10) The Hebbian learning rule: 'Neurons that fire together wire together' [plus corollaries] (Hebb – 1949)
D. O. Hebb's most famous idea, that neurons with correlated activity increase their synaptic connection strength, was based on the more general concept of association of correlated ideas by philosopher David Hume (1739) and others. Hebb expanded on this by postulating the 'cell assembly', in which networks of neurons representing features associate to form distributed chains of percepts, actions, and/or concepts.
Hebb, who was a student of Lashley (see previous post), followed in the tradition of distributed processing (discounting localizationist views).
The above figure illustrates Hebb's most original hypothesis (which is yet to be proven): The reverbatory cell assembly formed via correlated activity. Hebb theorized that increasing connection strength due to correlated activity would cause chains of association to form, some of which could maintain subsequent activation for some period of time as a form of short term memory (due to autoassociation).
Implication: The mind, largely governed by reward-seeking behavior, is implemented in an electro-chemical organ with distributed and modular function consisting of excitatory and inhibitory neurons communicating via convergent and divergent synaptic connections strengthened by correlated activity.
[This post is part of a series chronicling history's top brain computation insights (see the first of the series for a detailed description)]
-MC
It is imporant to note that Hebb’s hypothesis is inaccurate because it doesn’t specify the kind of correlation required for strengthening It turns out that the correlation required for strengthening is not that of firing at the same time, but instead, at two slightly different times. The directionality of the connection between the cells and the order of firing matters: If cell A fires before cell B, then the connection from cell A to cell B is likely to be strengthened. If the reverse firing sequence occurs, that connection is likely to be weakened.
This fact is known as spike-timing-dependent plasticity — STDP. It was shown as early as 1983 in the paper Temporal Contiguity Requirements for Long-Term Associative Potentiation/Depression in the Hippocampus, by Levy & Steward (1983). Many neuroscientists have replicated the finding since then.
It’s unfortunate that the “fire together, wire together” phrase is so catchy because it really does relay an incorrect idea.
-PL
Actually, Hebb's original proposal (for his master's thesis) said that neuron A had to take part in causing neuron B to fire. He later backed down from this requirement, likely because he couldn't think of a mechanism for going beyond correlation. With the discovery of STDP that mechanism is now clear. By the way, an entry on STDP is coming up soon…
Excellent, looking forward to it.
And you’re right — I do remember hearing of Hebb’s original statement about “neuron A repeatedly taking part in causing neuron B to fire”. I suppose the unfortunate thing is really the popularization of the phrase “fire together, wire together” since it misconstrues Hebb’s original idea.
I don’t understand how spike-timing-dependent plasticity would work, and I wonder if someone could explain it to me. I imagine it would have to involve a synapse between neuron A and neuron B increasing in strength when A and B fire simultaneously (or A slightly before B).
But if the synapse is on B’s cell body or one of it’s dendrites, how can it ever “know” if neuron B has fired, given that action potentials are generated at the axon hillock and propagate only down the axon? How does the fact that B has fired or not, crucial to influencing plasticity, reach the synapse?
Nikhil,
Great question.
In reality, it appears that there are several different mechanisms which make up STDP/LTP. But here's one that answers your question:
When the postsynaptic cell fires, the electrical properties of the cell can result in a backpropogating action potential (bAP) back along its dendrites. (Note that this bAP will *not* cross the synapse into other presynaptic cells). So not only do action potentials travel forward along the axon, some depolarization also goes back along the dendrites.
The idea then is that the bAP causes depolarization at a particular synapse, repelling the magnesium block from the NMDA-type glutamate receptors. LTP can then occur when the presynaptic cell releases glutamate into the synapse.
Search for "distal dendrites" "depolarize" and "backpropagating action potential" with your LTP/STDP queries to find out more.
-PL