Spikes can transmit neurons’ subthreshold membrane potentials

Authors: Valentin Schmutz

Presenting author: Valentin Schmutz

Presentation type: Flash talk at SNUFA 2025 online workshop (5-6 Nov 2025)

Abstract

Neurons in the mammalian brain mainly communicate through the emission of large, pulse-like depolarizations of their membrane potential called spikes. To fire a spike, a neuron’s membrane potential needs to cross a threshold. Between spikes, the subthreshold membrane potential fluctuations of a neuron are, by definition, not transmitted to other neurons, but they carry precious information as they reflect the total synaptic input received by the neuron. The richness of sensory and behavioral information contained in the subthreshold membrane potential dynamics, compared to that contained in the timing of spikes, has put forward the idea that spikes are just “the ‘tip of the iceberg’ in terms of neuronal activity” (Petersen, 2017). But if a single presynaptic neuron does not transmit any information about its subthreshold membrane potential fluctuations, does this imply that subthreshold information is lost, in the sense that it is not accessible to a postsynaptic neuron? At the single-neuron level, the loss is real; optimal estimation of the membrane potential from spikes only allows for partial recovery of the membrane potential (Pfister, Dayan, and Lengyel 2010). Adopting a population-level perspective, we prove that the membrane potential fluctuations of a presynaptic population of neurons emitting sparse spikes can be fully and perfectly transmitted to a postsynaptic population of neurons. Our proof combines ideas from high-dimensional probability with recent results on concentration of measure in networks of spiking neurons. Incidentally, this mathematical result provides a possible explanation for why the processing of sensory signals is improved in the ‘desynchronized state’ of cortical activity: weak correlations in the desynchronized state may be the reflection of a high-dimensional regime enabling population-level transmission of subthreshold information.