BCM theory


BCM theory

BCM theory, BCM synaptic modification, or the BCM rule, named for Elie Bienenstock, Leon Cooper, and Paul Munro, is a physical theory of learning in the visual cortex developed in 1981. Due to its successful experimental predictions, the theory is arguably the most accurate model of synaptic plasticity to date.

Development

In 1949, Donald Hebb proposed a working mechanism for memory and computational adaption in the brain called Hebbian learning, or the maxim that "cells that fire together, wire together". This law formed the basis of the brain as the modern neural network, theoretically capable of Turing complete computational complexityFact|date=July 2008, and thus became a standard materialist model for the mind.

However, Hebb's rule has problems, namely that it has no mechanism for connections to get weaker and no upper bound for how strong they can get. In other words, the model is unstable, both theoretically and computationally. Later modifications gradually improved Hebb's rule, normalizing it and allowing for decay of synapses, where no activity or unsynchronized activity between neurons results in a loss of connection strength. New biological evidence brought this activity to a peak in the 1970s, where theorists formalized various approximations in the theory, such as the use of firing frequency instead of potential in determining neuron excitation, and the assumption of ideal and, more importantly, linear synaptic integration of signals. That is, there is no unexpected behavior in the adding of input currents to determine whether or not a cell will fire.

These approximations resulted in the basic form of BCM below in 1979, but the final step came in the form of mathematical analysis to prove stability and computational analysis to prove applicability, culminating in Bienenstock, Cooper, and Munro's 1982 paper.

Since then, experiments have shown evidence for BCM behavior in both the visual cortex and the hippocampus, the latter of which plays an important role in the formation and storage of memories. Both of these areas are well-studied experimentally, but both theory and experiment have yet to establish conclusive synaptic behavior in other areas of the brain. Furthermore, a biological mechanism for synaptic plasticity in BCM has yet to be established. [cite journal |last=Cooper |first=L.N. |authorlink=Leon Cooper |year=2000 |month= |title=Memories and memory: A physicist's approach to the brain |journal=International Journal of Modern Physics A |volume=15 |issue=26 |pages=4069–4082 |id= |url=http://physics.brown.edu/physics/researchpages/Ibns/Lab%20Publications%20(PDF)/memoriesandmemory.pdf |accessdate= 2007-11-11 |quote= ]

Theory

The basic BCM rule takes the form

:frac{d m_j(t)}{d t}=phi( extbf{c}(t))d_j(t)-epsilon m_j(t),

where m_j is the synaptic weight of the jth synapse, d_j is that synapse's input current, extbf{c} is the weighted presynaptic output vector, phi is the postsynaptic activation function that changes sign at some output threshold heta_M, and epsilon is the (often negligible) time constant of uniform decay of all synapses.

This model is merely a modified form of the Hebbian learning rule, dot{m_j}=c d_j, and requires a suitable choice of activation function, or rather, the output threshold, to avoid the Hebbian problems of instability. This threshold was derived rigorously in BCM noting that with c(t)= extbf{m}(t)cdot extbf{d}(t) and the approximation of the average output ar{ extbf{c(t) approx extbf{m}(t)ar{d}(t), for one to have stable learning it is sufficient that

:sgnphi(c,ar{c})=sgnleft(c-left(frac{ar{c{c_0} ight)^par{c} ight) ~~ extrm{for} ~~ c>0, and:phi(0,ar{c})=0 ~~ extrm{for} ~ extrm{all} ~ ar{c},

or equivalently, that the threshold heta_M(ar{c}) = (ar{c}/c_0)^par{c}, where p and c_0 are fixed positive constants. [cite journal |last=Bienenstock |first=Elie L. |authorlink=Elie Bienenstock |coauthors=Leon Cooper, Paul Munro |year=1982 |month=January |title=Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex |journal=The Journal of Neuroscience |volume=2 |issue=1 |pages=32–48 |id= |url=http://www.physics.brown.edu/physics/researchpages/Ibns/Cooper%20Pubs/070_TheoryDevelopment_82.pdf |accessdate= 2007-11-11 |quote= ]

When implemented, the theory is often taken such that

:phi(c,ar{c})=c(c- heta_M) ~~ extrm{and} ~~ heta_M=langle c^2 angle = frac{1}{ au}int_{-infty}^t c^2(t^prime)e^{-(t-t^prime)/ au}d t^prime,

where angle brackets are a time average and au is the time constant of selectivity.

The model has drawbacks, as it requires both long-term potentiation and long-term depression, or increases and decreases in synaptic strength, something which has not been observed in all cortical systems. Further, it requires a variable activation threshold and depends strongly on stability of the selected fixed points c_0 and p. However, the model's strength is that it incorporates all these requirements from independently-derived rules of stability, such as normalizability and a decay function with time proportional to the square of the output. [cite web |url=http://www.cs.tau.ac.il/~nin/Courses/NC05/BCM.ppt |title=The BCM theory of synaptic plasticity |accessdate=2007-11-11 |last=Intrator |first=Nathan |coauthors= |date=2006-2007 |work=Neural Computation |publisher=School of Computer Science, Tel-Aviv University ]

Experiment

The first major experimental confirmation of BCM came in 1992 in investigating LTP and LTD in the hippocampus. The data showed qualitative agreement with the final form of the BCM activation function. [cite journal |last=Dudek |first=Serena M. |authorlink=Serena Dudek |coauthors=Mark Bear |year=1992 |month= |title=Homosynaptic long-term depression in area CAl of hippocampus and effects of N-methyl-D-aspartate receptor blockade |journal=Proc. Natl. Acad. Sci. |volume=89 |issue= |pages=4363–4367 |id= |url=http://www.pnas.org/cgi/reprint/89/10/4363.pdf |accessdate= 2007-11-11 |quote=|doi=10.1073/pnas.89.10.4363 ] This experiment was later replicated in the visual cortex, where BCM was originally designed to model [cite journal |last=Kirkwood |first=Alfredo |authorlink=Alfredo Kirkwood |coauthors=Marc G. Rioult, Mark F. Bear |year=1996 |month= |title=Experience-dependent modification of synaptic plasticity in rat visual cortex |journal=Nature |volume=381 |issue= |pages=526–528 |id= |url=http://www.nature.com/nature/journal/v381/n6582/abs/381526a0.html |accessdate= 2007-11-11 |quote=|doi=10.1038/381526a0 ] This work provided further evidence of the necessity for a variable threshold function for stability in Hebbian-type learning (BCM or others).

