A mechanism for the Hebb and the anti-Hebb processes underlying learning and memory.

In a previous paper, a model was presented showing how the group of Ca2+/calmodulin-dependent protein kinase II molecules contained within a postsynaptic density could stably store a graded synaptic weight. This paper completes the model by showing how bidirectional control of synaptic weight could be achieved. It is proposed that the quantitative level of the activity-dependent rise in postsynaptic Ca2+ determines whether the synaptic weight will increase or decrease. It is further proposed that reduction of synaptic weight is governed by protein phosphatase 1, an enzyme indirectly controlled by Ca2+ through reactions involving phosphatase inhibitor 1, cAMP-dependent protein kinase, calcineurin, and adenylate cyclase. Modeling of this biochemical system shows that it can function as an analog computer that can store a synaptic weight and modify it in accord with the Hebb and anti-Hebb learning rules.