Inhibition enhances spatially-specific calcium encoding of synaptic input patterns in a biologically constrained model

Abstract

Synaptic plasticity, which underlies learning and memory, depends on calcium elevation in neurons, but the precise relationship between calcium and spatiotemporal patterns of synaptic inputs is unclear. Here, we develop a biologically realistic computational model of striatal spiny projection neurons with sophisticated calcium dynamics, based on data from rodents of both sexes, to investigate how spatiotemporally clustered and distributed excitatory and inhibitory inputs affect spine calcium. We demonstrate that coordinated excitatory synaptic inputs evoke enhanced calcium elevation specific to stimulated spines, with lower but physiologically relevant calcium elevation in nearby non-stimulated spines. Results further show a novel and important function of inhibition—to enhance the difference in calcium between stimulated and non-stimulated spines. These findings suggest that spine calcium dynamics encode synaptic input patterns and may serve as a signal for both stimulus-specific potentiation and heterosynaptic depression, maintaining balanced activity in a dendritic branch while inducing pattern-specific plasticity.

Dopaminergic modulation of striatal function and Parkinson’s disease

“…The proximal and distal dendrites of SPNs differ in their regenerative capacity: brief glutamate uncaging at distal dendrites, but not proximal dendrites, evokes dendritic regenerative events that depolarize somatic potential for hundreds of milliseconds-so termed dendritic plateau potential (Plotkin et al 2011). Modeling predicts that spine calcium transients and dendritic plateau potentials are most effectively attenuated by distal GABAergic inputs at the site of excitatory inputs, whose primary function is to enhance input-specificity of local calcium elevation (Dorman et al 2018;Du et al 2017). Supporting this theoretical model, Du et al showed that dendritic plateau potentials are effectively attenuated by optogenetic activation of GABAergic synapses or GABA uncaging at "on-branch" dendritic site within a defined temporal window, but not by perisomatic GABAergic inhibition (Du et al 2017).…”

Deciphering the postsynaptic calcium-mediated energy homeostasis through mitochondria-endoplasmic reticulum contact sites using systems modeling

“…We recognize that a stochastic approach would likely give rise to insights in limit cases of few molecules or in noisy environments [37,38]. Finally, we note that a more complete Ca 2+ influx model would account for voltage-gated calcium channel dynamics as well [22,24,60,61]. This is a focus of technical development and ongoing effort in our group.…”

Synaptic plasticity is predicted by spatiotemporal firing rate patterns and robust to in vivo-like variability

“…Our finding that plasticity at a single synapse is influenced by neighboring synaptic activity is consistent with many studies (both in vivo and in vitro) showing relationships between spatial synaptic input patterns and plasticity. For instance, in SPNs in vitro, spatially clustered synaptic inputs can produce supralinear depolarization (termed NMDA-spikes or plateau potentials) and synaptic calcium transients (Plotkin et al, 2011;Dorman et al, 2018;Prager et al, 2020;Du et al, 2017), and synaptic activation of 2-4 neighboring spines at depolarized potentials can produce nonlinear enhancement of spine calcium transients (Carter et al, 2007). Building on these findings, our work predicts that neighboring synaptic interactions can influence the direction and magnitude of corticostriatal synaptic plasticity in vivo, with high neighboring activity producing LTP and low neighbor- Observations suggest that functional synaptic clustering is a key component in plasticity and learning.…”

“…Prior work has shown that synaptic plasticity can be affected by spatiotemporally cooperative synaptic activity-that is, multiple synapses on the same dendritic branch active within a limited time window (Govindarajan et al, 2011 shown that spatiotemporal activity patterns have nonlinear, spatially specific effects on calcium transients in dendrites and spines (Dorman et al, 2018). Thus, it is likely that nearby synaptic activity can cooperatively influence plasticity in our calcium-based model.…”

“…In spiny projection neurons (SPNs) of the striatum-input nucleus of the basal ganglia and the focus of this paper-calcium signaling may provide an eligibility trace for corticostriatal plasticity (either LTP or LTD) underlying reinforcement learning for goaldirected or habitual behavior (He et al, 2015;Kreitzer and Malenka, 2008). We have previously shown that synaptic calcium transients are highly sensitive to spatiotemporal patterns of synaptic input and can encode synapse-specificity and cooperativity of neighboring synaptic activity in an experimentally-validated biophysical SPN model with dendritic branches (Dorman et al, 2018). We have also shown that a calcium based plasticity rule with dual amplitude-and duration-thresholds can predict LTP and LTD outcomes of several in vitro plasticity experiments (Jędrzejewska-Szmek et al, 2017).…”