In the past decade, major effort has been devoted to understanding the structure and the electrochemical benefit of lithium-rich transition metal oxides for high-energy density lithium-ion batteries because of their potentially high specific capacity of up to 250 mAh g^-1 when the materials are electrochemically activated. However, a consensus has not been established yet on the local structure of this class of materials. It has been reported that the excess portion of lithium was homogeneously dissolved in the matrix of the transition metal layer (3b sites of space group 166 or R-3m), forming a solid solution. On the other hand, Thackeray and others believed that the excess lithium on 3b sites has a strong tendency to pair with manganese and segregate to form Li₂MnO₃ domains. Despite the discrepancy in these researchers’ findings on the local structure, both sides agreed that the excess lithium can be electrochemically activated at a potential above 4.5 V vs. Li⁺/Li to deliver more reversible specific capacity, offering great opportunity for use of these lithium-rich transition metal oxides as high-capacity cathode materials. However, the structural instability of the material after electrochemical activation causes a continuous decrease in its working potential, a phenomenon which we will call voltage fade, that hinders the commercial deployment of this class of high-capacity cathode materials. The migration mechanism of transition metal in the oxygen framework was investigated here to illustrate the voltage decay of lithium-rich transition metal oxides.  For the first time, a voltage increase was observed during the cycling of a lithium manganese oxide spinel within a wide potential window.  It was demonstrated that the voltage variation of these lithium transition metal oxides was caused by the continuous migration of transition metal ions, mostly Mn²⁺ and Co⁴⁺, between the octahedral and tetrahedral sites of oxygen framework, resulting in the ration change between the spinel component and layered component.