1. This paper revisits the so-called "receptor mosaic hypothesis" for memory trace formation in the light of recent findings in "functional (or interaction) proteomics." The receptor mosaic hypothesis maintains that receptors may form molecular aggregates at the plasma membrane level representing part of the computational molecular networks. 2. Specific interactions between receptors occur as a consequence of the pattern of transmitter release from the source neurons, which release the chemical code impinging on the receptor mosaics of the target neuron. Thus, the decoding of the chemical message depends on the receptors forming the receptor mosaics and on the type of interactions among receptors and other proteins in the molecular network with novel long-term mosaics formed by their stabilization via adapter proteins formed in target neurons through the incoming neurotransmitter code. The internalized receptor heteromeric complexes or parts of them may act as transcription factors for the formation of such adapter proteins. 3. Receptor mosaics are formed both at the pre- and postsynaptic level of the plasma membranes and this phenomenon can play a role in the Hebbian behavior of some synaptic contacts. The appropriate "matching" of the pre- with the postsynaptic receptor mosaic can be thought of as the "clamping of the synapse to the external teaching signal." According to our hypothesis the behavior of the molecular networks at plasma membrane level to which the receptor mosaics belong can be set in a "frozen" conformation (i.e. in a frozen functional state) and this may represent a mechanism to maintain constant the input to a neuron. 4. Thus, we are suggesting that molecular networks at plasma membrane level may display multiple "attractors" each of which stores the memory of a specific neurotransmitter code due to a unique firing pattern. Hence, this mechanism may play a role in learning processes where the input to a neuron is likely to remain constant for a while.