Rethinking Single Neuron Electrical Compartmentalization: Dendritic Contributions to Network Computation In Vivo

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cellular Neuroscience

Background:

• Established theory posits active dendritic processing enhances neuronal computation via electrical compartmentalization.
• This compartmentalization allows dendrites to perform semi-independent operations, creating a "network-in-a-neuron" model.
• Ex vivo studies consistently show high dendritic compartmentalization.

Purpose of the Study:

• To review and contextualize recent in vivo findings on dendritic processing in mouse cortex.
• To discuss the implications of these findings for the established theory of neuronal computation.
• To explore how coordinated somatic and dendritic activity contributes to cortical functions.

Main Methods:

• Review of recent in vivo functional imaging experiments in mouse cortex.
• Theoretical analysis of neuronal computation models.
• Discussion of experimental and theoretical neuroscience literature.

Main Results:

• Recent in vivo imaging in mouse cortex shows surprisingly little evidence for strong dendritic compartmentalization.
• This contrasts with findings from ex vivo preparations.
• Coordinated activity in soma and dendrites may underlie complex cortical computations.

Conclusions:

• The degree of dendritic compartmentalization in vivo may be less than previously assumed from ex vivo studies.
• Revisiting the "network-in-a-neuron" model is necessary in light of new in vivo data.
• Understanding coordinated somatic-dendritic activity is crucial for explaining cortical computations like nonlinear mixed selectivity, prediction, multiplexing, and credit assignment.