ePoster

An inhibitory network model explains the transient dynamics of hippocampal ripple oscillations

Natalie Schieferstein,Tilo Schwalger,Richard Kempter,Benjamin Lindner
COSYNE 2022(2022)
Lisbon, Portugal
Presented: Mar 17, 2022

Conference

COSYNE 2022

Lisbon, Portugal

Resources

Authors & Affiliations

Natalie Schieferstein,Tilo Schwalger,Richard Kempter,Benjamin Lindner

Abstract

Hippocampal ripple oscillations (140–220 Hz) have been implicated in important cognitive functions such as memory consolidation (Buzsáki 1989). Their generating mechanism however remains unclear, although various models have been proposed. We suggest that transient features of ripples can guide model selection: So far only the “bifurcation-based” inhibitory network model (Donoso et al. 2018, Brunel and Hakim 1999) could reproduce the experimentally observed intra-ripple frequency accommodation (IFA) — an asymmetry in the instantaneous ripple frequency in response to transient, sharp wave-like (50–100 ms) stimulation (Ponomarenko et al. 2004). This model assumes that the recurrent CA1 PV⁺ interneuron network generates ripples for strong enough excitatory drive. Here we explain IFA using a theoretical mean-field approach and numerical simulations. In the mean-field limit we describe the density of membrane potentials as a gaussian with time-dependent mean and consider only drift-based spiking. This framework allows us to approximate analytically the frequency and amplitude of the network oscillation as a function of the external drive. Numerical simulations verify that the approximation works in a large parameter regime. We show that for fast changing, sharp wave-like drive the network frequency response is asymmetric due to a speed-dependent hysteresis effect in the oscillation amplitude of the mean membrane potential. We predict that IFA vanishes in the limit of slowly changing transient drive, which can be tested optogenetically. With no further parameter dependencies, IFA is an inherent feature of this model. Conversely, we find that the alternative, “perturbation-based”, inhibitory ripple model (Malerba et al. 2016) cannot exhibit IFA by default. From a theoretical perspective our work introduces a new ansatz to describe strongly nonlinear oscillation dynamics in recurrent spiking networks with interesting links to escape noise formalisms. For experimentalists we provide new evidence to advance the search for the true generating mechanism of ripples.

Unique ID: cosyne-22/inhibitory-network-model-explains-transient-dc2e60f0