Special Bernstein Seminar with the Bernstein-CorTec Awardees Katja Kleespies and Christian Follert: Save the Date!

Katja Kleespies: Structuring the world: Brain oscillatory activity involved in representing naturalistic event schemas

 Prior knowledge shapes how we perceive the world and store new information about it. Part of our prior knowledge is thought to be stored in schemas, which are flexible structures of knowledge generalized from multiple experiences (e.g., eating at a restaurant). While most neuroimaging research on schemas has relied on functional magnetic resonance imaging (fMRI), much less is known about the oscillatory signatures captured by electrophysiological recordings involved in processing and representing schematic information. 

 To address this gap, we measured oscillatory power with scalp electroencephalography (EEG) while participants repeatedly encoded and retrieved six audio-visual movies. The movies were set either in a supermarket or a restaurant (schema context). Within each context, one movie closely followed the typical sequence of events associated with the schema, one followed it only loosely, and one contained no typical schematic events (schema strength). 

 In my talk, I will present both behavioral and neural evidence suggestive of schema-related memory processing. Behaviorally, movies with higher schema strength led to more false recall of schema-consistent details, indicating an ‘overgeneralization’ effect when experiences strongly align with stored schematic knowledge. Neurally, multivariate pattern analyses (MVPA) showed that schema context evoked reliable and generalizable representations across all encoding repetitions, which were reinstated during free recall and reflected in lower frequency bands. Taken together, these findings suggest that oscillatory power plays a key role in how we structure and retrieve memories of complex, everyday-like experiences.

 Katja Kleespies now works as a PhD Student in the Lab of her previous and current Supervisor, Prof. Monika Schönauer, in the Institut for Psychology, Dept. of Neuropsychology.

Christian Follert: Associative memory in self-organizing neuronal networks - Competition for synaptic resources and blocking

 Associative memory in biological neural networks is formed by synaptic plasticity, which leads to changes in connectivity between neurons in the network. This process is generally attributed to Hebbian correlation-based plasticity, while homeostatic plasticity is thought to stabilize the network dynamics.

 We have studied how associative learning arises in self-organizing neural networks with limited synaptic resources based solely on homeostatic principles. To this end, we modeled network connectivity using the directed configuration model for random graphs and developed a mathematical description for changes in connectivity when the network is perturbed.

 In this talk, I will introduce the model and show how competition for synaptic resources shapes associative learning and, under certain conditions, leads to a “blocking-like” effect that has similar properties to the Kamin blocking effect known from behavioral research on associative learning.

Christian Follert is currently finishing his MSc studies in Applied Physics.