The spike-timing dependence of plasticity

In spike-timing-dependent plasticity (STDP), the order and precise temporal interval between presynaptic and postsynaptic spikes determine the sign and magnitude of long-term potentiation (LTP) or depression (LTD). STDP is widely utilized in models of circuit-level plasticity, development, and learning. However, spike timing is just one of several factors (including firing rate, synaptic cooperativity, and depolarization) that govern plasticity induction, and its relative importance varies across synapses and activity regimes. This review summarizes this broader view of plasticity, including the forms and cellular mechanisms for the spike-timing dependence of plasticity, and, the evidence that spike timing is an important determinant of plasticity in?vivo.     

Neuromodulators control the polarity of spike-timing-dependent synaptic plasticity

Near coincidental pre- and postsynaptic action potentials induce associative long-term potentiation (LTP) or long-term depression (LTD), depending on the order of their timing. Here, we show that in visual cortex the rules of this spike-timing-dependent plasticity are not rigid, but shaped by neuromodulator receptors coupled to adenylyl cyclase (AC) and phospholipase C (PLC) signaling cascades. Activation of the AC and PLC cascades results in phosphorylation of postsynaptic glutamate receptors at sites that serve as specific "tags" for LTP and LTD. As a consequence, the outcome (i.e., whether LTP or LTD) of a given pattern of pre- and postsynaptic firing depends not only on the order of the timing, but also on the relative activation of neuromodulator receptors coupled to AC and PLC. These findings indicate that cholinergic and adrenergic neuromodulation associated with the behavioral state of the animal can control the gating and the polarity of cortical plasticity.     

Local field potential modeling predicts dense activation in cerebellar granule cells clusters under LTP and LTD control

Local field-potentials (LFPs) are generated by neuronal ensembles and contain information about the activity of single neurons. Here, the LFPs of the cerebellar granular layer and their changes during long-term synaptic plasticity (LTP and LTD) were recorded in response to punctate facial stimulation in the rat in vivo. The LFP comprised a trigeminal (T) and a cortical (C) wave. T and C, which derived from independent granule cell clusters, co-varied during LTP and LTD. To extract information about the underlying cellular activities, the LFP was reconstructed using a repetitive convolution (ReConv) of the extracellular potential generated by a detailed multicompartmental model of the granule cell. The mossy fiber input patterns were determined using a Blind Source Separation (BSS) algorithm. The major component of the LFP was generated by the granule cell spike Na(+) current, which caused a powerful sink in the axon initial segment with the source located in the soma and dendrites. Reproducing the LFP changes observed during LTP and LTD required modifications in both release probability and intrinsic excitability at the mossy fiber-granule cells relay. Synaptic plasticity and Golgi cell feed-forward inhibition proved critical for controlling the percentage of active granule cells, which was 11% in standard conditions but ranged from 3% during LTD to 21% during LTP and raised over 50% when inhibition was reduced. The emerging picture is that of independent (but neighboring) trigeminal and cortical channels, in which synaptic plasticity and feed-forward inhibition effectively regulate the number of discharging granule cells and emitted spikes generating "dense" activity clusters in the cerebellar granular layer.     

Tag-Trigger-Consolidation: A Model of Early and Late Long-Term-Potentiation and Depression

Changes in synaptic efficacies need to be long-lasting in order to serve as a substrate for memory. Experimentally, synaptic plasticity exhibits phases covering the induction of long-term potentiation and depression (LTP/LTD) during the early phase of synaptic plasticity, the setting of synaptic tags, a trigger process for protein synthesis, and a slow transition leading to synaptic consolidation during the late phase of synaptic plasticity. We present a mathematical model that describes these different phases of synaptic plasticity. The model explains a large body of experimental data on synaptic tagging and capture, cross-tagging, and the late phases of LTP and LTD. Moreover, the model accounts for the dependence of LTP and LTD induction on voltage and presynaptic stimulation frequency. The stabilization of potentiated synapses during the transition from early to late LTP occurs by protein synthesis dynamics that are shared by groups of synapses. The functional consequence of this shared process is that previously stabilized patterns of strong or weak synapses onto the same postsynaptic neuron are well protected against later changes induced by LTP/LTD protocols at individual synapses.     

