Spreading depression (SD) as well as hypoxia-induced SD-like depolarization in forebrain gray matter are characterized by near complete depolarization of neurons. The biophysical mechanism of the depolarization is not known. Earlier we reported that simultaneous pharmacological blockade of all known major Na⁺ and Ca²⁺ channels prevents hypoxic SD. We now recorded extracellular voltage, Na⁺, and K⁺concentrations and the intracellular potential of individual CA1 pyramidal neurons during hypoxia of rat hippocampal tissue slices after substituting Na⁺ in the bath by an impermeant cation, or in the presence of channel blocking drugs applied individually and in combination. Reducing extracellular Na⁺ concentration [Na⁺]_o to 90 mM postponed the hypoxia-induced extracellular DC-potential deflection (ΔV _o) and reduced its amplitude, and it also postponed the SD-like depolarization of neurons. After lowering [Na⁺]_o to 25 mM, SD-like ΔV _o became very small, indicating that an influx of Na⁺ is required for SD; influx of Ca²⁺ ions alone is not sufficient. We then asked whether the SD-related Na⁺ current flows through glutamate-controlled and/or through voltage-gated Na⁺channels. Administration of either the non-N-methyl-d-aspartate (NMDA) receptor antagonist 6,7-dinitroquinoxaline-2,3-dione (DNQX), or the NMDA receptor antagonist (±)-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP) postponed the hypoxic ΔV _o and depressed its amplitude but the effect of the combined administration of these two drugs was not greater than that of either alone. During the early phase of hypoxia, before SD onset, [K⁺]_oincreased faster and reached a much higher level in the presence of glutamate antagonists than in their absence. The [K⁺]_o level reached at the height of hypoxic SD was, however, not affected. When TTX was added to DNQX and CPP, SD was prevented in half the trials. When SD did occur, it was greatly delayed, yet eventually neurons depolarized to the same extent as in normal solution. The SD-related sudden drop in [Na⁺]_o was depressed by only 19% in the presence of the three drugs, indicating that Na⁺ can flow into cells through pathways other than ionotropic glutamate receptors and TTX-sensitive Na⁺ channels. We conclude that, when they are functional, glutamate-receptor-mediated and voltage-gated Na⁺ currents are the major generators of the self-regenerative rapid depolarization, but in their absence other pathways can sometimes take their place. The final level of SD-like depolarization is determined by positive feedback and not by the number of channels available. A schematic flow chart of the events generating hypoxic SD is discussed.