Skip to main content

Temporal filters in response to presynaptic spike trains: interplay of cellular, synaptic and short-term plasticity time scales

Abstract

Temporal filters, the ability of postsynaptic neurons to preferentially select certain presynaptic input patterns over others, have been shown to be associated with the notion of information filtering and coding of sensory inputs. Short-term plasticity (depression and facilitation; STP) has been proposed to be an important player in the generation of temporal filters. We carry out a systematic modeling, analysis and computational study to understand how characteristic postsynaptic (low-, high- and band-pass) temporal filters are generated in response to periodic presynaptic spike trains in the presence STP. We investigate how the dynamic properties of these filters depend on the interplay of a hierarchy of processes, including the arrival of the presynaptic spikes, the activation of STP, its effect on the excitatory synaptic connection efficacy, and the response of the postsynaptic cell. These mechanisms involve the interplay of a collection of time scales that operate at the single-event level (roughly, during each presynaptic interspike-interval) and control the long-term development of the temporal filters over multiple presynaptic events. These time scales are generated at the levels of the presynaptic cell (captured by the presynaptic interspike-intervals), short-term depression and facilitation, synaptic dynamics and the post-synaptic cellular currents. We develop mathematical tools to link the single-event time scales with the time scales governing the long-term dynamics of the resulting temporal filters for a relatively simple model where depression and facilitation interact at the level of the synaptic efficacy change. We extend our results and tools to account for more complex models. These include multiple STP time scales and non-periodic presynaptic inputs. The results and ideas we develop have implications for the understanding of the generation of temporal filters in complex networks for which the simple feedforward network we investigate here is a building block.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14

References

  • Abbott, L., & Regehr, W. G. (2004). Synaptic computation. Nature, 431, 796–803.

    CAS  PubMed  Article  Google Scholar 

  • Abbott, L. F., Varela, J. A., Sen, K., & Nelson, S. B. (1997). Synaptic depression and cortical gain control. Science, 275, 220–224.

    CAS  PubMed  Article  Google Scholar 

  • Abernet, L., Jadhav, S. P., Feldman, D. E., Carandini M., & Scanziani, M. (2005). Somatosensory integration controlled by thalamocortical feed-forward inhibition. Neuron, 48:315–327.

  • Amari, S.-I. (1977). Dynamics of pattern formation in lateral-inhibition type neural fields. Biological Cybernetics, 27, 77–87.

    CAS  PubMed  Article  Google Scholar 

  • Barak, O., & Tsodyks, M. (2007). Persistent activity in neural networks with dynamic synapses. PLoS Computational Biology, 3, e35.

  • Barak, O., Tsodyks, M., & Romo, R. (2008). Neuronal population coding of parametric working memory. Journal of Neuroscience, 319, 1543–1546.

    Google Scholar 

  • Bernstein, J. G., & Boyden, E. S. (2012). Optogenetic tools for analyzing the neural circuits of behavior. Current Opinion in Neurobiology, 22, 61–71.

    CAS  PubMed  Article  Google Scholar 

  • Bliss, T. V., & Gardner-Medwin, A. R. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the unanaestetized rabbit following stimulation of the perforant path. Journal of Physiology, 232, 357–374.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Bourjaily, M. A., & Miller, P. (2012). Dynamic afferent synapses to decision-making networks improve performance in tasks requiring stimulus associations and discriminations. Journal of Neurophysiology, 108, 513–527.

    PubMed  PubMed Central  Article  Google Scholar 

  • Buonomano, D., & Maass, W. (2009). State-dependent computations: spatiotemporal processing in cortical networks. Nature Reviews Neuroscience, 10, 113–125.

    CAS  PubMed  Article  Google Scholar 

  • Buonomano, D. V. (2000). Decoding temporal information: A model based on short-term synaptic plasticity. Journal of Neuroscience, 20, 1129–1141.

    CAS  PubMed  Article  Google Scholar 

  • Burden, R. L., & Faires, J. D. (1980) Numerical analysis. PWS Publishing Company - Boston.

  • Carver, S., Roth, E., Cowan, N. J., & Fortune, E. S. (2008). Synaptic plasticity can produce and enhance direction selectivity. PLoS Computational Biology, 4, e32.

  • Chance, F. S., Nelson, S. B., & Abbott, L. F. (1998). Synaptic depression and the temporal response characteristics of V1 cells. Journal of Neuroscience, 18, 4785–4799.

    CAS  PubMed  Article  Google Scholar 

  • Clements, J. D., Lester, R. A., Tong, G., Jahr, C. E., & Westbrook, G. L. (1992). The time course of glutamate in the synaptic cleft. Science, 258, 1498–1501.

    CAS  PubMed  Article  Google Scholar 

  • Colquhoun, D., Jonas, P., & Sakmann, B. (1992). Action of brief pulses of glutamate on AMPA/kainate receptors in patches from different neurones of rat hippocampal slices. Journal of Physiology, 458, 261–287.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Cook, D. L., Schwindt, P. C., Grande, L. A., & Spain, W. J. (2003). Synaptic depression in the localization of sound. Nature, 421, 66–70.

    CAS  PubMed  Article  Google Scholar 

  • Dayan, P., & Abbott, L. F. (2001). Theoretical Neuroscience. Cambridge, Massachusetts: The MIT Press.

    Google Scholar 

  • De Pitta, M., Volman, V., Berry, H., Parpura, V., Volterra, A., & Ben-Jacob, E. (2012). Computational quest for understanding the role of astrocyte signaling in synaptic transmission and plasticity. Frontiers in Computational Neuroscience, 6, 98.

    PubMed  PubMed Central  Article  Google Scholar 

  • De Pitta, M., Volman, V., Berry, H., & Ben-Jacob, E. (2011). A tale of two stories: Astrocyte regulation of synaptic depression and facilitation. PLoS Computational Biology, 7, e1002293.

  • Deco, G., Rolls, E., & Romo, R. (2010). Synaptic dynamics and decision making. Proceedings of the National academy of Sciences of the United States of America, 107, 7547–7549.

