State-dependent memory or state-dependent learning is the phenomenon where people remember more information if their physical or mental state is the same at time of encoding and time of recall. State-dependent memory is heavily researched in regards to its employment both in regards to synthetic states of consciousness (such as under the effects of psychoactive drugs) as well as organic states of consciousness such as mood. While state-dependent memory may seem rather similar to context-dependent memory, context-dependent memory involves an individual's external environment and conditions (such as the room used for study and to take the test) while state-dependent memory applies to the individual's internal conditions (such as use of substances or mood).

In 1784, a French aristocrat named Marquis de Puységur, realized that when people were put in a hypnotic state then awoken, they had no recollection of what they were told. However, when they were put back under hypnosis, in the state they would be able to recall everything from the last time.

In 1910, a man named Morton Prince came to a realization about dreams. He hypothesized that the reason we have a hard time remembering our dreams when we wake up is not due to the fact that we are unable to, but because dreams are not like the real world.

In 1937, at the University of Illinois, Edward Girden and Elmer Culler conducted an experiment on conditioned responses in dogs under the influence of the drug curare. In the experiment, dogs were taught a conditioned muscular response – to draw their paw away from the ground when they heard a buzzer. The buzzer was often accompanied by a small electric shock, which motivated the response. For dogs that had been under the influence of curare when they first learned the response, after the curare was no longer in their system, they were less likely to remember to draw their paw away on hearing the buzzer. Once they were given curare again, the response returned. This result indicated that the dogs' ability to recall the responses was connected to their state of consciousness. Girden and Culler's research opened the door for further investigation of the influences of state of consciousness on an organism's ability to encode memory.

Following this discovery, other researchers looked into the effect of different states of being on the ability to learn and remember responses or information. In 1964, Donald Overton conducted a study as a direct response to Girden and Culler's 1937 experiment. The study tested the effects of sodium pentobarbital on rats' abilities to learn and remember certain taught responses. These rats were randomly assigned to one of two groups – substance administered or no substance administered (the control condition) – and then placed in a simple maze and taught to escape an electrical shock. Overton found that the rats that had been administered 25 mg of sodium pentobarbital could no longer remember the proper escape response when they were later placed in the maze without the drug. However, if these rats were administered sodium pentobarbital once again and placed in the maze, they recalled the escape response they had been taught. Similarly, when Overton taught a rat the escape response under the control condition (no sodium pentobarbital administered), it could not recall that behavior when it was administered the drug and asked to perform later on. Results strongly indicated that rats performed the learned response more efficiently when in the either sodium pentobarbital or control state that they were in when they first learned it. In regard to this idea the study specifically stated "a response learned under the influence of a particular drug will subsequently reoccur (with maximum strength) only when that drug condition is reinstated".

In 1969, Hoine, Bremer, and Stern conducted a test with two main parts. The participants were given time to study and just before they were tested they were asked to consume 10 ounces of Vodka. The very next day they did the same thing except some were intoxicated, while others remained sober. The results found that whether the students were sober or intoxicated they did well, but only if the state they were in was the same when they studied and when they were tested. In other words if they were intoxicated while they studied, then they did better taking the test in the same state. If they were sober when they studied then they got their best results while sober.

In later years, similar studies confirmed that learning could be state-dependent. In 1971, Terry Devietti and Raymond Larson conducted a similar study in rats, seeing how memory was affected by various levels of electric shock. Their results supported the idea that the rats' ability to remember a learned response was influenced by their state. The phenomenon continued to be studied more than thirty years later. In 2004, Mohammad-Reza Zarrindast and Ameneh Rezayof studied mice, to see how memory and learning were affected by morphine. They found that when mice learned a response under the influence of morphine, they later performed it most efficiently under the influence of morphine. When mice learned the response free of morphine, they recalled it best when similarly sober. And for mice that were taught the response under the influence of morphine, once the drug wore off, they suffered amnestic effects; they could no longer remember the learned response.

The results of each of these studies points to the existence of a state-dependent memory phenomenon. Further research on the subject continues to be carried out today in order to discover further implications of state-dependent memory or other situations in which state-dependent memory might take place.