Single-Neuron Criticality Optimizes Analog Dendritic Computation

TL;DR

This paper models large dendritic trees as active media that operate near a critical phase transition, optimizing their ability to process complex stimuli through analog computation at the single-neuron level.

Contribution

It introduces a probabilistic cellular automaton model showing how non-deterministic dendritic spikes lead to a critical state, enhancing understanding of dendritic computation.

Findings

Dendritic arbors can undergo a phase transition to a self-sustained active state.

Neurons optimize stimulus discrimination at the critical point.

Dendritic computation is better described as analog near criticality.

Abstract

Neurons are thought of as the building blocks of excitable brain tissue. However, at the single neuron level, the neuronal membrane, the dendritic arbor and the axonal projections can also be considered an extended active medium. Active dendritic branchlets enable the propagation of dendritic spikes, whose computational functions, despite several proposals, remain an open question. Here we propose a concrete function to the active channels in large dendritic trees. By using a probabilistic cellular automaton approach, we model the input-output response of large active dendritic arbors subjected to complex spatio-temporal inputs and exhibiting non-stereotyped dendritic spikes. We find that, if dendritic spikes have a non-deterministic duration, the dendritic arbor can undergo a continuous phase transition from a quiescent to an active state, thereby exhibiting spontaneous and self-sustained localized activity as suggested by experiments. Analogously to the critical brain hypothesis, which states that neuronal networks self-organize near a phase transition to take advantage of specific properties of the critical state, here we propose that neurons with large dendritic arbors optimize their capacity to distinguish incoming stimuli at the critical state. We suggest that "computation at the edge of a phase transition" is more compatible with the view that dendritic arbors perform an analog rather than a digital dendritic computation.