Elsevier

Journal of Physiology-Paris

Volume 97, Issues 4–6, July–November 2003, Pages 391-402
Journal of Physiology-Paris

Homeostatic plasticity in the CNS: synaptic and intrinsic forms

https://doi.org/10.1016/j.jphysparis.2004.01.005Get rights and content

Abstract

The study of experience-dependent plasticity has been dominated by questions of how Hebbian plasticity mechanisms act during learning and development. This is unsurprising as Hebbian plasticity constitutes the most fully developed and influential model of how information is stored in neural circuits and how neural circuitry can develop without extensive genetic instructions. Yet Hebbian plasticity may not be sufficient for understanding either learning or development: the dramatic changes in synapse number and strength that can be produced by this kind of plasticity tend to threaten the stability of neural circuits. Recent work has suggested that, in addition to Hebbian plasticity, homeostatic regulatory mechanisms are active in a variety of preparations. These mechanisms alter both the synaptic connections between neurons and the intrinsic electrical properties of individual neurons, in such a way as to maintain some constancy in neuronal properties despite the changes wrought by Hebbian mechanisms. Here we review the evidence for homeostatic plasticity in the central nervous system, with special emphasis on results from cortical preparations.

Introduction

How the properties of neurons are regulated by experience, by their own activity and that of their neighbors, is crucial for our understanding of the nervous system. In the absence of implausibly extensive genetic instructions it is difficult to imagine how neural circuits form correctly except by means of experience-dependent modifications; and, once formed, circuits must be capable of adapting their properties and those of their constituent parts, in order to allow animals to learn and function in a dynamic environment. Consequently neural plasticity has been the subject of an enormous and varied research effort. The great bulk of this effort has been focused on what might be called “Hebbian plasticity,” the idea that correlated presynaptic and postsynaptic activity strengthens synapses whereas uncorrelated activity weakens them. Partly, this is because a compelling body of experimental evidence indicates that synapses can in fact be modified in this way, most notably through long-term potentiation and depression (LTP and LTD), with some of the evidence even linking LTP and LTD to learned behavior [7], [40], [52], [53]. And partly, Hebbian plasticity has been the object of intense study because it is so intuitively attractive: it invests neural circuits with immense computational power because each of the hundreds or thousands of synapses onto a given neuron can, in principle, be modified independently by experience. This intuition has been borne out by theoretical models which explore the consequences of Hebbian plasticity rules, and find that they can be used not only to store information in neural networks but to explain the development of those networks [18], [44].

However, these models have also demonstrated that Hebbian plasticity alone is unlikely to account for the stable functioning of neural circuits. One problem is created by the nature of the correlation-based rules themselves: instability due to positive feedback. Consider two synaptically coupled neurons. Strengthening the synapse between them would allow the presynaptic neuron to drive the postsynaptic one more strongly; this would tend to strengthen the synapse even more, by a correlation-based rule; this would lead to more driving; this would lead to more strengthening; and so on. Absent constraints of some kind Hebbian rules can lead to situations in which all synapses are saturated (or have fallen to 0). A second––related but distinct––problem arises as a result of both Hebbian change and activity-independent development: instability due to fluctuations in average synaptic input. At many points in its life, but especially during development, the baseline amount of synaptic input received by a neuron might vary substantially, because synapses are continually being formed, eliminated, and reconfigured, and because developmental events, like eye-opening, change the amount of sensory drive. This brings up the question of how neurons keep their firing rates from saturating or falling silent if average input rises too high or falls too low. Simple correlation-based modification rules are not well poised to answer this question.

These considerations have led researchers to augment Hebbian rules in computational models with additional rules for constraining synaptic weights and/or neuronal firing [43], [45]. Among these are schemes that make the ability of synapses to undergo Hebbian modification depend upon their history of use, that limit the total synaptic strength over a neuron, or that modulate its intrinsic excitability as function of average activity. These schemes have sometimes had an ad hoc quality because they have not been well-constrained by biological evidence; instead the additional rules have simply been introduced by hand, without a strong appeal to empirical data. But within the last decade or so considerable progress has been made in giving an experimental grounding to this notion of “homeostatic plasticity,” i.e., regulatory processes that work to maintain the stability and functionality of neural networks subject to Hebbian synaptic change [1], [16], [59]. In this article we will review the major experimental evidence for homeostatic forms of plasticity in the mammalian central nervous system. Three candidate plasticity mechanisms will be covered: synaptic scaling, the Bienenstock–Cooper–Munro (BCM) synapse, and regulation of intrinsic excitability.

Section snippets

Synaptic scaling

A frequent approach to the problem of stability is constraining the total synaptic strength over a neuron. This is particularly effective when the constraints are imposed as a function of postsynaptic activity or total synaptic efficacy. For example, one might scale the strengths of all excitatory synapses onto a given neuron up or down depending upon whether average spiking activity (with the average taken over some appropriately slow time scale) is below or above a target value. In this way

BCM model

A different approach to the problem of stability in Hebbian networks is to make the capacity of a synapse to undergo Hebbian modification depend upon its history of use or upon the history of neuronal activity [2], [3]. This idea was put into its best-known form 20 years ago by Bienenstock, Cooper, and Munro (BCM) [9], and has since been the subject of considerable experimental and theoretical exploration [7], [13], [14], [30], [33], [64].

