Alcohol and memory: General
       In order to evaluate the effects of alcohol, or any other drug, on memory, one must first identify a model of memory formation and storage to use as a reference.  One classic, oft cited model, initially proposed by Atkinson and Shiffrin (1968), posits that memory formation and storage takes place in several stages, proceeding from sensory memory (which lasts up to a few seconds) to short-term memory (which lasts from seconds to minutes depending upon whether the information is rehearsed) to long-term storage. This model is often referred to as the modal model of memory, as it captures key elements of several other major models. Indeed, elements of this model can still be seen in virtually all models of memory formation. In the model, when one attends to sensory information, it is transferred from a sensory memory store to short-term memory.  The likelihood that information will be transferred from short-term to long-term storage, or be encoded into long-term memory, was thought to depend primarily on how long the individual keeps the information active in short-term memory via rehearsal. While rehearsal clearly influences the transfer of information into long-term storage, it is important to note that other factors, such as the depth of processing (i.e., the level of true understanding and manipulation of the information), motivation, and arousal also play important roles (Craik and Lockhart 1972; Otten et al. 2001; Eichenbaum, 2002) .¹

General model of memory showing the primary effects of alcohol
The figure displays a general model of memory formation, storage, and retrieval based on the modal model of memory originally proposed by Atkinson and Shiffrin (1968). Alcohol seems to influence most stages of the process to some degree. However, its primary effect appears to be on the transfer of information from short-term to long-term storage. Intoxicated subjects are typically able to recall information immediately after its presentation and even keep it active in short-term memory for one minute or more if they are not distracted. Subjects are also normally able to recall long-term memories formed prior to becoming intoxicated. However, beginning with just one or two drinks, subjects begin to show impairments in the ability to transfer information into long-term storage. Under some circumstances, alcohol can impact this process so severely that, once sober again, subjects are unable to recall critical elements of events, or even entire events, that occurred while intoxicated. Such impairments are known as blackouts.

       Variability in the use of terms, particularly in operational definitions of short-term memory, makes it difficult to formulate a simple synopsis of the literature on alcohol-induced memory impairments. As Mello (1973) stated three decades ago with regard to the memory literature in general, "The inconsistent use of descriptive terms has been a recurrent source of confusion in the ´short-term´ memory literature and ´short-term´ memory has been variously defined as 5 seconds, 5 minutes, and 30 minutes" (p. 333).  In spite of this inconsistency, several conclusions can be drawn from research on alcohol-induced memory impairments. One conclusion is that the impact of alcohol on the formation of new long-term "explicit" memories-that is, memories of facts (e.g., names and phone numbers) and events-is far greater than the drug's impact on the ability to recall previously established memories or to hold new information in short-term memory (Lister et al. 1991). (See figure 1 for a diagram depicting the stages of memory and where alcohol interferes with memory.) Intoxicated subjects are typically able to repeat new information immediately after its presentation and can often keep it active in short-term storage for up to a few minutes if they are not distracted (for an early review, see Ryback 1971), though this is not always the case (Nordby et al. 1999). Similarly, subjects normally are capable of retrieving information placed in long-term storage prior to acute intoxication. In contrast, alcohol impairs the ability to store information across delays longer than a few seconds if subjects are distracted between the time that they are given the new information and the time that they are tested, thus preventing rehearsal. In a classic study, Parker and colleagues (1976) reported that when intoxicated subjects were presented with "paired associates"-for example, the letter "B" paired with the month "January", they were impaired when asked to recall the items after delays of a minute or more. However, subjects could recall paired associates that they had learned before becoming intoxicated. More recently, Acheson and colleagues (1998) observed that intoxicated subjects could recall items on words lists immediately after the lists were presented but were impaired when asked to recall the items 20 minutes later.
       Ryback (1971) characterized the impact of alcohol on memory formation as a dose-related continuum with minor impairments at one end and large impairments at the other, and with all impairments representing the same fundamental deficit in the ability to transfer new information from short-term to long-term storage. When doses of alcohol are small to moderate (producing blood alcohol concentrations (BACs) below 0.15%), memory impairments tend to be small to moderate as well. At these levels, alcohol produces what Ryback (1971) referred to as cocktail party memory deficits, lapses in memory that people might experience after having a few drinks at a cocktail party, often manifested as problems remembering what another person said or where they were in conversation. Several studies have revealed that alcohol at such levels causes difficulty forming memories for items on word lists or learning to recognize new faces (Westrick et al. 1988; Mintzer et al. 2002). As the dose increases, the resulting memory impairments can become much more profound, sometimes culminating in blackouts-periods for which a person is unable to remember critical elements of events, or even entire events, that transpired while he or she was intoxicated.

1 It is well beyond the scope of this review to assess the impact of alcohol on memory utilizing multiple perspectives on information processing and storage. For simplicity, this review will characterize the effects of alcohol on memory using a three-stage process of memory formation akin to the modal model. However, the interpretation of the effects of alcohol on memory likely would vary somewhat depending on the memory model that one uses.