A theory is proposed that biochemical changes at the synapse that occur as a result of stimulation of specific neuronal circuits can lead to long-term changes only if alterations occur in synaptic structures in these circuits. The main synaptic structure that is thought to undergo this alteration is the postsynaptic density (PSD). There are many reports in the literature of overall structural changes at the synapse, including the PSD, resulting from various neuronal stimuli. These structural changes are here envisaged to include those of concentration and conformation of PSD proteins, changes that could alter the neural physiology of dendritic spines and even that of the presynaptic terminal.When a cell biologist looks at the central nervous system (CNS), he or she is immediately struck by the singular occurrence of a structure called the postsynaptic density (PSD), which is not found in other tissues. During the last decade, due to biochemical results from work on isolated preparations and due to immunological studies on tissue in situ, much has been learned about the protein composition and properties of this structure (cf. ref. 1). I will not enumerate all of these results, but instead focus on some which, together with published work in other neuroscience fields, could lead one to the conclusion that this structure has great meaning for changes in CNS function that last for comparatively long periods of time.As seen in countless electron micrographs, the cerebral cortex PSD is a disc, ranging from 300 to 500 nm in diameter and from 50 to 60 mm in thickness, lying tightly apposed to the postsynaptic membrane (2, 3). In many cases it has a large perforation in the center, thus resembling a doughnut (4, 5). The PSD seems to be composed of filaments and particles, with some of the filaments apparently extending into the interior of the cell (5-7). One of the postulated functions of this structure has been as a constrictor of movement of the membrane proteins, neurotransmitter receptors, and ion channels, which undoubtedly occupy that part of the membrane to which the PSD is attached (8,9). Of the proteins that make up the filamentous structure, actin (2, 10-12) and fodrin (13, 14) or brain spectrin, an actin-and calmodulin-binding protein, are found in the PSD, with tubulin probably also being a part of the structure (2,10,11,(15)(16)(17)(18). These proteins could account for the tightness of binding of the PSD to the membrane, a tightness that can be inferred from the observations that, in a synaptosomal membrane preparation, only those portions of the presynaptic and postsynaptic membranes at the synapse are connected to each other (3), as if bound together through the proteins of the PSD linking it to the postsynaptic membrane, and probably via the synaptic cleft, linking the pre-and postsynaptic membranes. Detergent disruption of the membranes allows one to isolate PSDs (3) free of their membrane attachment. The binding of the PSD structure to the postsynaptic membrane could conceivably come about th...