Coexistence of Reward and Unsupervised Learning During the Operant Conditioning of Neural Firing Rates

A fundamental goal of neuroscience is to understand how cognitive processes, such as operant conditioning, are performed by the brain. Typical and well studied examples of operant conditioning, in which the firing rates of individual cortical neurons in monkeys are increased using rewards, provide an opportunity for insight into this. Studies of reward-modulated spike-timing-dependent plasticity (RSTDP), and of other models such as R-max, have reproduced this learning behavior, but they have assumed that no unsupervised learning is present (i.e., no learning occurs without, or independent of, rewards). We show that these models cannot elicit firing rate reinforcement while exhibiting both reward learning and ongoing, stable unsupervised learning. To fix this issue, we propose a new RSTDP model of synaptic plasticity based upon the observed effects that dopamine has on long-term potentiation and depression (LTP and LTD). We show, both analytically and through simulations, that our new model can exhibit unsupervised learning and lead to firing rate reinforcement. This requires that the strengthening of LTP by the reward signal is greater than the strengthening of LTD and that the reinforced neuron exhibits irregular firing. We show the robustness of our findings to spike-timing correlations, to the synaptic weight dependence that is assumed, and to changes in the mean reward. We also consider our model in the differential reinforcement of two nearby neurons. Our model aligns more strongly with experimental studies than previous models and makes testable predictions for future experiments.     

Spike timing-dependent LTP/LTD mediates visual experience-dependent plasticity in a developing retinotectal system

Sensory experience plays an instructive role in the development of the nervous system. Here we showed that visual experience can induce persistent modification of developing retinotectal circuits via spike timing-dependent plasticity (STDP). Pairing light stimuli with spiking of the tectal cell induced persistent enhancement or reduction of light-evoked responses, with a dependence on the relative timing between light stimulus and postsynaptic spiking similar to that for STDP. Using precisely timed sequential three-bar stimulation to mimic a moving bar, we showed that spike timing-dependent LTP/LTD can account for the asymmetric modification of the tectal cell receptive field induced by moving bar. Furthermore, selective inhibition of signaling mediated by brain-derived neurotrophic factor and nitric oxide, which are respectively required for light-induced LTP and LTD, interfered with moving bar-induced temporally specific changes in the tectal cell responses. Together, these findings suggest that STDP can mediate sensory experience-dependent circuit refinement in the developing nervous system.     

Phosphorylation of the AMPA receptor GluR1 subunit is required for synaptic plasticity and retention of spatial memory

Plasticity of the nervous system is dependent on mechanisms that regulate the strength of synaptic transmission. Excitatory synapses in the brain undergo long-term potentiation (LTP) and long-term depression (LTD), cellular models of learning and memory. Protein phosphorylation is required for the induction of many forms of synaptic plasticity, including LTP and LTD. However, the critical kinase substrates that mediate plasticity have not been identified. We previously reported that phosphorylation of the GluR1 subunit of AMPA receptors, which mediate rapid excitatory transmission in the brain, is modulated during LTP and LTD. To test if GluR1 phosphorylation is necessary for plasticity and learning and memory, we generated mice with knockin mutations in the GluR1 phosphorylation sites. The phosphomutant mice show deficits in LTD and LTP and have memory defects in spatial learning tasks. These results demonstrate that phosphorylation of GluR1 is critical for LTD and LTP expression and the retention of memories.     

Coactivation of pre- and postsynaptic signaling mechanisms determines cell-specific spike-timing-dependent plasticity

Synapses may undergo long-term increases or decreases in synaptic strength dependent on critical differences in the timing between pre-and postsynaptic activity. Such spike-timing-dependent plasticity (STDP) follows rules that govern how patterns of neural activity induce changes in synaptic strength. Synaptic plasticity in the dorsal cochlear nucleus (DCN) follows Hebbian and anti-Hebbian patterns in a cell-specific manner. Here we show that these opposing responses to synaptic activity result from differential expression of two signaling pathways. Ca2+/calmodulin-dependent protein kinase II (CaMKII) signaling underlies Hebbian postsynaptic LTP in principal cells. By contrast, in interneurons, a temporally precise anti-Hebbian synaptic spike-timing rule results from the combined effects of postsynaptic CaMKII-dependent LTP and endocannabinoid-dependent presynaptic LTD. Cell specificity in the circuit arises from selective targeting of presynaptic CB1 receptors in different axonal terminals. Hence, pre- and postsynaptic sites of expression determine both the sign and timing requirements of long-term plasticity in interneurons.     

Emergence of preferred firing sequences in large spiking neural networks during simulated neuronal development

Two main processes concurrently refine the nervous system over the course of development: cell death and selective synaptic pruning. We simulated large spiking neural networks (100 x 100 neurons "at birth") characterized by an early developmental phase with cell death due to excessive firing rate, followed by the onset of spike timing dependent synaptic plasticity (STDP), driven by spatiotemporal patterns of stimulation. The cell death affected the inhibitory units more than the excitatory units during the early developmental phase. The network activity showed the appearance of recurrent spatiotemporal firing patterns along the STDP phase, thus suggesting the emergence of cell assemblies from the initially randomly connected networks. Some of these patterns were detected throughout the simulation despite the activity-driven network modifications while others disappeared.     

