholinergic interneurons in the nucleus accumbens core. Indeed, these interneurons are know to control the expression of general PIT. Our hypothesis is that a lack of insulin receptors activity on these interneurons amplify general PIT and therefore promote overeating, which is a key feature of obesity.
Appetitive predictions and action selection
A stimulus that predicts food can exert a more selective influence over our actions than the one observed in general PIT. This is demonstrated via specific PIT, during which a stimulus predicting a particular food biases choice towards an action earning the same, but not a different food. In collaboration with the Integrative Neuroscience Group, we aim to identify how the brain allows such a selective influence of predictive learning on choice between actions. We have recently found that specific PIT requires the durable accumulation of delta-opioid receptors on the membrane of cholinergic interneurons located in the nucleus accumbens shell. Our current efforts intend to describe how this memory is acquired, updated when needed, and how its expression is implemented.