51 min
Seminar, Hosted by van Vreeswijk Theoretical Neuroscience Seminar
From Cajal’s Spines to Digital Neurons: Connectomics View on Dendritic Computations
ELSC, The Hebrew University of Jerusalem
Recording
Abstract
Dendritic spines are among the smallest structures in the brain, yet they hold some of the deepest clues to how neurons compute and store memories. This lecture revisits the classical question of spine biophysics in the age of connectomics, where nanometer-resolution electron microscopy allows whole neurons with thousands of synapses and spines to be reconstructed. It shows how spine density, spine-neck resistance and the irregular geometry of individual spines shape local synaptic signals, generate fast voltage dynamics, influence excitatory/inhibitory gating and affect communication between neighboring spines.
The lecture then discusses how automated reconstruction pipelines and biophysical modeling make it possible to build digital neurons at spine and synapse resolution, and how AI-based approaches, such as the new TwinProp algorithm, allow a fundamental question to be asked and answered: what can a neuron compute? It closes by arguing that a new age of connectomics now provides an essential constraint on understanding how synapses, neurons and circuits perform their computational functions.