Acetylcholine-gated current translates wake neuronal firing rate information into a spike timing-based code in Non-REM sleep,stabilizing neural network dynamics during memory consolidation

Sleep is critical for memory consolidation, although the exact mechanisms mediating this process are unknown. Combining reduced network models and analysis of in vivo recordings, we tested the hypothesis that neuromodulatory changes in acetylcholine (ACh) levels during non-rapid eye movement (NREM) sleep mediate stabilization of network-wide firing patterns, with temporal order of neurons’ firing dependent on their mean firing rate during wake. In both reduced models and in vivo recordings from mouse hippocampus, we find that the relative order of firing among neurons during NREM sleep reflects their relative firing rates during prior wake. Our modeling results show that this remapping of wake-associated, firing frequency-based representations is based on NREM-associated changes in neuronal excitability mediated by ACh-gated potassium current. We also show that learning-dependent reordering of sequential firing during NREM sleep, together with spike timing-dependent plasticity (STDP), reconfigures neuronal firing rates across the network. This rescaling of firing rates has been reported in multiple brain circuits across periods of sleep. Our model and experimental data both suggest that this effect is amplified in neural circuits following learning. Together our data suggest that sleep may bias neural networks from firing rate-based towards phase-based information encoding to consolidate memories.     

Pull-push neuromodulation of LTP and LTD enables bidirectional experience-induced synaptic scaling in visual cortex

Neuromodulatory input, acting on G protein-coupled receptors, is essential for the induction of experience-dependent cortical plasticity. Here we report that G-coupled receptors in layer II/III of visual cortex control the polarity of synaptic plasticity through a pull-push regulation of LTP and LTD. In slices, receptors coupled to Gs promote LTP while suppressing LTD; conversely, receptors coupled to Gq11 promote LTD and suppress LTP. In vivo, the selective stimulation of Gs- or Gq11-coupled receptors brings the cortex into LTP-only or LTD-only states, which allows the potentiation or depression of targeted synapses with visual stimulation. The pull-push regulation of LTP/LTD occurs via direct control of the synaptic plasticity machinery and it is independent of changes in NMDAR activation or neuronal excitability. We propose these simple rules governing the pull-push control of LTP/LTD form a general metaplasticity mechanism that may contribute to neuromodulation of plasticity in other cortical circuits.     

Computational investigation of the changing patterns of subtype specific NMDA receptor activation during physiological glutamatergic neurotransmission

NMDA receptors (NMDARs) are the major mediator of the postsynaptic response during synaptic neurotransmission. The diversity of roles for NMDARs in influencing synaptic plasticity and neuronal survival is often linked to selective activation of multiple NMDAR subtypes (NR1/NR2A-NMDARs, NR1/NR2B-NMDARs, and triheteromeric NR1/NR2A/NR2B-NMDARs). However, the lack of available pharmacological tools to block specific NMDAR populations leads to debates on the potential role for each NMDAR subtype in physiological signaling, including different models of synaptic plasticity. Here, we developed a computational model of glutamatergic signaling at a prototypical dendritic spine to examine the patterns of NMDAR subtype activation at temporal and spatial resolutions that are difficult to obtain experimentally. We demonstrate that NMDAR subtypes have different dynamic ranges of activation, with NR1/NR2A-NMDAR activation sensitive at univesicular glutamate release conditions, and NR2B containing NMDARs contributing at conditions of multivesicular release. We further show that NR1/NR2A-NMDAR signaling dominates in conditions simulating long-term depression (LTD), while the contribution of NR2B containing NMDAR significantly increases for stimulation frequencies that approximate long-term potentiation (LTP). Finally, we show that NR1/NR2A-NMDAR content significantly enhances response magnitude and fidelity at single synapses during chemical LTP and spike timed dependent plasticity induction, pointing out an important developmental switch in synaptic maturation. Together, our model suggests that NMDAR subtypes are differentially activated during different types of physiological glutamatergic signaling, enhancing the ability for individual spines to produce unique responses to these different inputs.     

