#ResultatScientifique🔎| L’hippocampe combine signaux externes et états émotionnels pour adapter le comportement face à l’environnement.
✍️ @mnpompili.bsky.social
📕 buff.ly/DulC2E9
▶️ buff.ly/WwfQgfo
🌐 Overall, our results show how distinct hippocampal subregions encode complementary aspects of associative learning — learning and expression — and how population-level coordination across the HPC dorsoventral axis integrates these signals to support adaptive behavior.
12/12
🧠✨ These multifactorial engrams, distributed across the dorsoventral axis of the hippocampus, provide a population-level substrate for integrating emotional learning with its associated behavioral response — offering a mechanism for coordinating associative learning.
11/12
🔗 These mixed assemblies combined neurons encoding fear expression with neurons encoding fear learning — effectively binding behavior and associative value at the population level.
10/12
🧩 Crucially, these signals did not remain segregated and converged in cell assemblies composed of both dHPC and vHPC neurons that emerged during conditioning.
9/12
🧭 Population activity revealed a dorsoventral division of labor in the hippocampus:
• dHPC → fear expression
• vHPC → fear learning / cue valence
with representations mapped onto complementary components of the emotional experience.
This raised a key question: how are these signals combined?
8/12
🔔 In contrast, vHPC neurons developed robust excitatory responses to the conditioned stimulus, independent of freezing behavior — tracking the acquisition of conditioned fear rather than its expression.
7/12
🧊 Importantly, dHPC responses were specific to fearful freezing — not immobility or motor confounds — and were context-dependent, pointing to a role for dHPC (but not vHPC) in encoding contextual aspects of learned fear expression.
6/12
⚠️➡️ Unexpectedly, neuronal correlates of fear expression (freezing) were more prominent in the dorsal hippocampus than in the ventral hippocampus — challenging the long-standing view that vHPC is the main locus of fear regulation.
5/12
📈 As expected, fear conditioning induced large-scale shifts in firing rates in both regions, with neurons increasing or decreasing their activity following acquisition of the CS–US association.
4/12
🔬 Here, we performed simultaneous single-unit recordings in dorsal and ventral hippocampus during fear conditioning — to our knowledge, the first time this has been done.
This let us directly compare population dynamics as fear is learned and expressed.
3/12
📚 Classically, the dorsal hippocampus (dHPC) is linked to contextual/spatial processing, while the ventral hippocampus (vHPC) is associated with emotion and anxiety.
Both are required for fear conditioning — but how their contributions are integrated remained unknown.
2/12
Everyone knows dorsal and ventral hippocampus do different things.
But how do neurons in these regions differ in function — and how are their contributions integrated?
We tackled this with @noehamou.bsky.social and Sid Wiener in a paper just out in @pnas.org 👇🧵
1/12
www.pnas.org/doi/abs/10.1...
I am more than pleased to share our new work with you:
"Intra- and Interhemispheric Signatures of Criticality at the Onset of Synchronization"
biorxiv.org/content/10.6...
Short 🧵 1/9:
New preprint out!
How does criticality propagate from local neuronal circuits to whole-brain dynamics?
We tackle this with a multiscale, connectome-based mouse model @ldallap.bsky.social . 🧵/n
👉 biorxiv.org/content/10.6...
Happy to announce the 2nd Hippocampus Green Meeting
📍 Barcelona, Spain
🗓️ May 11–12, 2026
Organized together with Manu Valero, Lisa Roux and Dan Bendor
🎤 Keynotes: Nachum Ulanovsky & György Buzsáki
‼️ Call for abstracts now open
🔗 hippocampusgreen.net/wp/
#HippocampusGreen
…and we have another work coming where we propose a way to study cell assemblies without binning or z-scoring. Stay tuned!
Binning is necessary for the most common method of cell assembly detection that we also used in this study (ICA), but we agree this does have its limitations, which we discuss here: doi.org/10.1007/978-...
It remains unclear whether such co-firing truly matters for brain function. Here, we show that in the prefronto-amygdalar circuit, “cell assemblies” matter on timescales of up to ~40 ms: synchronous activity over longer windows fails to trigger a supralinear downstream response.
What cell assemblies are — and what they do for the brain — is a profound question and the central motivation behind this work! Most people define an assembly as a set of neurons that fire together (thereby enabling the plasticity associated with synchrony).
In sum: (1) this is, to our knowledge, the first experimental support for the “reader-centric” framework of cell assemblies; (2) assembly reading implements both pattern completion and pattern separation; and (3) assembly–reader relationships can be flexibly modified by learning. 10/10
Finally, we expected that assembly–reader relationships could flexibly change to reflect learning processes. Indeed, fear conditioning induced changes in AMY assembly → mPFC reader couplings, and conversely, changes in mPFC → AMY reading followed fear extinction. 9/10
The assembly–reader mechanism implements both pattern completion (responses did not simply increase proportionally with the number of active members) and pattern separation (responses were selective for specific assemblies and discriminated between overlapping assemblies). 8/10
Assembly activations exerted synergistic effects on reader neurons (not just a linear summation of responses to individual assembly members), and the identity of participating members mattered beyond their compound activity (spikes from neurons A and B > two spikes from A alone). 7/10
We recorded large neural ensembles from the medial prefrontal cortex (mPFC) and the amygdala (AMY), which are reciprocally interconnected, and showed that assembly activation could trigger firing responses in downstream neurons. 6/10
With rtodorova.bsky.social, Céline Boucly, Sidney Wiener, and Michaël Zugaro, we evaluated the downstream impact of cell assemblies: how is the co-activation of their members special? And how does the “reading” of this activity relate to brain computation and behavior? 5/10
György Buzsáki suggested that activation of a cell assembly may be meaningful only if it triggers specific responses in one or more target neurons, and proposed studying cell assemblies in terms of their impact on downstream “reader” neurons. doi.org/10.1016/j.ne... 4/10
Indeed, cell assemblies have been found to be involved in perception, memory, and executive and motor functions. Nevertheless, it is still unknown whether such “representations” in specific neuronal populations are relevant for the rest of the brain. 3/10
Groups of neurons firing together have been proposed to constitute a computational unit of the brain since Donald Hebb’s theory (1949). Today, the notion of a cell assembly is used to designate a group of cells involved in an action or representing a percept or concept. 2/10
Cell assemblies are drawing increasing attention in neuroscience, but one could argue that they are just an epiphenomenon. Is the activity of cell assemblies relevant for the brain?
The short answer is yes. The long answer is in our paper, now online at PLOS Biology. 🧵👇 1/10
doi.org/10.1371/jour...