    Google Scholar 

  • Deng, P.-Y., & Klyachko, A. (2011). The diverse functions of short-term plasticity components in synaptic computations. Communicative & Integrative Biology, 4, 543–548.

    Article  Google Scholar 

  • Destexhe, A., Mainen, Z. F., & Sejnowski, T. (1994). An efficient method for computing synaptic conductances based on a kinetic model of receptor binding. Neural Computation, 6, 14–18.

    Article  Google Scholar 

  • Destexhe, A., Mainen, Z. F., & Sejnowski, T. (1998). Kinetic models of synaptic transmission. In Methods in Neural Modeling. Koch, C. and Segev, I., editors, second edition. MIT Press: Cambridge, Massachusetts, pages 1–25.

  • Destexhe, A., & Marder, E. (2004). Plasticity in single neuron and circuit computations. Nature, 431, 785–795.

    Article  CAS  Google Scholar 

  • Dittman, J. S., Kreitzer, A. C., & Regehr, W. G. (2000). Interplay between facilitation, depression, and residual calcium at three presynaptic terminals. Journal of Neuroscience, 20, 1374–1385.

    CAS  PubMed  Article  Google Scholar 

  • Drover, J. D., Tohidi, V., Bose, A., & Nadim, F. (2007). Combining synaptic and cellular resonance in a feedforward neuronal network. Neurocomputing, 70, 2041–2045.

    PubMed  PubMed Central  Article  Google Scholar 

  • Dudel, J., & Kuffler, S. W. (1961). Presynaptic inhibition at the crayfish neuromuscular junction. Journal of Physiology, 155, 543–562.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Ermentrout, G. B., & Terman, D. (2010). Mathematical Foundations of Neuroscience. Springer.

    Book  Google Scholar 

  • Fioravante, D., & Regehr, W. G. (2011). Short-term forms of presynaptic plasticity. Current Opinion in Neurobiology, 21, 260–274.

    Article  CAS  Google Scholar 

  • Fortune, E., & Rose, G. (1997a). Passive and active membrane properties contribute to the temporal filtering properties of midbrain neurons in vivo. Journal of Neuroscience, 17, 3815–3825.

    CAS  PubMed  Article  Google Scholar 

  • Fortune, E., & Rose, G. (1997b). Temporal filtering properties of ampullary electrosensory neurons in the torus semicircularis of eigenmannia: evolutionary and computational implications. Brain, Behavior and Evolution, 49, 312–323.

    CAS  PubMed  Article  Google Scholar 

  • Fortune, E., & Rose, G. (2000). Short-term synaptic plasticity contributes to the temporal filtering of electrosensory information. Journal of Neuroscience, 20, 7122–7130.

    CAS  PubMed  Article  Google Scholar 

  • Fortune, E., & Rose, G. (2001). Short-term plasticity as a temporal filter. Trends in Neurosciences, 24, 381–385.

    CAS  PubMed  Article  Google Scholar 

  • Fortune, E., & Rose, G. (2002). Roles of short-term plasticity in behavior. Journal of Physiology - Paris, 96, 539–545.

    PubMed  Article  Google Scholar 

  • Fuhrmann, G., Segev, I., & Markram, H. (2004). Coding of temporal information by activity-dependent synapses. Journal of Physiology, 556, 19–27.

    Article  CAS  Google Scholar 

  • George, A. A., Lyons-Warren, A. M., Ma, X., & Carlson, B. A. (2011). A diversity of synaptic filters are created by temporal summation of excitation and inhibition. Journal of Neuroscience, 31, 14721–14734.

    CAS  PubMed  Article  Google Scholar 

  • Gerstner, W., Kistler, W. M., Naud, R., & Paninski, L. (2014). Neuronal dynamics: From single neurons to networks and models of cognition. Cambridge University Press.

    Book  Google Scholar 

  • Goldman, M. S., Maldonado, P., & Abbott, L. F. (2002). Redundancy reduction and sustained firing with stochastic depressing synapses. Journal of Neuroscience, 22, 584–591.

    CAS  PubMed  Article  Google Scholar 

  • Gupta, A., Wang, Y., & Markram, H. (2000). Organizing principles for a diversity of GABAergic interneurons and synapses in the neocortex. Science, 287, 273–278.

    CAS  PubMed  Article  Google Scholar 

  • Hebb, D. O. (1949). The Organization of Behavior: A neuropsychological theory. New York: Wiley.

    Google Scholar 

  • Hennig, M. H. (2013). Theoretical models of synaptic short term plasticity. Frontiers in Computational Neuroscience, 7, 45.

    PubMed  PubMed Central  Article  Google Scholar 

  • Hennig, M. H., Postlethwaite, M., Forsythe, I. D., & Graham, B. P. (2008). Interactions between multiple sources of short-term plasticity during evoked and spontaneous activity at the rat calyx of held. Journal of Physiology, 586, 3129–3146.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Holcman, D., & Tsodyks, M. (2006). The emergence of up and down states in cortical networks. PLoS Computational Biology, 2(3), e23.

  • Hu, H., Vervaeke, K., & Storm, J. F. (2002). Two forms of electrical resonance at theta frequencies generated by M-current, h-current and persistent Na+ current in rat hippocampal pyramidal cells. Journal of Physiology, 545(3), 783–805.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Hutcheon, B., Miura, R. M., & Puil, E. (1996). Subthreshold membrane resonance in neocortical neurons. Journal of Neurophysiology, 76, 683–697.

    CAS  PubMed  Article  Google Scholar 

  • Hutcheon, B., & Yarom, Y. (2000). Resonance, oscillations and the intrinsic frequency preferences in neurons. Trends in Neurosciences, 23, 216–222.

    CAS  PubMed  Article  Google Scholar 

  • Izhikevich, E. M., Desai, N. S., Walcott, E. C., & Hoppensteadt, F. C. (2003). Bursts as a unit of neural information: Selective communication via resonance. Trends in Neurosciences, 26, 161–167.

    CAS  PubMed  Article  Google Scholar 

  • Kandaswamy, U., Deng, P.-Y., Stevens, C., & Klyachko, V. A. (2010). The role of presynaptic dynamics in processing of natural spike trains in hippocampal synapses. Journal of Neuroscience, 30, 15904–15914.