Disruption of MET Receptor Tyrosine Kinase,an Autism Risk Factor,Impairs Developmental Synaptic Plasticity in the Hippocampus

As more genes conferring risks to neurodevelopmental disorders are identified, translating these genetic risk factors into biological mechanisms that impact the trajectory of the developing brain is a critical next step. Here, we report that disrupted signaling mediated MET receptor tyrosine kinase (RTK), an established risk factor for autism spectrum disorders, in the developing hippocampus glutamatergic circuit leads to profound deficits in neural development, synaptic transmission, and plasticity. In cultured hippocampus slices prepared from neonatal mice, pharmacological inhibition of MET kinase activity suppresses dendritic arborization and disrupts normal dendritic spine development. In addition, single‐neuron knockdown (RNAi) or overexpression of Met in the developing hippocampal CA1 neurons leads to alterations, opposite in nature, in basal synaptic transmission and long‐term plasticity. In forebrain‐specific Met conditional knockout mice (Met^fx/fx;emx1^cre), an enhanced long‐term potentiation (LTP) and long‐term depression (LTD) were observed at early developmental stages (P12–14) at the Schaffer collateral to CA1 synapses compared with wild‐type littermates. In contrast, LTP and LTD were markedly reduced at young adult stage (P56–70) during which wild‐type mice show robust LTP and LTD. The altered trajectory of synaptic plasticity revealed by this study indicate that temporally regulated MET signaling as an intrinsic, cell autonomous, and pleiotropic mechanism not only critical for neuronal growth and functional maturation, but also for the timing of synaptic plasticity during forebrain glutamatergic circuits development.     

Interplay of the magnitude and time-course of postsynaptic Ca<Superscript>2 + </Superscript> concentration in producing spike timing-dependent plasticity

Synaptic strength can be modified by the relative timing of pre- and postsynaptic activity, a phenomenon termed spike timing-dependent plasticity (STDP). Studies of neurons in the hippocampus and in other regions have found that when presynaptic activity occurs within a narrow time window, typically 10 or 20 ms, before postsynaptic activity, long-term potentiation (LTP) is induced, while if presynaptic activity occurs within a similar time window after postsynaptic activity, long-term depression (LTD) results. The mechanisms underlying these modifications are not completely understood, although there is strong evidence that the postsynaptic Ca ^2 + concentration plays a central role. Some previous modeling of STDP has focused on the dynamics of the postsynaptic Ca ^2 + concentration, while other work has studied biophysical mechanisms of how a synapse can exist in, and switch between, different states corresponding to LTP and LTD. Building on previous work in these two areas we have developed the first low level STDP model of a tristable biochemical system that incorporates induction and maintenance of both LTP and LTD. Our model is able to explain the STDP observed in hippocampal neurons in response to pre- and postsynaptic pulse pairs, using only parameters derived from previous work and without the need for parameter fine-tuning. Our results also give insight into how and why the time course of the postsynaptic Ca ^2 + concentration can lead to either LTP or LTD, and suggest that voltage dependent calcium channels play a key role.     

Coincident activity of converging pathways enables simultaneous long-term potentiation and long-term depression in hippocampal CA1 network in vivo

Memory is believed to depend on activity-dependent changes in the strength of synapses, e.g. long-term potentiation (LTP) and long-term depression (LTD), which can be determined by the sequence of coincident pre- and postsynaptic activity, respectively. It remains unclear, however, whether and how coincident activity of converging efferent pathways can enable LTP and LTD in the pathways simultaneously. Here, we report that, in pentobarbital-anesthetized rats, stimulation (600 pulses, 5 Hz) to Schaffer preceding to commissural pathway within a 40-ms timing window induced similar magnitudes of LTP in both pathways onto synapses of CA1 neurons, with varied LTP magnitudes after reversal of the stimulation sequence. In contrast, in urethane-anesthetized or freely-moving rats, the stimulation to Schaffer preceding to commissural pathway induced Schaffer LTP and commissural LTD simultaneously within a 40-ms timing window, without affecting synaptic efficacy in the reversed stimulation sequence. Coincident activity of Schaffer pathways confirmed the above findings under pentobarbital and urethane anesthesia. Thus, coincident activity of converging afferent pathways tends to switch the pathways to be LTP only or LTP/LTD depending on the activity states of the hippocampus. This network rule strengthens the view that activity-dependent synaptic plasticity may well contribute to memory process of the hippocampal network with flexibility or stability from one state to another.     

