{
  "id": 6810135,
  "title": "Plateau-gated one-shot plasticity supports continual recognition memory",
  "url": "https://urgent.news/2026/09/11/plateau-gated-one-shot-plasticity-supports-continual-recognition",
  "topic": "science",
  "section": "Science",
  "published": "2026-09-11T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.09.04.749460v1?rss=1"
  },
  "original_language": "en",
  "account": "Biological memory systems retain individual experiences while simultaneously acquiring new knowledge, yet the mechanisms behind how one-time synaptic modifications avoid interfering with pre-existing memories remain elusive. Synaptic plasticity known as behavioral timescale synaptic plasticity (BTSP) rapidly alters synapses within seconds of a dendritic plateau. To better understand this phenomenon, researchers isolated the plateau-triggered component in a model where plastic weights and a consistent guiding pathway jointly dictate whether a plateau occurs, effectively creating a feedback loop between the synaptic state and the plastic event that shapes it.\n\nWhen confronted with unstructured inputs, represented by independent random signed patterns, the system's dynamics align precisely with the guidance pathway, which in turn establishes the accuracy of the instructed representation, the rate at which synapses are replaced, and the average interval between rewrites. Conversely, when presented with structured inputs, embodied by correlated bimodal Curie--Weiss patterns, the guiding pathway influences which aspect of the input structure is incorporated into the synaptic state.\n\nIn a BTSP-inspired continual-recognition network, the combined guidance and plastic influences determine which memory unit is chosen for each one-time encoding, while a Hebbian control uses the same plastic weights for assigning credit and storing memories. The BTSP-inspired network exhibited superior performance over extended time intervals compared to Hebbian controls, with the advantage increasing as the network grew larger, and each architecture was optimized independently for each time interval.\n\nA simplified theory accurately predicted the network's effectiveness, capacity for handling delays, and the strength of memory traces directly from the optimized parameters. This theory elucidated why a balance of proximal and distal coupling proved optimal: proximal plastic drive steered the generation of plateaus towards the intended memory units, limiting synaptic turnover but hindering new learning, while distal guidance strengthened familiar responses but also had the potential to make novel inputs seem familiar. The memory trace strength in the Hebbian control decayed more rapidly and exhibited less efficient credit assignment compared to the BTSP-inspired network. These findings establish a connection between dendritic plateau physiology and the persistence of memory, and offer support for the partial separation of allocation and storage as a strategy to minimize interference during continuous learning.",
  "summary": "Biological memory systems store single experiences while continuing to learn, but how one-shot plasticity limits interference with existing memories is unclear. Behavioral timescale synaptic plasticity (BTSP) rapidly modifies synapses active within seconds of a dendritic plateau. We isolate its plateau-triggered component in a model where plastic weights and a stable instructive pathway jointly…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}