    CAS  PubMed  Article  Google Scholar 

  • Karmarkar, U. R., & Buonomano, D. V. (2007). Timing in the absence of clocks: Encoding time in neural network states. Neuron, 53, 427–438.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Klyachko, V. A., & Stevens, C. F. (2006). Excitatory and feed-forward inhibitory hippocampal synapses work synergistically as an adaptive filter of natural spike trains. PLoS Computational Biology, 4, e207.

  • Latorre, R., Torres, J. J., & Varona, P. (2016). Interplay between subthreshold oscillations and depressing synapses in single neurons. PLoS ONE, 11, e0145830.

  • Lewis, J. E., & Maler, L. (2002). Dynamics of electrosensory feedback: short-term plasticity and inhibition in a parallel fiber pathway. Journal of Neurophysiology, 88, 1695–1702.

    PubMed  Article  Google Scholar 

  • Lindner, B., Gangloff, D., Longtin, A., & Lewis, J. E. (2009). Broadband coding with dynamic synapses. Journal of Neuroscience, 29, 2076–2088.

    CAS  PubMed  Article  Google Scholar 

  • Lisman, J. E. (1997). Bursts as a unit of neural information: making unreliable synapses reliable. Trends in Neurosciences, 20, 38–43.

    CAS  PubMed  Article  Google Scholar 

  • Loebel, A., & Tsodyks, M. (2002). Computation by ensemble synchronization in recurrent networks with synaptic depression. Journal of Computational Neuroscience, 13, 111–124.

    PubMed  Article  Google Scholar 

  • Maass, A., & Zador, W. (1999). Dynamic stochastic synapses as computational units. Neural Computation, 11, 903–917.

    CAS  PubMed  Article  Google Scholar 

  • Magleby, K. L., & Zengel, J. E. (1982). A quantitative description of stimulation-induced changes in transmitter release at the frog neuromuscular junction. Journal of General Physiology, 80, 613–638.

    CAS  PubMed  Article  Google Scholar 

  • Manor, Y., & Nadim, F. (2001). Synaptic depression mediates bistability in neuronal networks with recurrent inhibitory connectivity. Journal of Neuroscience, 21, 9460–9470.

    CAS  PubMed  Article  Google Scholar 

  • Manor, Y., Nadim, F., Abbott, L., & Marder, E. (1997). Temporal dynamics of graded synaptic transmission in the lobster stomatogastric ganglion. Journal of Neuroscience, 17, 5610–5621.

    CAS  PubMed  Article  Google Scholar 

  • Marder, E. (2012). Neuromodulation of neuronal circuits : Back to the future. Neuron, 76, 1–11.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Marder, E., & Thirumalai, V. (2002). Cellular, synaptic and network effects of neuromodulation. Neural Networks, 15, 479–493.

    PubMed  Article  Google Scholar 

  • Markram, H., Gupta, A., Uziel, A., Wang, Y., & Tsodyks, M. (1998a). Information processing with frequency-dependent synaptic connections. Neurobiology of Learning and Memory, 70, 101–112.

    CAS  PubMed  Article  Google Scholar 

  • Markram, H., Pikus, D., Gupta, A., & Tsodyks, M. (1998b). Potential for multiple mechanisms, phenomena and algorithms for synaptic plasticity at single synapses. Neuropharmacology, 37, 489–500.

    CAS  PubMed  Article  Google Scholar 

  • Markram, H., & Tsodyks, M. (1996). Redistribution of synaptic efficacy between neocortical pyramidal neurons. Nature, 382, 807–810.

    CAS  PubMed  Article  Google Scholar 

  • Markram, H., Wang, Y., & Tsodyks, M. (1998c). Differential signaling via the same axon of neocortical pyramidal neurons. Proceedings of the National academy of Sciences of the United States of America, 95, 5323–5328.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Martin, S. J., Grimwood, P. D., & Morris, R. G. (2000). Synaptic plasticity and memory: an evaluation of the hypothesis. Annual Review of Neuroscience, 23, 649–711.

    CAS  PubMed  Article  Google Scholar 

  • Mejias, J. F., & Torres, J. J. (2008). The role of synaptic facilitation in spike coincidence detection. Journal of Computational Neuroscience, 24, 222–234.

    PubMed  Article  Google Scholar 

  • Mejias, J. F., & Torres, J. J. (2009). Maximum memory capacity on neural networks with short-term synaptic depression and facilitation. Neural Computation, 21, 851–871.

    PubMed  Article  Google Scholar 

  • Mongillo, G., Barak, O., & Tsodyks, M. (2015). Synaptic theory of working memory. Science, 319, 1543–1546.

    Article  CAS  Google Scholar 

  • Morrison, A., Diesmann, M., & Gerstner, W. (2008). Phenomenological models of synaptic plasticity based on spike timing. Biological Cybernetics, 98, 459–478.

    PubMed  PubMed Central  Article  Google Scholar 

  • O’Donnell, C., & Nolan, M. F. (2010). Tuning of synaptic responses: an organizing principle for optimization of neural circuits. Trends in Neurosciences, 34, 51–60.

    PubMed  Article  CAS  Google Scholar 

  • Pouille, F., & Scanziani, M. (2004). Routing of spike series by dynamic circuits in the hippocampus. Nature, 429, 717–723.

    CAS  PubMed  Article  Google Scholar 

  • Richardson, M. J. E., Brunel, N., & Hakim, V. (2003). From subthreshold to firing-rate resonance. Journal of Neurophysiology, 89, 2538–2554.

    PubMed  Article  Google Scholar 

  • Rosenbaum, R., Rubin, J., & Doiron, B. (2012). Short term synaptic depression imposes a frequency dependent filter on synaptic information transfer. PLoS Compuatational Biology, 8, e1002557.

  • Rotman, Z., Deng, P.-Y., & Klyachko, V. A. (2011). Short-term plasticity optimizes synaptic information transmission. Journal of Neuroscience, 31, 14800–14809.

