Lab of Rebecca Jordan at the University of Edinburgh
The locus coeruleus projects densely throughout the cortex and wider brain, receives diverse long range and local inputs, and responds to many stimuli, including discrepancies in expected input. However, little is known about what this system computes. We aim to understand the computational role of the locus coeruleus in predictive learning, and determine how its circuitry enables its function.
Computation in the locus coeruleus
Prediction errors in cortical learning
Prediction errors are discrepancies between expected and actual sensory input and are thought to drive corrective changes to the brain's predictions. Prediction errors have been found in layer 2/3 cortical neurons and in neuromodulatory systems. We aim to test the involvement of these signals in driving plasticity using virtual reality systems and manipulations of neural activity.
Predictive learning in autism models
A prevalent model of autism spectrum disorder is a reduced influence of predictions in the brain, leading to perception being over-dominated by the sensory input. This could explain many of the features of autism spectrum disorder, from sensory overload to repetitive behaviours. Reduced predictions could arise from improper neuromodulation, excitatory-inhibitory imbalance or reduced capacity to learn predictions. We aim to assess these ideas in mouse models of autism.