Self-organization of Computation in Neural Systems: biological brains to intelligent machines

Abstract:

When learning a complex task our nervous system self-organizes large groups of neurons into coherent dynamic activity patterns. During this, a cell assembly network with multiple, simultaneously active, and computationally powerful assemblies are formed; a process which is so far not completely understood. Using computational methods it can be shown that the combinations of plasticity mechanisms at different timescales (synaptic plasticity and synaptic scaling) enable the formation of such assembly networks. Can such self-organization process enable efficient and fast learning in artificial robotic agents ? We show this type of self-organization allows executing difficult manipulation tasks on multi-degree of freedom manipulation robots. Where in, such assemblies need to learn computing complex non-linear transforms and - for execution - must cooperate with each other without interference. Such mechanism, thus, permits the guided self-organization of computationally powerful sub-structures in dynamic networks for behavior control. In this talk I will explain the theory behind such self-organized computation and show examples of real world applications within an industrial setting.

Molecular Mechanisms of Gene X Environment Interactions: Possible Relevance for Prevention, Diagnosis and Treatment of Psychiatric Disorders

Abstract:

Early adverse exposures, such as maternal stress during pregnancy and child abuse, are thought to result in long-lasting consequences on neural circuit function and stress hormone regulation and ultimately in an increased risk for psychiatric but also medical disorders later in life. Overall, this presentation will describe putative molecular mechanisms how genetic variants and exposure to adversity interact to shape risk and resilience for psychiatric disorders, with a focus on stress hormones. These glucocorticoids (GCs) have been shown to alter gene expression pattern and to induce long-lasting epigenetic changes in specific loci through binding of the glucocorticoid receptor (GR) to glucocorticoid responsive elements (GREs).

The talk will first highlight data from a human hippocampal cell line that identify long lasting epigenetic changes in DNA methylation in response to GCs. These lasting epigenetic changes are located in brain development- and disease-relevant gene enhancer regions and lead to increased transcriptional sensitivity to future stress exposure. Data from human brain organoids and single cell sequencing will then delineate whether specific subtypes of cells show differential sensitivity to early GC exposure during brain development. The second focus will be on common genetic variants in long-range enhancer elements that can modulate the transcriptional and epigenetic response to GR activation and early life adversity. These functional genetic variants associate with increased risk for psychiatric phenotypes and differences in neural correlated of stress processing and are in turn enriched among the loci showing lasting DNA methylation changes with GC in the hippocampal cell line described above.

Overall, the presentation will outline how stress-exposure can have lasting effects on cell and tissue function and how this relates to risk or resilience to stress-related disorders in the context of common genetic variation.