    CAS  PubMed  Article  Google Scholar 

  • Rotstein, H. G. (2015). Subthreshold amplitude and phase resonance in models of quadratic type: nonlinear effects generated by the interplay of resonant and amplifying currents. Journal of Computational Neuroscience, 38, 325–354.

    PubMed  Article  Google Scholar 

  • Rotstein, H. G., & Nadim, F. (2014). Frequency preference in two-dimensional neural models: a linear analysis of the interaction between resonant and amplifying currents. Journal of Computational Neuroscience, 37, 9–28.

    PubMed  Article  Google Scholar 

  • Rotstein, H. G., & Tabak, E. G. (2019). Analysis of spike-driven processes through attributable components. Communications in Mathematical Sciences, 17, 1177–1192.

  • Silberberg, G., Wu, C., & Markram, H. (2004). Synaptic dynamics control the timing of neuronal excitation in the activated neocortical microcircuit. Journal of Physiology, 556, 19–27.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Stevens, C., & Wang, Y. (1995). Facilitation and depression at single central synapses. Neuron, 14, 795–802.

    CAS  PubMed  Article  Google Scholar 

  • Suh, B., & Baccus, S. A. (2014). Building blocks of temporal filters in retinal synapses. PLoS Biology, 12, e1001973.

  • Tauffer, L., & Kumar, A. (2021). Short-term synaptic plasticity makes neurons sensitive to the distribution of presynaptic population firing rates. Eneuro, 8(2).

  • Thomson, A. (2003). Presynaptic frequency- and pattern-dependent filtering. Journal of Computational Neuroscience, 15, 159–202.

    PubMed  Article  Google Scholar 

  • Tsodyks, M., & Markram, H. (1996). Plasticity of neocortical synapses enables transitions between rate and temporal coding. Lecture Notes in Computer Science, 1112, 445–450.

    Article  Google Scholar 

  • Tsodyks, M., & Markram, H. (1997). The neural code between neocortical pyramidal neurons depends on neurotransmitter release probability. Proceedings of the National academy of Sciences of the United States of America, 94, 719–723.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Tsodyks, M., Pawelzik, K., & Markram, H. (1998). Neural networks with dynamic synapses. Neural Computation, 10, 821–835.

    CAS  PubMed  Article  Google Scholar 

  • Tsodyks, M., Uziel, A., & Markram, H. (2000). Synchrony generation in recurrent networks with frequency-dependent synapses. Journal of Neuroscience, 20, 1–5.

    Article  Google Scholar 

  • Tsodyks, M., & Wu, S. (2013). Short-term synaptic plasticity. Scholarpedia, 8, 3153.

    Article  Google Scholar 

  • Turrigiano, G. G., & Nelson, S. B. (2004). Homeostatic plasticity in the developing nervous system. Nature Reviews Neuroscience, 5, 97–107.

    CAS  PubMed  Article  Google Scholar 

  • Varela, J. A., Sen, K., Gibson, J., Fost, J., Abbott, L. F., & Nelson, S. B. (1997). A quantitative description of short-term plasticity at excitatory synapses in layer 2/3 of rat primary visual cortex. Journal of Neuroscience, 17, 7926–7940.

    CAS  PubMed  Article  Google Scholar 

  • Varela, J. A., Song, S., Turrigiano, G. G., & Nelson, S. B. (1999). Differential depression at excitatory and inhibitory synapses in visual cortex. Journal of Neuroscience, 19, 4293–4304.

    CAS  PubMed  Article  Google Scholar 

  • Yuan, W., Dimigen, O., Sommer, W., & Zhou, C. (2013). A model of microsaccade-related neural responses induced by short-term depression in thalamocortical synapses. Frontiers in Computational Neuroscience, 7, 47.

    PubMed  PubMed Central  Article  Google Scholar 

  • Zador, A., & Dobrunz, L. (1997). Dynamic synapses in the cortex. Neuron, 19, 1–4.

    CAS  PubMed  Article  Google Scholar 

  • Zhang, F., Gradinaru, V., Adamantidis, A. R., Durand, R., Airan, R. D., de Lecea, L., & Deisseroth, K. (2010). Optogenetic interrogation of neural circuits: Technology for probing mammalian brain structures. Nature Protocols, 5, 439–456.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  • Zucker, R. S. (1989). Short-term synaptic plasticity. Annual Review of Neuroscience, 12, 13–31.

    CAS  PubMed  Article  Google Scholar 

  • Zucker, R. S., & Regehr, W. G. (2002). Short-term synaptic plasticity. Annual Review of Physiology, 64, 355–405.

    CAS  PubMed  Article  Google Scholar 

Download references

Acknowledgements

This work was partially supported by the National Science Foundation grant DMS-1608077 (HGR) and an NSF Graduate Research Fellowship (YM). The authors are grateful to Allen Tannenbaum for useful comments and support, and to Farzan Nadim, Dirk Bucher and Nelly Daur for useful discussions.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Horacio G. Rotstein.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Action Editor: J. Rinzel

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 1010 KB)

Appendices

A. 1D linear difference equations

A.1 Constant coefficients

Consider the following linear difference equation

$$\begin{aligned} w_{n+1} = \alpha \, w_n + \beta , n=1, 2, \ldots \end{aligned}$$
(73)

where \(\alpha\) and \(\beta\) are constants. The steady-state for this equation, if it exists, is given by

$$\begin{aligned} \bar{w} = \frac{\beta }{1-\alpha }. \end{aligned}$$
(74)

By solving (73) recurrently and using

$$\begin{aligned} \sum _{n=0}^{N} a^n = \frac{a^{N+1}-1}{a-1} \end{aligned}$$
(75)

where \(a \ne 1\) is a real number, one gets

$$\begin{aligned} w_n = \alpha ^{n-1} w_1 + \beta \, \frac{\alpha ^{n-1}-1}{\alpha -1} . \end{aligned}$$
(76)

Substitution of (74) into this equation yields

$$\begin{aligned} w_n = \bar{w} + \alpha ^{n-1}\, (w_1 - \bar{w}). \end{aligned}$$
(77)

