A form of synaptic plasticity operating on behavioral timescales has been proposed as a basis for learning and memory.1
This plasticity is a candidate bridge from biochemistry to cognition and is a focus of current exploration.11
Deep research (2026-07-20)
BTSP is induced by a single dendritic plateau potential (not repeated spike pairings), sufficient to create a new hippocampal CA1 place field in one trial, unlike Hebbian/STDP which needs many correlated spikes.23
Original demonstration (Bittner et al., Science 2017): five pairings of subthreshold presynaptic input with Ca2+ plateau potentials in hippocampal slices produced large, asymmetric, seconds-long potentiation.4
The plasticity window operates over seconds (not milliseconds), bidirectionally modifying synaptic weight for inputs active seconds before/after the plateau, described as a “credit assignment” rule for memory.32
Plateau potential initiation is gated by local feedback inhibition plus an instructive higher-order-region input, tying plasticity timing to behavioral/experiential context.3
Molecular mechanism (Jain et al., Nature, Max Planck Florida Institute; NIH + Max Planck Society funding): CaMKII is required for BTSP but shows delayed, stochastic activation 10–100 s after induction across a broad dendritic region rather than at the specific potentiated synapse — termed “dendritic, delayed, stochastic CaMKII activation” (DDSC).5
Optogenetically blocking CaMKII 15–30 s after the BTSP induction protocol abolished synaptic potentiation, confirming DDSC’s causal role.5
BTSP occurs in CA1 more frequently than in CA3 and is more prevalent during exploration of novel/new contexts than familiar ones, per computational modeling matched to in vivo place-field shifting dynamics (Madar et al., Nat Neurosci 2025).6
BTSP has since been shown to generate non-spatial representations too: two-photon Ca2+ imaging of mouse CA1 during an odor-cued working-memory task revealed plateau-like Ca2+ events forming stable odor representations, modulated differentially by medial vs. lateral entorhinal cortex input.7
Clinical/BCI relevance: because BTSP enables one-shot, non-Hebbian synaptic weight updates over a seconds-long eligibility trace, it is being explored as a candidate biologically plausible learning rule for neuromorphic and few-shot online BCI decoders, replacing millisecond-scale STDP credit assignment.83
A related computational model (Nat Commun 2024) shows a simplified BTSP rule provides content-addressable memory with binary synapses and one-shot learning, useful for in-memory computing hardware — directly relevant to neuromorphic BCI chips.9
Related work
Bittner KC, Milstein AD, Grienberger C, Romani S, Magee JC. “Behavioral time scale synaptic plasticity underlies CA1 place fields.” Science, 2017.4
Madar AD, et al. “Synaptic plasticity rules driving representational shifting in the hippocampus.” Nat Neurosci, 2025.6
Jain A, Nakahata Y, Yasuda R, et al. “Dendritic, delayed, stochastic CaMKII activation in behavioural time scale plasticity.” Nature, 2024.5