The processing of a target speech under noisy environment is influenced by the presence of concurrent sounds, this phenomenon is referred to as auditory masking. Previous research showed that there are two kinds of masking, energetic masking and informational masking. Energetic masking originated from auditory periphery, where the masking sound overlaps with the target sound in time and frequency, resulting in an incomplete representation of the latter. On the other hand, informational masking is believed to occur in the central auditory system. However, for a long time, the concept of informational masking has been used as a “suitcase” or “catchall” word. All the influencing effect that can not be explained by energetic masking is attributed to informational masking, though its underlying mechanism is still largely unknown. In this article, a new conceptual framework of informational masking is proposed. In this framework, informational masking is divided into two subcomponents, perceptual informational masking and cognitive informational masking. Perceptual masking is caused by the competition between the sound stream/object, while cognitive informational masking is caused by the higher cognitive information included in the masking sounds, which could be automatically processed by the auditory or linguistic system. Psychophysical and neural imaging studies are proposed to test the above hypothesis. In behavioral studies, the perceived spatial separation between the target and the masker, the intelligibility and the voice number in the masking sound, as well as the perceptual similarity between the target and masker streams are manipulated to demonstrate that perceptual masking and cognitive masking could be double-dissociated. In the neural imaging study, fMRI technique is used to show that perceptual and cognitive informational masking may have different neural mechanism.