Application of formula (77) to the difference Eqs. (7) and (8) gives, respectively,

$$\begin{aligned} X_n = \bar{X} + (1 - \bar{X})\, (1-a_d)^{n-1} e^{-(n-1) \Delta _{spk} / \tau _{dep}} = \end{aligned}$$
$$\begin{aligned} \bar{X} + (1 - \bar{X})\, e^{-(n-1) [\, \Delta _{spk} / \tau _{dep} - \ln (1-a_d) ]} \end{aligned}$$
(78)

and

$$\begin{aligned} Z_n = \bar{Z} + (1 - \bar{Z})\, (1-a_f)^{n-1} e^{-(n-1) \Delta _{spk} / \tau _{fac}} = \end{aligned}$$
$$\begin{aligned} \bar{Z} + (1 - \bar{Z})\, e^{-(n-1) [\, \Delta _{spk} / \tau _{fac} - \ln (1-a_f) ]}. \end{aligned}$$
(79)

A.2 Variable (n-dependent) coefficients

Consider the following linear difference equation

$$\begin{aligned} w_{n+1} = \alpha _n\, w_n + \beta _n, n=1, 2, \ldots \end{aligned}$$
(80)

By solving (73) recurrently one gets

$$\begin{aligned} w_n = \left( \prod _{k=1}^{n-1} \alpha _k \right) \, x_1 + \sum _{k=1}^{n-1} \left( \prod _{j=k+1}^{n-1} \alpha _j \right) \, \beta _k \end{aligned}$$
(81)

where we are using the convention \(\prod _{j_1}^{j_2} = 1\) if \(j_1 > j_2\). Equation (81) reduces to Eq. (79) if both coefficients in (81) are constant.

Consider now Eq. (80) where the coefficients are expressed as small perturbations \(\delta _{\alpha ,n} \ll 1\) and \(\delta _{\beta ,n} \ll 1\) (\(n = 1, 2, \ldots\)), respectively, of constant coefficients

$$\begin{aligned} \alpha _n = \alpha + \delta _{\alpha ,n} \text { and } \beta _n = \beta + \delta _{\beta ,n}. \end{aligned}$$
(82)

To the first order approximation, the solution (81) reads

$$\begin{aligned} w_n = &\alpha ^{n-1} w_1 + \beta \, \frac{\alpha ^{n-1}-1}{\alpha -1} + \alpha ^{n-2} w_1\, \sum _{k=1}^{n-1} \delta _{\alpha ,k}\\&+ \beta \, \sum _{k=1}^{n-1} \alpha ^{n-k-2} \sum _{j=k+1}^{n-1} \delta _{\alpha ,j} + \sum _{k=1}^{n-1} \alpha ^{n-k-1}\, \delta _{\beta ,k} = . \end{aligned}$$
$$\begin{aligned} = &\bar{w} + \alpha ^{n-1}\, (w_1 - \bar{w}) + \alpha ^{n-2} w_1\, \sum _{k=1}^{n-1} \delta _{\alpha ,k} \\&+ \beta \, \sum _{k=1}^{n-1} \alpha ^{n-k-2} \sum _{j=k+1}^{n-1} \delta _{\alpha ,j} + \sum _{k=1}^{n-1} \alpha ^{n-k-1}\, \delta _{\beta ,k}. \end{aligned}$$
(83)

B. Some properties of \(\bar{X}\) and \(\bar{Z}\) and their dependence with \(\Delta _{spk}\) and \(\tau _{dep/fac}\)

Consider \(\bar{X}\) and \(\bar{Z}\) given by (9) and (10), respectively.

B.1 Monotonic dependence of \(\bar{X}\) and \(\bar{Z}\) with \(\Delta _{spk}\)

If \(a_d > 0\) and \(x_{\infty } > 0\), then \(\bar{X}\) is an increasing function of \(\Delta _{spk}\) and a decreasing function of \(f_{spk}\). This results from

$$\begin{aligned} \frac{\partial \bar{X}}{\partial \Delta _{spk}} = \frac{x_{\infty }\, a_d\, e^{-\Delta _{spk}/\tau _{dep}}}{\tau _{dep}\, [1 - (1-a_d) e^{-\Delta _{spk}/\tau _{dep}}]^2} > 0. \end{aligned}$$
(84)

If \(a_f < 1\) and \(z_{\infty } < 1\), then \(\bar{Z}\) is a decreasing function of \(\Delta _{spk}\) and an increasing function of \(f_{spk}\). This results from

$$\begin{aligned} \frac{\partial \bar{Z}}{\partial \Delta _{spk}} = \frac{a_f\, (1-a_f)\, (z_{\infty }-1)}{\tau _{fac}\, [1 - (1-a_f) e^{-\Delta _{spk}/\tau _{fac}}]^2} < 0. \end{aligned}$$
(85)

B.2 Monotonic dependence of \(\bar{X}\) and \(\bar{Z}\) with \(\tau _{dep/fac}\)

If \(a_d > 0\) and \(x_{\infty } > 0\), then \(\bar{X}\) is a decreasing function of \(\tau _{dep}\). This results from

$$\begin{aligned} \frac{\partial \bar{X}}{\partial \Delta _{spk}} = -\frac{\Delta _{spk}\ x_{\infty }\, a_d\, e^{-\Delta _{spk}/\tau _{dep}}}{\tau _{dep}^2\, [1 - (1-a_d) e^{-\Delta _{spk}/\tau _{dep}}]^2} < 0. \end{aligned}$$
(86)

If \(a_f < 1\) and \(z_{\infty } < 1\), then \(\bar{Z}\) is a decreasing function of \(\tau _{fac}\). This results from

$$\begin{aligned} \frac{\partial \bar{Z}}{\partial \Delta _{spk}} = -\frac{\Delta _{spk}\, a_f\, (1-a_f)\, (z_{\infty }-1)}{\tau _{fac}^2\, [1 - (1-a_f) e^{-\Delta _{spk}/\tau _{fac}}]^2} > 0. \end{aligned}$$
(87)

C. Models of synaptic depression and facilitation

C.1 Depression - facilitation model used in Latorre et al. (2016)

Following Destexhe et al. (19941998), the synaptic variables S obey a kinetic equation of the form

$$\begin{aligned} \frac{dS}{dt} = N(V)\, \frac{ (1 - S)}{\tau _{r}} - \frac{S}{\tau _{d}}, \end{aligned}$$
(88)

where N(V) (mM) representes the neurotransmitter concentration in the synaptic cleft. Neurotransmitters are assumed to be released quickly upon the arrival of a presynaptic spike and remain in the synaptic cleft for the duration of the spike (\(\sim 1\) ms). This can be modeled by either using a sigmoid function

$$\begin{aligned} N(V) = \frac{1 + \tanh (V/4)}{2}, \end{aligned}$$
(89)

or a step function if the release is assumed to be instantaneous. The parameters \(\tau _r\) and \(\tau _d\) are the rise and decay time constants respectively (msec).

