Subproject 4

Signatures of network reset during dynamic belief updating in volatile and stochastic environments

PI:	Dr. Jan Gläscher

How do we detect changes in a dynamically changing, but also environment? This is a fundamental problem encompassing grand and small aspects of our existence. An example for the grand aspects of human life on the planet is the relationship between weather and climate: Individual weather events that show a large variability are unfolding on the canvas of climate, but how do we detect a change in climate from the individual (and noisy) weather events? A tennis player rehearsing his return with a ball throwing machine is an example for the small-scale problems: the machine inherently shoots the tennis ball to slightly different (noisy) locations, but how does the tennis player detect, when the overall direction of machine has changed? And who would the behavior change if the machine had two kinds of balls, one easy to return (and thus more rewarding) and another more difficult? Would the tennis player focus on the easy balls, while ignoring the others?

In this project of the RU we investigate how the human brain prepares and responds to changes in the environment (e.g. the orientation of the ball throwing machine has shifted), when our world model is no longer valid and has to be adapted and compare these processes to the situation, when just the inherent variability of noisy environments changes (e.g. when the directedness of the machine is reduced due some technical failure). Detecting these different kinds of variability due to environmental changes vs. a change in the noisiness of the information is a difficult problem, because a differentiation is not possible with one piece of information alone. Rather it requires the accumulation of evidence for environmental changes over time. Nevertheless, solving this problem is of crucial importance for effective and adaptive behavior. Environmental changes are often accompanied by cascade of neural events starting from change detection in the mid-brain which is broadcast to cortical areas eliciting large-scale changes in oscillatory activity and triggering and arousal response detectable in the increase in pupil size. In this project we will used a combination of synchronous recordings EEG, fMRI, and eye-tracking to measure activity in all stages of the network reset cascade, which will allow us to characterize the neural computations surrounding change detection and noise processing in detail and to find out how the human brain is solving this fundamental problem.