Particle-swarm based modelling reveals two distinct classes of CRH^PVN neurons

Area of Science:

• Neuroscience
• Computational Neuroscience
• Endocrinology

Background:

• Electrophysiological recordings offer insights into neuronal dynamics but rapid modeling for network inference is challenging.
• The paraventricular nucleus of the hypothalamus (PVN) is critical for the mammalian stress response, housing corticotropin-releasing hormone (CRH) neurons.
• Understanding CRH-PVN neuron dynamics is key to deciphering stress response mechanisms.

Purpose of the Study:

• To develop a rapid computational modeling platform for electrophysiological data of CRH-PVN neurons.
• To investigate how single CRH-PVN neuron dynamics influence network-level responses within the PVN.
• To classify CRH-PVN neurons based on their firing dynamics.

Main Methods:

• Whole-cell current-clamp recordings from CRH-PVN neurons.
• Rapid and automatic fitting of neuron dynamics using a modified adaptive exponential integrate-and-fire (AdEx) model with particle swarm optimization (PSO).
• Construction and simulation of CRH-PVN neuron networks.

Main Results:

• The AdEx model with PSO accurately fit all recorded CRH-PVN neurons, identifying multiple parameter sets with stable dynamical features.
• CRH-PVN neurons were classified into two subtypes: 'integrators' and 'resonators', based on their firing bifurcation.
• CRH-PVN resonators were experimentally confirmed and shown to maintain network firing, even under inhibitory input.

Conclusions:

• A novel computational approach enables rapid biophysical modeling of CRH-PVN neurons.
• The discovery of CRH-PVN integrator and resonator subtypes provides a new framework for understanding PVN function.
• CRH-PVN resonators play a vital role in maintaining the baseline activity of the PVN network during stress response.