Nonlinear cellular computations contribute to spatiotemporally precise storage and expression of memory in single neurons of the hippocampus

Ching-Lung Hsu^1,2,3,4*

¹Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan
²Neuroscience Program of Academia Sinica (NPAS), Academia Sinica, Taipei, Taiwan
³Department of Life Science, National Taiwan University, Taipei, Taiwan
⁴Graduate Institute of Life Sciences, National Medical Defense Center, Taipei, Taiwan

* Presenter:Ching-Lung Hsu, email:hsuc@ibms.sinica.edu.tw

Our survival relies on the amazing capabilities of rapidly encode important information in the brain, even in response to single exposure of experiences. The underlying mechanisms remain elusive. Such one-shot learning creates challenges for neural circuits to balance the needs for memory specificity and storage capacity. A standard picture of deep neural networks uses single-compartment, passive processes as model neurons to solve the problem, based on architectures and algorithms hardly generalizable to biological systems. In contrast, actual neural circuits use morphologically complex neurons with compartmentalized nonlinear dynamics. Focusing on the hippocampus, a brain region crucial for spatial and episodic memories, our lab has applied quantitative approaches to behavioral and electrophysiological analyses for the understanding of learning mechanisms. Here, I will demonstrate how single hippocampal neurons utilize nonlinear properties to implement learning rules that are spatiotemporally precise at the levels of inputs to individual cells. The subcellular computational properties may contribute to rapid, robust and flexible circuit coding of events and environments.

Keywords: Computational Neuroscience, Neuron, Learning rule, Hippocampus, Synaptic plasticity