Postsynaptic burst reactivation of hippocampal neurons enables associative plasticity of temporally discontiguous inputs

Tanja Fuchsberger, Claudia Clopath, Przemysław Jarzębowski, Zuzanna Brzosko, Hongbing Wang, Ole Paulsen

TL;DR

This paper introduces a novel mechanism for how the brain forms associations between events separated in time, a key challenge in memory formation. They demonstrate that a specific pattern of neuronal activity – a "burst reactivation" – in the presence of dopamine can strengthen synaptic connections that were previously "primed" but inactive for minutes. This discovery offers a new perspective on reinforcement learning and memory encoding in the hippocampus.

What Problem Does This Paper Address?

The brain's ability to link events that occur minutes or even hours apart is fundamental to learning and memory. Traditional models of synaptic plasticity, like spike-timing-dependent plasticity (STDP), typically operate on much shorter timescales (milliseconds). While neuromodulators like dopamine are known to play a role in reinforcement learning by signaling reward, it's unclear how they help bridge these temporal gaps and assign credit to specific past events. This paper tackles the question of how the brain can establish long-lasting associations between temporally discontiguous events, proposing a mechanism that leverages neuronal "reactivation" and dopamine signaling.

Methods

The researchers employed a multi-pronged approach:

  • Electrophysiology in Brain Slices: They used whole-cell patch-clamp recordings from mouse hippocampal CA1 neurons. A key technique was a "priming" protocol (post-before-pre pairing) that induced synaptic depression, creating a "silent eligibility trace." They then tested how subsequent "burst reactivation" (a rapid series of action potentials) in the presence of dopamine affected these primed synapses.
  • Pharmacological Manipulations: They used various drugs to block specific molecular pathways (e.g., NMDARs, mGluRs, PKA, protein synthesis inhibitors) to pinpoint the signaling cascades involved in this novel plasticity.
  • Computational Modeling: They built a neural network model to simulate their proposed synaptic learning rule. This allowed them to explore how this rule could influence reinforcement learning, particularly in terms of memory allocation and specificity.
  • In Vivo Calcium Imaging: They used head-mounted microscopes to record neuronal activity in freely moving mice performing a reward-based navigation task. This allowed them to observe neuronal activity patterns associated with reward approach and immobility, and to identify "reactivated" neurons.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • The "Silent Eligibility Trace": The priming protocol doesn't immediately change synaptic strength but leaves a latent trace that can be modified later. This is a crucial concept for bridging temporal gaps.
  • Burst Reactivation as a "Tag": Postsynaptic bursts, especially when paired with dopamine, act as a signal that "activates" these silent traces, leading to potentiation. This suggests that not just the timing of spikes, but their pattern (bursting) is critical.
  • Dopamine as a Coincidence Detector: Dopamine, in conjunction with intracellular calcium influx triggered by bursts, acts as a coincidence detector, enabling the conversion of the eligibility trace into long-term potentiation (LTP).
  • A Novel Signaling Pathway: The mechanism involves a pathway distinct from conventional LTP, highlighting the complexity of synaptic plasticity. It requires adenylyl cyclases (AC1/AC8), PKA, and protein synthesis, but the induction is surprisingly rapid.
  • Enhanced Specificity in Reinforcement Learning: The computational model showed that this burst-dependent plasticity makes reinforcement learning more specific, allowing the network to allocate memories more precisely and avoid "overwriting" information.

Key Results

  • Synaptic Potentiation: Postsynaptic bursts in the presence of dopamine, 10 minutes after a priming event, induced input-specific LTP in hippocampal synapses.
  • Long-Lasting Eligibility Traces: The "silent eligibility trace" created by the priming protocol persisted for at least 10 minutes, significantly longer than previously reported.
  • Role of Postsynaptic NMDARs: Postsynaptic NMDARs were essential for setting the eligibility trace, while mGluRs and non-postsynaptic NMDARs were involved in the initial depression.
  • Calcium Channels and ACs/PKA: Voltage-gated calcium channels and AC1/AC8 were required for DA- and burst-induced potentiation, as was PKA. Protein synthesis was also necessary for the expression of this potentiation.
  • In Vivo Validation: In freely moving mice, neurons that reactivated at reward locations showed significantly higher calcium activity and carried more spatial information than non-reactivated neurons, supporting the in vitro findings.

Limitations and Future Work

  • External Calcium Concentration: Experiments were conducted at 2 mM external calcium, which is higher than physiological levels. Further investigation is needed to see if the mechanism holds at lower concentrations.
  • Temporal Control of Dopamine: While dopamine is crucial, precise temporal control of its release in vivo was not achieved in these experiments. Future work could use optogenetic or uncaging techniques for better temporal resolution.
  • Mechanism of Protein Synthesis Involvement: The rapid induction of protein synthesis-dependent potentiation is surprising and warrants further investigation into the underlying molecular mechanisms.
  • Role of Other Brain Areas: The study focused on the hippocampus. Understanding how this mechanism interacts with other brain regions involved in reward and memory would be valuable.
  • In Vivo Stability: While reactivated cells showed increased activity and spatial information, the study did not observe a significant increase in place cell stability after a single reactivation event. Further research is needed to understand the conditions under which this mechanism influences long-term stability.

Summary updated

Abstract

A fundamental unresolved problem in neuroscience is how the brain associates in memory events that are separated in time. Here, we propose that reactivation-induced synaptic plasticity can solve this problem. Previously, we reported that the reinforcement signal dopamine converts hippocampal spike timing-dependent depression into potentiation during continued synaptic activity (Brzosko et al., 2015). Here, we report that postsynaptic bursts in the presence of dopamine produce input-specific LTP in mouse hippocampal synapses 10 min after they were primed with coincident pre- and post-synaptic activity (post-before-pre pairing; Δt = -20 ms). This priming activity induces synaptic depression and sets an NMDA receptor-dependent silent eligibility trace which, through the cAMP-PKA cascade, is rapidly converted into protein synthesis-dependent synaptic potentiation, mediated by a signaling pathway distinct from that of conventional LTP. This synaptic learning rule was incorporated into a computational model, and we found that it adds specificity to reinforcement learning by controlling memory allocation and enabling both 'instructive' and 'supervised' reinforcement learning. We predicted that this mechanism would make reactivated neurons activate more strongly and carry more spatial information than non-reactivated cells, which was confirmed in freely moving mice performing a reward-based navigation task.
BibTeX
@inproceedings{fuchsberger2022postsynaptic,
  author = {Tanja Fuchsberger and Claudia Clopath and Przemysław Jarzębowski and Zuzanna Brzosko and Hongbing Wang and Ole Paulsen},
  title = {{Postsynaptic burst reactivation of hippocampal neurons enables associative plasticity of temporally discontiguous inputs}},
  booktitle = {eLife},
  year = {2022},
  doi = {10.7554/elife.81071},
}
Figure