Hebbian Learning
A mechanism for neural plasticity: 'cells that fire together wire together.'
In Plain Language
Two nerve cells sit joined by a tiny junction. Whenever one fires and helps set off the other, that junction grows a little stronger, so next time the first cell has an easier job triggering the second. Nobody supervises this — the two cells only know their own activity. Do it enough and things that keep happening together become linked, which is how one smell can drag up a whole memory.
Consider a single junction between two nerve cells. If one of them repeatedly takes part in firing the other, Hebb proposed, some growth process strengthens the connection between them. The rule is local: it depends only on the activity of the two cells either side of the junction, with no teacher, no target answer and no signal arriving from elsewhere. Translated into artificial networks it becomes an instruction to increase the weight between two units in proportion to the product of their activities, so units that are active together get reinforced and the network gradually comes to encode which inputs tend to co-occur. That also yields associative memory — present part of a stored pattern and the strengthened links pull the remainder back. Left as stated the rule only ever adds, so weights grow without limit; working versions bolt on a decay term, a ceiling, or some form of normalisation to keep them in check.
Drafted with AI and edited by hand (claude-opus-5, reviewed 2026-08).
Why It Was Important
Donald Hebb introduced this neurophysiological postulate proposing how learning occurs in the brain through the strengthening of synapses between repeatedly co-activated neurons. In AI, this became the basis for unsupervised learning algorithms and weight-updating mechanisms in artificial neural networks.
Who Invented It
Donald Hebb
Canadian psychologist who profoundly influenced the area of neuropsychology.
Applications
- Unsupervised Learning
- Synaptic Plasticity Modeling
- Associative Memory
Key Papers
- The Organization of Behavior: A Neuropsychological Theory
Donald O. Hebb · Wiley · 1949