TL;DR
This paper explores how the hippocampus might store information by adjusting the strength of connections between neurons. They demonstrate a novel form of synaptic plasticity called "associative long-term depression" (LTD) that works in tandem with known "associative long-term potentiation" (LTP). Together, these mechanisms allow neurons to encode the temporal relationships (covariance) between incoming signals, a crucial aspect for memory formation.
What Problem Does This Paper Address?
The prevailing idea in neural network models for memory storage is that synaptic strengths need to be selectively enhanced (potentiation) and weakened (depression) to efficiently encode information. While long-term potentiation (LTP), a persistent strengthening of synapses, is well-established in the hippocampus (a brain region critical for long-term memory), a corresponding mechanism for persistent synaptic weakening that could balance LTP has been elusive. This paper aims to investigate whether such a mechanism exists and how it might function in the hippocampus, specifically looking for conditions that could lead to a decrease in synaptic strength.
Methods
The researchers used an in vitro hippocampal slice preparation from rats. They stimulated two distinct input pathways onto CA1 pyramidal neurons: a "strong" input (Schaffer collaterals) and a "weak" input (commissural afferents). The key experimental manipulation involved the temporal relationship between the stimulation of these two inputs.
- Homosynaptic LTP: They first confirmed that stimulating the strong pathway alone produced homosynaptic LTP, a well-known phenomenon where a pathway strengthens itself after high-frequency stimulation.
- Associative LTP: When the weak input was stimulated in phase with the bursts of the strong input, it led to an increase in the weak pathway's synaptic strength (associative LTP). This aligns with Hebbian principles ("neurons that fire together, wire together").
- Associative LTD: Crucially, when the weak input was stimulated out of phase with the strong input bursts, it resulted in a persistent decrease in the weak pathway's synaptic strength (associative LTD). This is a novel finding, suggesting an "anti-Hebbian" form of plasticity.
- Intracellular Recordings: To further investigate the cellular mechanisms, they used intracellular recordings. This allowed them to directly measure postsynaptic potentials and observe how synaptic strength changes affected neuronal firing. They also manipulated postsynaptic membrane potential (depolarization or hyperpolarization) while stimulating inputs to probe the conditions necessary for LTP and LTD.
Datasets
Not discussed in the paper. The experiments were conducted using in vitro hippocampal slices from rats.
Key Insights and Ideas
- Dual Plasticity for Covariance: The most significant insight is the demonstration of associative LTD, which, alongside associative LTP, provides a mechanism for storing covariance between inputs. This means the hippocampus can encode not just the presence of signals but also their temporal relationships.
- Anti-Hebbian Plasticity: Associative LTD is characterized as "anti-Hebbian," occurring when presynaptic firing is paired with postsynaptic hyperpolarization. This contrasts with Hebbian LTP, which requires postsynaptic depolarization.
- Temporal Correlation is Key: The precise timing of weak input stimulation relative to strong input bursts is critical. Being "in phase" leads to strengthening, while being "out of phase" leads to weakening.
- NMDA Receptor Dependence: Associative LTP induction appears to depend on NMDA receptors, a common finding. However, associative LTD induction does not seem to require NMDA receptors, suggesting different underlying molecular mechanisms.
- Cellular Mechanisms for LTD: Postsynaptic hyperpolarization, in conjunction with presynaptic activity, is sufficient to induce LTD. This provides a cellular basis for how the "out of phase" timing leads to synaptic weakening.
Key Results
- Associative LTP: Stimulation of weak input in phase with strong input bursts led to a significant, long-lasting enhancement of weak input synaptic strength (e.g., +49.8% for EPSP, +65.4% for population spike).
- Associative LTD: Stimulation of weak input out of phase with strong input bursts resulted in a marked, persistent reduction of weak input synaptic strength (e.g., -46.5% for population spike, -13.8% for EPSP).
- Input Specificity: Both associative LTP and LTD were specific to the activated pathway. Unstimulated inputs remained unchanged.
- Control Experiments: Weak input stimulation alone did not induce lasting changes. Strong input stimulation alone induced homosynaptic LTP but did not affect the weak input. A control pattern with zero covariance between inputs also produced no net change.
- Intracellular Evidence: Intracellular recordings confirmed that associative LTP increased EPSP amplitude and lowered action potential threshold, while associative LTD reduced EPSP amplitude and firing ability.
Limitations and Future Work
- Model vs. Reality: While the study demonstrates these mechanisms in hippocampal slices, translating these findings directly to complex memory formation in a living brain requires further investigation.
- Molecular Mechanisms: The paper notes that LTD induction differs from LTP in its NMDA receptor dependence, suggesting different downstream molecular pathways. Future work could delve deeper into these specific mechanisms.
- Broader Implications: The authors suggest that similar Hebbian and anti-Hebbian mechanisms might be present in other brain areas and are crucial for higher cognitive functions. Further research could explore these across different brain regions and their role in more complex behaviors.
- Duration and Magnitude: The paper mentions that the amplitude and duration of associative LTD or LTP could be increased by stimulating input pathways with more trains of shocks, hinting at further parameters to explore.
Summary updated