Growth Dynamics Explain the Development of Spatiotemporal Burst Activity of Young Cultured Neuronal Networks in Detail

A typical property of isolated cultured neuronal networks of dissociated rat cortical cells is synchronized spiking, called bursting, starting about one week after plating, when the dissociated cells have sufficiently sent out their neurites and formed enough synaptic connections. This paper is the third in a series of three on simulation models of cultured networks. Our two previous studies [26], [27] have shown that random recurrent network activity models generate intra- and inter-bursting patterns similar to experimental data. The networks were noise or pacemaker-driven and had Izhikevich-neuronal elements with only short-term plastic (STP) synapses (so, no long-term potentiation, LTP, or depression, LTD, was included). However, elevated pre-phases (burst leaders) and after-phases of burst main shapes, that usually arise during the development of the network, were not yet simulated in sufficient detail. This lack of detail may be due to the fact that the random models completely missed network topology .and a growth model. Therefore, the present paper adds, for the first time, a growth model to the activity model, to give the network a time dependent topology and to explain burst shapes in more detail. Again, without LTP or LTD mechanisms. The integrated growth-activity model yielded realistic bursting patterns. The automatic adjustment of various mutually interdependent network parameters is one of the major advantages of our current approach. Spatio-temporal bursting activity was validated against experiment. Depending on network size, wave reverberation mechanisms were seen along the network boundaries, which may explain the generation of phases of elevated firing before and after the main phase of the burst shape.In summary, the results show that adding topology and growth explain burst shapes in great detail and suggest that young networks still lack/do not need LTP or LTD mechanisms.     

Signaling Pathways Involved in Striatal Synaptic Plasticity are Sensitive to Temporal Pattern and Exhibit Spatial Specificity

The basal ganglia is a brain region critically involved in reinforcement learning and motor control. Synaptic plasticity in the striatum of the basal ganglia is a cellular mechanism implicated in learning and neuronal information processing. Therefore, understanding how different spatio-temporal patterns of synaptic input select for different types of plasticity is key to understanding learning mechanisms. In striatal medium spiny projection neurons (MSPN), both long term potentiation (LTP) and long term depression (LTD) require an elevation in intracellular calcium concentration; however, it is unknown how the post-synaptic neuron discriminates between different patterns of calcium influx. Using computer modeling, we investigate the hypothesis that temporal pattern of stimulation can select for either endocannabinoid production (for LTD) or protein kinase C (PKC) activation (for LTP) in striatal MSPNs. We implement a stochastic model of the post-synaptic signaling pathways in a dendrite with one or more diffusionally coupled spines. The model is validated by comparison to experiments measuring endocannabinoid-dependent depolarization induced suppression of inhibition. Using the validated model, simulations demonstrate that theta burst stimulation, which produces LTP, increases the activation of PKC as compared to 20 Hz stimulation, which produces LTD. The model prediction that PKC activation is required for theta burst LTP is confirmed experimentally. Using the ratio of PKC to endocannabinoid production as an index of plasticity direction, model simulations demonstrate that LTP exhibits spine level spatial specificity, whereas LTD is more diffuse. These results suggest that spatio-temporal control of striatal information processing employs these Gq coupled pathways.     

Aquaporin-4 water channels and synaptic plasticity in the hippocampus

Aquaporin-4 (AQP4) is the major water channel expressed in the central nervous system (CNS) and is primarily expressed in glial cells. Many studies have shown that AQP4 regulates the response of the CNS to insults or injury, but far less is known about the potential for AQP4 to influence synaptic plasticity or behavior. Recent studies have examined long-term potentiation (LTP), long-term depression (LTD), and behavior in AQP4 knockout (KO) and wild-type mice to gain more insight into its potential role. The results showed a selective effect of AQP4 deletion on LTP of the Schaffer collateral pathway in hippocampus using an LTP induction protocol that simulates pyramidal cell firing during theta oscillations (theta-burst stimulation; TBS). However, LTP produced by a different induction protocol was unaffected. There was also a defect in LTD after low frequency stimulation (LFS) in AQP4 KO mice. Interestingly, some slices from AQP4 KO mice exhibited LTD after TBS instead of LTP, or LTP following LFS instead of LTD. These data suggest that AQP4 and astrocytes influence the polarity of long-term synaptic plasticity (potentiation or depression). These potentially powerful roles expand the influence of AQP4 and astrocytes beyond the original suggestions related to regulation of extracellular potassium and water balance. Remarkably, AQP4 KO mice did not show deficits in basal transmission, suggesting specificity for long-term synaptic plasticity. The mechanism appears to be related to neurotrophins and specifically brain-derived neurotrophic factor (BDNF) because pharmacological blockade of neurotrophin trk receptors or scavenging ligands such as BDNF restored plasticity. The in vitro studies predicted effects in vivo of AQP4 deletion because AQP4 KO mice performed worse using a task that requires memory for the location of objects (object placement). However, performance on other hippocampal-dependent tasks was spared. The results suggest an unanticipated and selective role of AQP4 in synaptic plasticity and spatial memory, and underscore the growing appreciation of the role of glial cells in functions typically attributed to neurons. Implications for epilepsy are discussed because of the previous evidence that AQP4 influences seizures, and the role of synaptic plasticity in epileptogenesis.