This model assumes N(V) is independent of the spiking history (the value of N(V) during a spike is constant, except possibly for the dependence on V). (There is evidence that this is not realistic Markram and Tsodyks (1996), Dudel and Kuffler (1961).) In Latorre et al. (2016), the “activated” time was 1 ms Clements et al. (1992), Colquhoun et al. (1992).

In Latorre et al. (2016), they followed the description of the synaptic short-term dynamics following Tsodyks and Markram (1997), Tsodyks et al. (1998) (Sect. 2.1.5). For the dynamics of the synaptic function S, they used a function \([T] = \kappa \, \Delta S_n\) during the release time and \([T] = 0\) otherwise, instead of N(V). The combination of the two formulations yields

$$\begin{aligned} \frac{dS}{dt} = \kappa \, \Delta S_n\, N(V)\, \frac{ (1 - S)}{\tau _{r}} - \frac{S}{\tau _{d}}. \end{aligned}$$
(90)

In the following alternative formulation Drover et al. (2007) \(\kappa \, \Delta S_n\) does not affect the effective rise time of the synaptic function S

$$\begin{aligned} \frac{dS}{dt} = N(V)\, \frac{ (\kappa \Delta S_n - S)}{\tau _{r}} - \frac{S}{\tau _{d}}. \end{aligned}$$
(91)

C.2 Depression model used in Manor and Nadim (2001)

Following experimental procedures described in Manor et al. (1997), the synaptic current is described by \(I_{syn}=G_{ex} a\, d (V-E_{ex})\) where a and d are variables that represent activation and depression processes, respectively. They follow the form:

$$\begin{aligned} \frac{dy}{dt} = \frac{y_{\infty }(V_\mathrm{pre}-y)}{\tau _y}, \end{aligned}$$
(92)

where \(y=a,d\). The steady-state of y is given by

$$\begin{aligned} y_{\infty } = \frac{1}{1+\exp ((V-V_x)/k)}, \end{aligned}$$
(93)

and its time constant follows

$$\begin{aligned} \tau _y = \tau _1 + \frac{\tau _h - \tau _1}{1 + \exp ((V-V_x)/k)}. \end{aligned}$$
(94)

This model is used in Manor and Nadim (2001) to describe bistability in pacemaker networks with recurrent inhibition and depressing synapses. Parameters in these equations are experimentally fitted from the pyloric network of the crab Cancer borealis.

D. Additional model formulations for multiple depression-facilitation processes

In Sect. 3.7 we discussed the model formulation (47) and (48) describing the interplay of two depression-facilitations processes. A number of additional, simplified formulations are possible based on different assumptions. The models we propose here are natural mathematical extensions of the single depression/facilitation processes discussed in the main body of this paper. They are phenomenological models, not based on any experimental observation or theoretical foundation, and they are limited in their general applicability. However, they are useful to explore the possible scenarios underlying the interplay of multiple depression and facilitation time scales affecting the PSP dynamics of a cell in response to presynaptic input trains.

D.1 Additive and multiplicative segregated-processes models

In the additive and multiplicative segregated models, the variable M is given, respectively, by

$$\begin{aligned} M^+(t) = (1 - \alpha )\, x_1(t) z_1(t) + \alpha \, x_2(t) z_2(t) \end{aligned}$$
(95)

and

$$\begin{aligned} M^*(t) = [x_1(t) z_1(t)]^{1 - \alpha }\, [x_2(t) z_2(t)]^\alpha \end{aligned}$$
(96)

where the parameter \(\alpha \in [0,1]\) controls the relative contribution of each of the processes. Correspondingly, the updates are given by

$$\begin{aligned} \Delta S_n^+ = (1 - \alpha )\, X_{1,n} Z_{1,n} + \alpha \, X_{2,n} Z_{2,n} \end{aligned}$$
(97)

and

$$\begin{aligned} \Delta S_n^* = [X_{1,n} Z_{1,n}]^{1 - \alpha }\, [X_{2,n} Z_{2,n}]^\alpha . \end{aligned}$$
(98)

For \(\alpha = 0\), \(\Delta S_n^+\) and \(\Delta S_n^*\) reduce to \(\Delta S_{1,n}\) (single depression-facilitation process). This accounts for the regimes where \(\tau _{dep,2}, \tau _{fac,2} \ll 1\). If the two processes are equal (\(\tau _{dep,1} = \tau _{dep,2}\) and \(\tau _{fac,1} = \tau _{fac,2}\)), then \(\Delta S_n^+\) and \(\Delta S_n^*\) also reduce to \(\Delta S_{1,n}\). However, these models fail to account for the reducibility in the situations where only \(\tau _{dep,2} \ll 1\) or \(\tau _{fac,2} \ll 1\), but not both. The option of considering depression to be described by \(x_1\) and facilitation by \(z_ 2\) (with \(\tau _{fac,1}, \tau _{dep,2} \ll 1\)) is technically possible in the context of the model, but it wouldn’t be consistent with the model description of single depression-facilitation processes, and it will make no sense to use the model in this way. In general, this model would be useful when the depression and facilitation time scales for each process 1 and 2 are comparable and the differences in these time scales across depression/facilitation processes should be large enough.

