Voltage- and Branch-specific Climbing Fiber Responses in Purkinje Cells

Yunliang Zang, Stéphane Dieudonné, Erik De SchutterbioRxiv 2018 · DOI

TL;DR

This paper introduces a new computational model of Purkinje cells (PCs) that reveals how voltage and dendritic branch structure dictate the way these neurons process climbing fiber (CF) input. The model explains previously conflicting experimental findings and suggests that PCs are not passive recipients of CF signals but actively integrate them, contributing to cerebellar learning.

What Problem Does This Paper Address?

The cerebellum is crucial for motor learning and coordination, and Purkinje cells (PCs) are the primary output neurons of the cerebellar cortex. Climbing fibers (CFs) provide PCs with "instructive" signals, thought to be related to movement errors, which are critical for learning. However, experimental studies have shown a surprising variability in how PCs respond to CFs, particularly in their dendrites. Some studies suggest CF-evoked calcium influx in distal dendrites is unreliable, while others show it's crucial. Furthermore, the exact mechanisms behind the variability in complex spike (CS) duration and spikelet number in PCs have remained unclear. This paper aims to resolve these discrepancies and understand the biophysical basis of CF processing in PCs by developing and utilizing a detailed computational model.

Methods

The authors developed a new, experimentally validated computational model of a Purkinje cell. This model incorporates detailed biophysical properties of various ion channels (Na+, K+, Ca2+) distributed across the PC's soma, axon initial segment (AIS), and dendritic tree. They used this model to systematically explore how different factors, such as voltage state (manipulated by somatic holding current), dendritic morphology, and concurrent synaptic inputs, influence CF-evoked responses. The model was meticulously tuned against a wide range of experimental data to ensure its accuracy.

Datasets

Not explicitly discussed as datasets for training/evaluation in the traditional machine learning sense. The model was built and validated against a comprehensive collection of in vitro experimental data from various studies on Purkinje cells, as detailed in the references.

Key Insights and Ideas

  • Voltage-gated dendritic spikes: The model demonstrates that the voltage state of the PC critically gates the initiation and propagation of dendritic spikes, which are essential for processing CF input.
  • Branch-specific computation: Purkinje cell dendrites are not uniform. Individual branches act as distinct computational units, with their own excitability and ability to process CF signals independently.
  • Reconciling conflicting data: The model successfully explains why different experimental studies have reported varying degrees of reliability for distal dendritic CF responses. This variability is attributed to differences in the PCs' voltage states during recordings.
  • Active integration of CF signals: PCs are not passive integrators of CF input. Their voltage- and branch-specific responses allow them to actively process and integrate these signals, rather than just passively receiving them.
  • Energy efficiency: The study also touches upon the metabolic cost of spiking, suggesting that dendritic processing might contribute to energy efficiency.

Key Results

  • The model accurately reproduces a wide range of experimental observations regarding simple spikes (SSs) and complex spikes (CSs) in PCs.
  • Dendritic spikes, crucial for CF processing, are shown to be highly dependent on the PC's membrane potential. Depolarization facilitates dendritic spike generation and propagation.
  • The model reveals that different dendritic branches have distinct excitabilities, determined by their morphology and ion channel distribution, allowing for branch-specific processing of CF inputs.
  • Somatic CS shapes (duration, spikelet number) are modulated by voltage state, the timing of CF input relative to preceding simple spikes (phase-dependency), and the rate of CF firing.
  • Clustered parallel fiber (PF) or stellate cell synaptic inputs can locally modulate dendritic responses, influencing the timing and occurrence of secondary dendritic spikelets and somatic CSs in a spatially precise manner.

Limitations and Future Work

  • In vitro vs. in vivo: The model's validation is primarily based on in vitro experimental data. The authors acknowledge that further experiments are needed to validate these findings in vivo.
  • Ionic channel details: While the model incorporates many ion channels, the authors note that the roles of dendritic SK2 and Kv1 currents in regulating dendritic excitability were not fully explored and could further refine the model.
  • Plasticity mechanisms: The paper touches upon how CF signals might influence synaptic plasticity (LTD/LTP) but does not simulate these plasticity mechanisms directly. Future work could explore how the voltage- and branch-specific CF responses interact with plasticity induction.
  • Other neuron types: The study's findings on energy efficiency and dendritic processing could be extended to other neuron types.

Summary updated

Abstract

Summary Climbing fibers (CFs) provide instructive signals driving cerebellar learning. However, conflicting experimental studies have been reported about the reliability of CF mediated Ca 2+ influx in Purkinje cell (PC) distal dendrites. Mechanisms causing the wide variation in duration and spikelet numbers of complex spikes (CSs) have not been explored systematically. Using a new experimentally validated PC model, we describe the full range of modifiability of CF responses to explain the experimental data and make new predictions. We find voltage state gates the initiation and propagation of dendritic spikes. PC dendrites exhibit inhomogeneous excitability with individual branches as computational units for CF input. Somatic CSs are regulated by voltage state, CF activation phase and instantaneous CF firing rate. Concurrent synaptic inputs can affect CSs by modulating dendritic responses in a spatially precise way. These voltage- and branch-specific CF responses will increase dendritic computational capacity and give PCs an active role in integrating CF signals.
BibTeX
@misc{zang2018voltage,
  author = {Yunliang Zang and Stéphane Dieudonné and Erik De Schutter},
  title = {{Voltage- and Branch-specific Climbing Fiber Responses in Purkinje Cells}},
  year = {2018},
  doi = {10.1101/284026},
  howpublished = {bioRxiv},
}
Figure