TL;DR
This study reveals how the lateral entorhinal cortex (LEC) uses both excitatory (glutamatergic) and inhibitory (GABAergic) projections to CA3 in the hippocampus to dynamically control memory processes like pattern separation (distinguishing similar experiences) and pattern completion (recalling a memory from partial cues). By manipulating these LEC inputs, the researchers found they could bias CA3 circuits towards either discrimination or generalization, offering a new mechanistic understanding of how the brain flexibly adapts to new and familiar environments.
What Problem Does This Paper Address?
The brain needs to do two seemingly opposite things with memories: distinguish between similar experiences (pattern separation) and generalize across similar ones (pattern completion). While the hippocampus, particularly area CA3, is known to be crucial for these processes, the precise neural circuits that enable this flexible control remain elusive. Specifically, how external inputs, like those from the entorhinal cortex (EC), interact with CA3's internal circuitry to achieve this balance is not well understood. This paper focuses on the lateral entorhinal cortex (LEC), a region known to process non-spatial contextual information, and investigates how its glutamatergic (excitatory) and GABAergic (inhibitory) projections to CA3 contribute to these fundamental memory operations.
Methods
The researchers employed a multi-pronged approach:
- Viral Tracing: They used genetically engineered viruses to label LEC neurons projecting to CA3, distinguishing between glutamatergic (LEC<sub>GLU</sub>) and GABAergic (LEC<sub>GABA</sub>) populations.
- Chemogenetics (DREADDs): To selectively silence LEC<sub>GLU</sub> or both LEC<sub>GLU</sub> and LEC<sub>GABA</sub> neurons, they used Designer Receptors Exclusively Activated by Designer Drugs (DREADDs). This allowed them to observe the behavioral and neural consequences of reducing LEC input.
- In Vivo Two-Photon Calcium Imaging: They imaged the activity of CA3 pyramidal neurons in head-fixed mice running on a treadmill. This allowed them to observe population-level neural dynamics in real-time under different conditions (familiar vs. novel environments, with or without LEC silencing).
- Ex Vivo Electrophysiology: They performed whole-cell patch-clamp recordings from CA3 neurons in acute brain slices. This enabled detailed investigation of the synaptic inputs from LEC<sub>GLU</sub> and LEC<sub>GABA</sub>, including their effects on neuronal excitability and integration of multiple inputs.
- Optogenetics: Used to precisely activate or inhibit specific neuronal populations and pathways, allowing for detailed circuit mapping.
- Computational Modeling: An attractor network model of CA3 was built and simulated to explore how the observed changes in population dynamics might influence pattern separation and completion.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- LEC's Bidirectional Control: The LEC doesn't just send one type of signal; it uses both excitatory (LEC<sub>GLU</sub>) and inhibitory (LEC<sub>GABA</sub>) pathways to CA3, and these pathways have distinct, yet complementary, roles.
- LEC<sub>GLU</sub> and Sparse Coding: Silencing LEC<sub>GLU</sub> led to more CA3 neurons being active but at lower rates, suggesting LEC<sub>GLU</sub> normally promotes sparse coding by recruiting feedforward inhibition. This inhibition seems to prevent CA3 neurons from firing excessively, thus creating a clearer signal for context discrimination.
- LEC<sub>GABA</sub> as a Disinhibitory Gate: LEC<sub>GABA</sub> doesn't directly excite CA3 neurons. Instead, it selectively inhibits interneurons that are themselves inhibiting CA3 neurons. This "disinhibition" allows CA3 neurons to integrate inputs more effectively, particularly promoting pattern completion.
- Compartment-Specific Effects: LEC<sub>GABA</sub>'s disinhibitory effect appears to be primarily somatic, boosting output from the neuron's cell body rather than its dendrites. This suggests a fine-tuned mechanism for controlling neuronal output.
- Dynamic Range for Memory: The interplay between LEC<sub>GLU</sub> (promoting discrimination/separation) and LEC<sub>GABA</sub> (promoting generalization/completion) provides CA3 with a dynamic range to flexibly switch between these crucial memory functions depending on the context and novelty of the information.
Key Results
- Behavioral: Silencing LEC<sub>GLU</sub> alone or together with LEC<sub>GABA</sub> impaired novel object recognition, suggesting a role for LEC inputs in distinguishing familiar from novel stimuli. Spatial learning in the Barnes maze was unaffected.
- Neural Activity (In Vivo): Silencing LEC<sub>GLU</sub> increased the number of active CA3 neurons but decreased their firing rates, indicating a role in sparsifying activity. It also reduced remapping between environments, suggesting it promotes generalization. Silencing both LEC<sub>GLU</sub> and LEC<sub>GABA</sub> reset activity patterns and increased remapping.
- Synaptic Physiology (Ex Vivo): LEC<sub>GLU</sub> stimulation evoked both excitatory (EPSCs) and inhibitory (IPSCs) responses in CA3, with the IPSCs mediated by feedforward inhibition. LEC<sub>GABA</sub> selectively inhibited CA3 interneurons, reducing LEC<sub>GLU</sub>-driven feedforward inhibition.
- Computational Model: The attractor network model, parameterized with experimental data, replicated the observed effects of LEC manipulation on CA3 population dynamics and predicted how these changes would bias the network towards pattern separation or completion.
Limitations and Future Work
- Compartment-Specific Mechanisms: While the study suggests LEC<sub>GABA</sub>'s disinhibitory effect is somatic, the precise molecular and cellular mechanisms for this compartment-specific action warrant further investigation. It's also unclear if LEC<sub>GABA</sub> targets other interneuron populations that might mediate dendritic disinhibition.
- Interplay with Other Inputs: The study focused on LEC inputs. Future work could explore how these LEC-driven mechanisms interact with inputs from other brain regions, such as the dentate gyrus (DG) and recurrent connections within CA3, to achieve a more complete understanding of CA3 function.
- Brain State and Task Dependence: The authors suggest future work should investigate how brain state and task demands specifically influence the contributions of LEC<sub>GLU</sub> and LEC<sub>GABA</sub> to CA3 plasticity and memory recall.
- LEC<sub>GABA</sub> Alone: The study couldn't experimentally silence LEC<sub>GABA</sub> alone. The computational model was used to predict its effects, highlighting the need for future experimental validation.
Summary updated