Experimental evidence has been non-specific to BCM until Rittenhouse "et al." confirmed BCM's prediction of synapse modification in the visual cortex when one eye is selectively closed. Specifically,

:logleft(frac{m_{ m closed}(t)}{m_{ m closed}(0)} ight) sim -overline{n^2}t,

where overline{n^2} describes the variance in spontaneous activity or noise in the closed eye and t is time since closure. Experiment agreed with the general shape of this prediction and provided an explanation for the dynamics of monocular eye closure versus binocular eye closure. [cite journal |last=Rittenhouse |first=Cynthia D. |authorlink= |coauthors=Harel Z. Shouval, Michael A. Paradiso, Mark F. Bear |year=1999 |month= |title=Monocular deprivation induces homosynaptic long-term depression in visual cortex |journal=Nature |volume=397 |issue= |pages=347 |id= |url=http://www.nature.com/nature/journal/v397/n6717/abs/397347a0.html |accessdate= 2007-11-11 |quote=|doi=10.1038/16922 ] The experimental results are far from conclusive, but so far have favored BCM over competing theories of plasticity.

Applications

While the algorithm of BCM is too complicated for large-scale parallel distributed processing, it has been put to use in lateral networks with some success. [cite web |url=http://www.cs.tau.ac.il/~nin/Courses/NC05/bcmppr.pdf |title=BCM Learning Rule, Comp Issues |accessdate=2007-11-11 |last=Intrator |first=Nathan |coauthors= |date=2006-2007 |work=Neural Computation |publisher=School of Computer Science, Tel-Aviv University ] Furthermore, some existing computational network learning algorithms have been made to correspond to BCM learning. [cite journal |last=Baras |first=Dorit |authorlink= |coauthors=Ron Meir |year=2007 |month= |title=Reinforcement Learning, Spike-Time-Dependent Plasticity, and the BCM Rule |journal=Neural Computation |volume= |issue=19 |pages=2245–2279 |id=2561 |url=http://eprints.pascal-network.org/archive/00002561/01/RL-STDP_Final.pdf |accessdate=2007-11-11 |quote=|doi=10.1162/neco.2007.19.8.2245 ]

References

External links

* [http://www.scholarpedia.org/article/BCM_rule Scholarpedia article]


Wikimedia Foundation. 2010.

Look at other dictionaries:

  • BCM — may refer to: * BCM, Business Consulting Master *Band Corporal Major, a warrant officer appointment in the bands of the British Household Cavalry * Banque Centrale de Madagascar , a financial institution in Madagascar * Banque Centrale de… …   Wikipedia

  • Hebbian theory — describes a basic mechanism for synaptic plasticity wherein an increase in synaptic efficacy arises from the presynaptic cell s repeated and persistent stimulation of the postsynaptic cell. Introduced by Donald Hebb in 1949, it is also called… …   Wikipedia

  • Disk encryption theory — Disk encryption is a special case of data at rest protection when the storage media is a sector addressable device (e.g., a hard disk). This article presents cryptographic aspects of the problem. For discussion of different software packages and… …   Wikipedia

  • Neural network — For other uses, see Neural network (disambiguation). Simplified view of a feedforward artificial neural network The term neural network was traditionally used to refer to a network or circuit of biological neurons.[1] The modern usage of the term …   Wikipedia

  • Artificial neural network — An artificial neural network (ANN), usually called neural network (NN), is a mathematical model or computational model that is inspired by the structure and/or functional aspects of biological neural networks. A neural network consists of an… …   Wikipedia

  • Long-term potentiation — (LTP) is a persistent increase in synaptic strength following high frequency stimulation of a chemical synapse. Studies of LTP are often carried out in slices of the hippocampus, an important organ for learning and memory. In such studies,… …   Wikipedia

  • Synaptic plasticity — In neuroscience, synaptic plasticity is the ability of the connection, or synapse, between two neurons to change in strength. There are several underlying mechanisms that cooperate to achieve synaptic plasticity, including changes in the quantity …   Wikipedia

  • Oja's rule — Oja s learning rule, or simply Oja s rule, named after a Finnish computer scientist Erkki Oja, is a model of how neurons in the brain or in artificial neural networks change connection strength, or learn, over time. It is a modification of the… …   Wikipedia

  • Synaptic weight — In neuroscience and computer science, synaptic weight refers to the strength or amplitude of a connection between two nodes, corresponding in biology to the amount of influence the firing of one neuron has on another. The term is typically used… …   Wikipedia

  • Black Consciousness Movement — The Black Consciousness Movement (BCM) was a grassroots anti Apartheid activist movement that emerged in South Africa in the mid 1960s out of the political vacuum created by the decimation of the African National Congress and Pan Africanist… …   Wikipedia


Share the article and excerpts

Direct link
Do a right-click on the link above
and select “Copy Link”

We are using cookies for the best presentation of our site. Continuing to use this site, you agree with this.