D.2 Fully multiplicative model

One natural way to extend the variable M to more than one process is by considering

$$\begin{aligned} M^\#(t) = x_1(t) z_1(t) x_2(t) z_2(t) \end{aligned}$$
(99)

and the synaptic update, given by

$$\begin{aligned} \Delta S_n^\# = X_{1,n} Z_{1,n} X_{2,n} Z_{2,n}. \end{aligned}$$
(100)

This formulation presents us with a number of consistency problems related to the reducibility (or lack of thereoff) to a single depression-facilitation process in some limiting cases when, for example, the two depression or facilitation time constants are very similar and therefore the associated processes are almost identical, or the depression or facilitation time constants are very small and therefore the envelopes of the associated processes are almost constant across cycles.

More specifically, first, if \(\tau _{dep,2}, \tau _{fac,2} \ll 1\) (almost no STD), then \(X_{2,n} Z_{2,n} \sim \bar{X}_2 \bar{Z}_2 = a_f\) for all n after a very short transient and therefore \(\Delta S_n^\# = X_1 Z_1 a_f \ne \Delta S_{1,n}\). One way, perhaps the simplest, to address this is to divide the expressions (99 ) and (100 ) by \(a_f^2\) and redefine \(\Delta S_{k,n}\) for the single depression-facilitation process accordingly. Specifically,

$$\begin{aligned} \Delta S_n^\# = \frac{X_{1,n} Z_{1,n} X_{2,n} Z_{2,n}}{a_f^2} = \frac{X_{1,n} Z_{1,n}}{a_f}\, \frac{X_{2,n} Z_{2,n}}{a_f} , \end{aligned}$$
(101)

where we use the notation

$$\begin{aligned} \Delta S_{1,n} = \frac{X_{1,n} Z_{1,n}}{a_f} \text{ and } \Delta S_{2,n} = \frac{X_{2,n} Z_{2,n}}{a_f}. \end{aligned}$$
(102)

The effect of redefining \(\Delta S_{k,n}\) by dividing the original expression (used in the previous sections) does not affect the time constants and the differences in the values between the two formulations is absorbed by the maximal synaptic conductance.

Second, if \(\tau _{dep,1} = \tau _{dep,2}\) and \(\tau _{fac,1} = \tau _{fac,2}\), then \(X_{1,n} = X_{2,n}\) and \(Z_{1,n} = Z_{2,n}\) for all n, and \(\Delta S_n^\# = \Delta S_{1,n}^2\) instead of \(\Delta S_n^\# = \Delta S_{1,n}\). In order to address this, the synaptic update can be modified to

$$\begin{aligned} \Delta S_n^\# = \left[ X_{1,n} X_{2,n} \right] ^{\lambda _{dep}} \left[ \frac{Z_{1,n} Z_{2,n}}{a_f^2} \right] ^{\lambda _{fac}} \end{aligned}$$
(103)

where

$$\begin{aligned} \lambda _{dep} = \frac{1}{H(| \tau _{dep,1}-\tau _{dep,2}|)} \text{ and } \lambda _{fac} = \frac{1}{H(| \tau _{fac,1}-\tau _{fac,2}|)} \end{aligned}$$
(104)

and \(H(\Delta \tau )\) is a rapidly decreasing function satisfying \(H(0) = 2\) and \(\lim _{\Delta \tau \rightarrow \infty } H(\Delta \tau ) = 1\). In our simulations we will use

$$\begin{aligned} H(\Delta \tau ) = 1 + e^{-\Delta \tau /\beta } \end{aligned}$$
(105)

with \(\beta > 0\). Correspondingly,

$$\begin{aligned} M^\# = \left[ x_1(t) x_2(t) \right] ^{\lambda _{dep}} \left[ \frac{z_1(t) z_2(t)}{a_f^2} \right] ^{\lambda _{fac}} \end{aligned}$$
(106)

In this way,

  • If \(\tau _{dep,1} = \tau _{dep,2}\), then \(X_{1,n} = X_{2,n}\) for all n and \(\lambda _{dep} = 1/2\). This gives

    $$\begin{aligned} \Delta S_n^\# = X_{1,n} \left[ \frac{Z_{1,n} Z_{2,n}}{a_f^2} \right] ^{\lambda _{fac}} . \end{aligned}$$

    If, in addition, \(\tau _{fac,1} \ne \tau _{fac,2}\) and \(| \tau _{fac,1}-\tau _{fac,2}| > 0\) is large enough, then \(\lambda _{fac} = 1\) and

    $$\begin{aligned} \Delta S_n^\# = X_{1,n} \frac{Z_{1,n} Z_{2,n}}{a_f^2} = \Delta S_{1,n} \frac{Z_{2,n}}{a_f}. \end{aligned}$$
  • If \(\tau _{fac,1} = \tau _{fac,2}\), then \(Z_{1,n} = Z_{2,n}\) for all n, \(\lambda _{fac} = 2\) and

    $$\begin{aligned} \Delta S_n^\# = \left[ X_{1,n} X_{2,n} \right] ^{\lambda _{dep}} \frac{Z_{1,n}}{a_f}. \end{aligned}$$

    If, in addition, \(\tau _{dep,1} \ne \tau _{dep,2}\) and \(| \tau _{dep,1}-\tau _{dep,2}| > 0\) is large enough, then \(\lambda _{dep} = 1\) and

    $$\begin{aligned} \Delta S_n^\# = X_{1,n} X_{2,n} \frac{Z_{1,n}}{a_f} = \Delta S_{1,n} X_{2,n}. \end{aligned}$$
  • It follows that if both \(\tau _{dep,1} = \tau _{dep,2}\) and \(\tau _{fac,1} = \tau _{fac,2}\), then \(X_{1,n} = X_{2,n}\) and \(Z_{1,n} = Z_{2,n}\) for all n, \(\lambda _{dep} = \lambda _{fac} = 2\) and

    $$\begin{aligned} \Delta S_n^\# = X_{1,n} \frac{Z_{1,n}}{a_f} = \Delta S_{1,n}. \end{aligned}$$
  • If \(\tau _{dep,2} \ll 1\) and \(| \tau _{dep,1}-\tau _{dep,2}|\) is large enough, then \(X_{2,n} = 1\) for all n (after a very short transient), \(\lambda _{dep} = 1\), and then

    $$\begin{aligned} \Delta S_n^\# = X_{1,n} \left[ \frac{Z_{1,n} Z_{2,n}}{a_f^2} \right] ^{\lambda _{fac}}. \end{aligned}$$