Intrinsic plasticity

Most efforts to understand activity-dependent plasticity in the nervous system have been concerned with synaptic plasticity, with how synaptic connections between neurons are modified by experience. This makes sense because a compelling body of evidence indicates that synaptic properties are plastic on several different time scales and in multiple ways (changes in strength, dynamics, distribution). But focusing on synaptic plasticity exclusively gives an incomplete and sometimes flawed picture

Open questions

As we have seen, experimental work from a number of preparations indicates the existence of non-Hebbian forms of long-lasting plasticity: synaptic scaling, the BCM synapse, and regulation of intrinsic excitability. Each of these has properties that would allow it to help stabilize networks subject to Hebbian and developmental change. Consequently all three have been labeled “homeostatic.” This is a fairly general term, which can mean different things in different contexts, and in using the term

Acknowledgements

I am grateful to Elisabeth C. Walcott for reading the manuscript. This work was supported by the Neurosciences Research Foundation.

References (65)

  • A. Rao et al.

    Activity regulates the synaptic localization of the NMDA receptor in hippocampal neurons

    Neuron

    (1997)
  • L.C. Rutherford et al.

    BDNF has opposite effects on the quantal amplitude of pyramidal neuron and interneuron excitatory synapses

    Neuron

    (1998)
  • G.G. Turrigiano

    Homeostatic plasticity in neuronal networks: the more things change, the more they remain the same

    Trends Neurosci.

    (1999)
  • A.J. Watt et al.

    Activity co-regulates quantal AMPA and NMDA currents at neocortical synapses

    Neuron

    (2000)
  • L.F. Abbott et al.

    Synaptic plasticity: taming the beast

    Nature Neurosci.

    (2000)
  • W.C. Abraham et al.

    Heterosynaptic metaplasticity in the hippocampus in vivo

    Proc. Natl. Acad. Sci. USA

    (2001)
  • C.D. Aizenman et al.

    Rapid, synaptically driven increases in the intrinsic excitability of cerebellar deep nuclear neurons

    Nature Neurosci.

    (2000)
  • S. Armano et al.

    Long-term potentiation of intrinsic excitability at the mossy fiber-granule cell synapse of rat cerebellum

    J. Neurosci.

    (2000)
  • G.-Q. Bi et al.

    Synaptic modification by correlated activity: Hebb's postulate revisited

    Ann. Rev. Neurosci.

    (2001)
  • E.L. Bienenstock et al.

    Theory for the development of neuron selectivity: orientation specificity and binocular interaction in the visual cortex

    J. Neurosci.

    (1982)
  • S.G. Brickley et al.

    Adaptive regulation of neuronal excitability by a voltage-independent potassium conductance

    Nature

    (2001)
  • W.B. Cannon

    A law of denervation

    Am. J. Med. Sci.

    (1939)
  • G.C. Castellani et al.

    Solutions of the BCM learning rule in a network of lateral interacting nonlinear neurons

    Network

    (1999)
  • E.E. Clothiaux et al.

    Synaptic plasticity in visual cortex: comparison of theory with experiment

    J. Neurophysiol.

    (1991)
  • H. Colman et al.

    Alterations in synaptic strength preceding axon withdrawal

    Science

    (1997)
  • G.W. Davis et al.

    Maintaining the stability of neuronal function: a homeostatic hypothesis

    Ann. Rev. Physiol.

    (2001)
  • G.W. Davis et al.

    Synapse-specific control of synaptic efficacy at the terminals of a single neurons

    Nature

    (1998)
  • P. Dayan et al.

    Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems

    (2001)
  • N.S. Desai et al.

    Activity-dependent regulation of intrinsic properties in rat visual cortical cultures

    Neurocomputing

    (1999)
  • N.S. Desai et al.

    Plasticity in the intrinsic excitability of cortical pyramidal neurons

    Nature Neurosci.

    (1999)
  • N.S. Desai et al.

    BDNF regulates the intrinsic excitability of cortical neurons

    Learn Mem.

    (1999)
  • N.S. Desai et al.

    Critical periods for experience-dependent synaptic scaling in visual cortex

    Nature Neurosci.

    (2002)
  • Cited by (124)

    • Repeated applications of high potassium elicit long-term changes in a motor circuit from the crab, Cancer borealis

      2022, iScience
      Citation Excerpt :

      The shortest activity-dependent adaptation processes, such as spike frequency adaptation, emerge directly because of the biophysical properties of channels and can occur on the millisecond timescale. Over longer time frames, activity-dependent homeostatic mechanisms actively regulate ion channel expression and synaptic weights to maintain stable function in the face of extended physiological perturbation (Brickley et al., 2001; Desai, 2003; Golowasch et al., 1999; Hengen et al., 2013; Marder and Goaillard, 2006; Mease et al., 2013; O'Leary and Marder, 2014; Temporal et al., 2014; Turrigiano, 2012; Turrigiano et al., 1998; Turrigiano and Nelson, 2004). These homeostatic processes are commonly thought to act over hours to days and require protein synthesis.

    • Homeostatic Intrinsic Plasticity Is Functionally Altered in Fmr1 KO Cortical Neurons

      2019, Cell Reports
      Citation Excerpt :

      Although it is clear that the mechanisms underlying cortical hyperexcitability in FXS are complicated and multifactorial, recognition of a guiding principle driving the expression of these altered mechanisms would be transformative. One set of mechanisms that is known to be important for ensuring appropriate levels of excitability is referred to as homeostatic plasticity (Davis, 2013; Desai, 2003; Turrigiano, 2011). Homeostatic mechanisms are expressed in the nervous system to maintain an appropriate level of neuronal activity despite ongoing challenges to the network.

    View all citing articles on Scopus
    View full text