时间：2026年10月8日（周四）中午 12:00
地点：清华大学吕大龙楼11层1100
报告主题：Deriving analytical theory from biophysically detailed neuron simulators reveals principles of dendritic inhibition
报告人： 北京大学博士研究生贺禹涛
内 容：
A central challenge in neuroscience is to derive general principles from biological complexity without simplifying away its underlying mechanisms. Biophysically detailed neuron models preserve many of these mechanisms, but extracting principles from their complex dynamics remains difficult. Here we introduce Simulator-to-Equation (S2E), a framework that transforms a detailed neuron simulator into a single mathematically analysable equation while retaining all mechanisms represented in the original model. This allows us to derive general, experimentally testable theory directly from the model’s dynamics. Using dendritic inhibition as a test case, we established how excitation shapes inhibitory control across neuronal compartments, unifying distinct, experimentally documented forms of inhibition—from perisomatic to distal dendritic inhibition—under a common biophysical theory. In a layer 5 pyramidal neuron model, the analysis further predicted that synchronous, rather than sustained, distal excitation favours event-locked inhibition at an apical-trunk choke point. Differentiating the S2E equation yielded a transfer–drive (Φ–a) law linking the first-order somatic voltage response to two factors: local inhibitory drive and state-dependent dendrite-to-soma coupling. Their alignment in space and time explained the inhibitory patterns. Across neocortical, hippocampal and striatal models, the same law revealed how electrically compartmentalized and compact dendritic architectures converged on similar inhibitory control through different balances of local drive and dendrite-to-soma coupling. S2E thus turns biophysically detailed simulations into a source of general theory, revealing how similar neuronal functions can arise from different underlying dynamics.