    If, in addition, \(\tau _{dep,1} \ll 1\) and \(\tau _{dep,2} \sim \tau _{dep,1}\) ( \(| \tau _{dep,1}-\tau _{dep,2}| \sim 0\) not large enough), then \(X_{1,n} = 1\) for all n (after a very short transient), \(\lambda _{dep} = 2\), and then

    $$\begin{aligned} \Delta S_n^\# = \left[ \frac{Z_{1,n} Z_{2,n}}{a_f^2} \right] ^{\lambda _{fac}}. \end{aligned}$$
  • If \(\tau _{fac,2} \ll 1\) and \(| \tau _{fac,1}-\tau _{fac,2}|\) is large enough, then \(Z_{2,n} = a_f\) for all n (after a very short transient), \(\lambda _{fac} = 1\), and then

    $$\begin{aligned} \Delta S_n^\# = \left[ X_{1,n} X_{2,n} \right] ^{\lambda _{dep}} \frac{Z_{1,n}}{a_f}. \end{aligned}$$

    If, in addition, \(\tau _{fac,1} \ll 1\) and \(\tau _{fac,2} \sim \tau _{face,1}\) ( \(| \tau _{fac,1}-\tau _{fac,2}| \sim 0\) not large enough), then \(Z_{1,n} = a_f\) for all n (after a very short transient), \(\lambda _{fac} = 2\), and then

    $$\begin{aligned} \Delta S_n^\# = \left[ X_{1,n} X_{2,n} \right] ^{\lambda _{dep}}. \end{aligned}$$
  • It follows that if \(\tau _{dep,1}, \tau _{dep,2} \ll 1\) ( \(| \tau _{dep,1}-\tau _{dep,2}| \sim 0\) not large enough) and \(\tau _{fac,1}, \tau _{fac,2} \ll 1\) ( \(| \tau _{fac,1}-\tau _{fac,2}| \sim 0\) not large enough), then

    $$\begin{aligned} \Delta S_n^\# = 1. \end{aligned}$$

E. Descriptive rules for the generation of temporal (envelope) band-pass filters from the interplay of the temporal (envelope) low- and high-pass filters

From a geometric perspective, temporal band-pass filters in response to periodic presynaptic inputs arise as the result of the product of two exponentially increasing and decreasing functions both decaying towards their steady-state (e.g., Fig. 6). At the descriptive level, this is captured by the temporal envelope functions (F, G and \(H = F G\)) discussed above whose parameters are not the result of a sequence of single events but are related to the biophysical model parameters by comparison with the developed temporal filters. These functions provide a geometric/dynamic way to interpret the generation of temporal filters in terms of the properties of depression (decreasing functions) and facilitation (increasing functions) in response to periodic inputs, although this interpretation uses the developed temporal filters and therefore is devoid from any biophysical mechanistic interpretation.

In order to investigate how the multiplicative interaction between F(t) and G(t) given by Eqs. (31) and (32) give rise to the temporal band-pass filters \(H = F G\), we consider a rescaled version of these functions

$$\begin{aligned} F(t) = A + (1-A) e^{-t/\eta } \end{aligned}$$
(107)

and

$$\begin{aligned} G(t) = B\, [\, 1 - C e^{-t}\, ] \end{aligned}$$
(108)

where \(B = 1\) and

$$\begin{aligned} \eta = \frac{\sigma _d}{\sigma _f}. \end{aligned}$$
(109)

The function G transitions from \(G(0) = 1 - C\) to \(\lim _{t \rightarrow \infty } G(t) = 1\) with a fixed time constant (Fig. 15, green curves). The function F transitions from \(F(0) = 1\) to \(\lim _{t \rightarrow \infty } F(t) = A\) with a time constant \(\eta\) (Fig. 15, red curves). It follows that H transitions from \(H(0) = 1 - C\) to \(\lim _{t \rightarrow \infty } H(t) = A\, B = A\) (Fig. 15, blue curves). A temporal band-pass filter is generated if H raises above A for a range of values of t. This requires F to decay slow enough so within that range \(H = F G > A\) (Fig. 15-A) or A to be small enough (Fig. 15-B). In fact, as A decreases, the values of \(\eta\) required to produce a band-pass temporal filter increases (compare Fig. 15-A2 and -B2).

Changes in the parameter B in (108) affect the height of the band-pass temporal filter, but not the generation mechanism described above. However, for certain ranges of parameter values H is a temporal low-pass filter (not shown).

Fig. 15
figure 15

Temporal band-pass filters generated as the result of the multiplicative interaction of temporal low- and high-pass filters: envelope functions approach. We used the envelope functions F and G defined by (40) and (108), respectively, and \(H = F G\). A. Increasing \(\eta\) contributes to the generation of a band-pass temporal filter. We used \(A = 0.5\), \(C = 0.8\) and A1. \(\eta = 0.1\). A2. \(\eta = 1\). A3. \(\eta = 10\). B. Decreasing A contributes to the generation of a band-pass temporal filter. We used \(\eta = 1\), \(C = 0.8\) and B1. \(A = 0.2\). B2. \(A = 0.4\). B3. \(A = 0.6\)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Mondal, Y., Pena, R.F.O. & Rotstein, H.G. Temporal filters in response to presynaptic spike trains: interplay of cellular, synaptic and short-term plasticity time scales. J Comput Neurosci (2022). https://doi.org/10.1007/s10827-022-00822-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s10827-022-00822-y

Keywords

  • Synaptic depression
  • Synaptic facilitation
  • Short-term synaptic dynamics
  • Frequency dependent filters
  • Temporal resonance
  • Low-pass temporal filter
  • High-pass temporal filter
  • Band-pass temporal filter
  • Interplay of time scales across levels of organization