A genuine layer 4 in motor cortex with prototypical synaptic circuit connectivity

Naoki Yamawaki, Katharine Borges, Benjamin A. Suter, Kenneth D. Harris, Gordon M. ShepherdeLife 2014 · DOI

TL;DR

The primary motor cortex (M1) has long been considered "agranular," lacking a distinct Layer 4 (L4) found in sensory areas. This study provides compelling evidence that M1 does possess a functional L4 equivalent, characterized by neurons at the L3/5A border exhibiting the same core synaptic connectivity as L4 neurons in sensory cortices, challenging traditional views of motor cortex organization.

What Problem Does This Paper Address?

The prevailing view in neuroscience is that the primary motor cortex (M1), responsible for planning and executing voluntary movements, is "agranular." This classification stems from its apparent lack of Layer 4 (L4), a distinct cortical layer that plays a crucial role in processing sensory information in sensory cortices. L4 neurons in sensory areas are characterized by specific input-output connections: strong excitatory input from the thalamus, unidirectional projections to superficial layers (L2/3), and limited long-range connections with other cortical areas. The absence of this layer in M1 has led to theories suggesting fundamental differences in the neural circuitry and function of motor versus sensory cortices. This paper aims to challenge this assumption by investigating whether M1 harbors neurons that, despite lacking the typical cytoarchitectural appearance of L4, exhibit its characteristic synaptic connectivity.

Methods

The researchers employed a multi-faceted approach combining optogenetics, electrophysiology, and advanced imaging techniques.

  • Optogenetic Tracing: They used adeno-associated viruses (AAVs) encoding channelrhodopsin-2 (ChR2) to express light-sensitive proteins in specific neuronal populations. Injections into the motor thalamus (ventrolateral nucleus, VL, and posterior nucleus, PO) allowed them to map thalamocortical (TC) inputs to M1. Similarly, injections into contralateral M1 and ipsilateral M2, S1, and S2 were used to map long-range corticocortical connections.
  • Electrophysiology: Whole-cell patch-clamp recordings were performed on neurons in acute brain slices of mouse M1. By stimulating ChR2-expressing axons with blue light, they could measure monosynaptic excitatory postsynaptic currents (EPSCs) and characterize the input-output connectivity of individual neurons across different cortical layers.
  • Morphological Reconstruction: Neurons were filled with biocytin during recordings, and their detailed dendritic and axonal structures were reconstructed using two-photon microscopy and custom software. This allowed for quantitative analysis of neuronal morphology and laminar projection patterns.
  • Molecular Markers: In situ hybridization for the Rorb gene was used to visualize potential laminar correlates of L4, as Rorb expression is known to be high in L4 of sensory cortices.

Datasets

The study utilized data from experiments performed on mice. No specific public datasets were used for training or evaluation. The primary "dataset" was generated through the experimental recordings and imaging from the research animals.

Key Insights and Ideas

  • M1's Hidden Layer 4: The most significant insight is the identification of a functionally distinct L4 in M1, despite its lack of clear cytoarchitectural definition. This suggests that cortical organization principles might be more conserved across areas than previously thought.
  • Prototypical Connectivity: The neurons identified at the L3/5A border in M1 exhibit the key input-output features of L4 neurons in sensory areas: strong thalamic input, unidirectional output to L2/3, and limited long-range cortical connections. This strongly supports their functional equivalence.
  • Pyramidal Neurons as L4: Unlike some sensory cortices where L4 can include specialized "stellate" cells, the putative M1-L4 neurons are consistently pyramidal. This suggests a different morphological substrate for L4 function in motor cortex.
  • Area-Specific Nuances: While sharing core L4 properties, M1-L4 neurons also show unique characteristics, such as receiving input from both VL and PO thalamic nuclei, and having axonal projections that extend into the white matter, similar to some S1 pyramidal neurons.

Key Results

  • Thalamocortical Input: Neurons at the L3/5A border in M1 received strong, monosynaptic excitatory input from the motor thalamus (VL), mirroring L4 input in sensory cortices. This input was also observed from the PO nucleus, though with a broader laminar distribution.
  • Output to L2/3: M1-L4 neurons projected to L2/3 pyramidal neurons, and this projection was largely unidirectional, with minimal reciprocal input from L2/3 back to L4. This is a hallmark of L4 connectivity in sensory areas.
  • Limited Long-Range Connections: M1-L4 neurons received and sent relatively little input from/to other cortical areas, particularly contralateral M1, consistent with the L4 pattern in sensory cortices.
  • Morphological Uniformity: While electrophysiologically diverse, the M1-L4 neurons were morphologically uniform, exhibiting pyramidal-like dendritic arbors and locally ramifying axons.
  • Rorb Expression: The study confirmed that Rorb, a marker for L4 in sensory areas, also shows a distinct laminar expression pattern at the L3/5A border in M1, further supporting the presence of an L4 equivalent.

Limitations and Future Work

  • Cytoarchitectural Absence: The primary limitation acknowledged is the lack of a distinct cytoarchitectural L4 in M1, meaning the identified neurons are functionally, not morphologically, defined as L4.
  • Cell-to-Cell Variability: While general patterns were observed, there was considerable cell-to-cell variability in electrophysiological properties, particularly in firing patterns.
  • Future Directions: The authors suggest that future work should shift from questioning the presence of L4 in M1 to investigating what specific types of information these M1-L4 neurons process and how this processing relates to motor behavior. Understanding the role of these neurons in integrating thalamic input for motor control is a key next step.

Summary updated

Abstract

The motor cortex (M1) is classically considered an agranular area, lacking a distinct layer 4 (L4). Here, we tested the idea that M1, despite lacking a cytoarchitecturally visible L4, nevertheless possesses its equivalent in the form of excitatory neurons with input-output circuits like those of the L4 neurons in sensory areas. Consistent with this idea, we found that neurons located in a thin laminar zone at the L3/5A border in the forelimb area of mouse M1 have multiple L4-like synaptic connections: excitatory input from thalamus, largely unidirectional excitatory outputs to L2/3 pyramidal neurons, and relatively weak long-range corticocortical inputs and outputs. M1-L4 neurons were electrophysiologically diverse but morphologically uniform, with pyramidal-type dendritic arbors and locally ramifying axons, including branches extending into L2/3. Our findings therefore identify pyramidal neurons in M1 with the expected prototypical circuit properties of excitatory L4 neurons, and question the traditional assumption that motor cortex lacks this layer.
BibTeX
@inproceedings{yamawaki2014genuine,
  author = {Naoki Yamawaki and Katharine Borges and Benjamin A. Suter and Kenneth D. Harris and Gordon M. Shepherd},
  title = {{A genuine layer 4 in motor cortex with prototypical synaptic circuit connectivity}},
  booktitle = {eLife},
  year = {2014},
  doi = {10.7554/elife.05422},
}
Figure