Our current understanding of amnesia owes much to the study of the patient HM, who in 1953, at the age of 26, underwent bilateral removal of structures in the medial temporal lobe (MTL) for treatment of intractable epilepsy. The surgery was largely successful in controlling his seizures, and it did not affect his sense of personal identity, his general fund of knowledge, or his social skills, but it left him with a global amnesia—he had a profound impairment in the ability to form new memories for events and information that he encountered after his surgery (Corkin, 1984). Thus, for example, he could not remember the plot of a movie he saw recently, could not recall what he ate at his last meal, and could not recognize the names or faces of new acquaintances. In the laboratory, HM showed profound impairments in recalling or recognizing recently presented lists of words, pictures, and other stimuli. He also showed impaired recall of events prior to his surgery, but could retrieve memories from his early life, prior to about age 16. With insights gleaned from studies of HM and subsequently from other amnesic individuals, we have achieved a more complete understanding not only of the amnesic syndrome, but also of the functional and neural architecture of normal human memory.
The cardinal feature of the amnesic syndrome is an impairment in new learning (anterograde amnesia), and this deficit is accompanied by a more variable impairment in the ability to retrieve memories formed prior to the onset of illness (retrograde amnesia). Amnesia occurs in the context of relatively preserved general intelligence and other cognitive abilities. Immediate or short-term memory is intact in amnesia. For example, an amnesic individual can hold in mind the current topic of conversation; can repeat a brief sequence of letters or numbers; and can address a new acquaintance by name upon initial introduction. After any distraction or delay, however, whether for minutes, days, or weeks, memory for this information is impaired. Thus, the anterograde memory impairment in amnesia characteristically affects long-term memory, rather than immediate or short-term memory.
Although amnesia can have a psychogenic origin, we focus here on organic amnesia, which results from structural brain damage. The impairment is typically associated with lesions of structures in the MTL (including the hippocampus and the adjacent entorhinal, perirhinal, and parahippocampal cortices), the diencephalon, and the basal forebrain. Amnesia has a variety of etiologies, including herpes simplex encephalitis, anoxia, Wernicke-Korsakoff syndrome, and cerebrovascular accidents (O’Connor, Verfaellie, & Cermak, 1995).
The anterograde memory deficit in amnesia cuts across all types of materials (e.g., verbal and visuospatial) and perceptual modalities (e.g, auditory, visual), and it includes personally experienced events (episodic memory) as well as impersonal facts and concepts (semantic memory). One of the most important insights to emerge from studies of amnesia, however, is that the impairment is not uniform across all forms of long-term memory. For example, whereas amnesic patients show deficits on “explicit” or “declarative” memory tasks, which require deliberate reference to a prior experience, they often show normal performance on “implicit” or “nondeclarative” memory tasks, which do not require deliberate reference to a prior experience, but in which memory for that experience is “implicit” in task performance.
Examples of preserved implicit memory effects in amnesia include classical eyeblink conditioning (learning to make an eyeblink in response to a tone if that tone was previously paired with an air-puff to the eye); skill learning (improved task performance with practice); and repetition priming (improved accuracy in identifying perceptually degraded words or pictures as a result of recent exposure to those stimuli). These findings suggest that, whereas explicit memory processes depend on neural circuits that are damaged in amnesia, implicit memory processes depend on neural circuits (e.g., in the cerebellum, the basal ganglia, or the neocortex) that are spared in amnesia (Gabrieli, 1998).
Even within the domain of explicit memory, not all forms of memory appear to be equally impaired in amnesia. For example, there is evidence in amnesia that memory for associations between items (e.g., remembering that the word “sample” was paired with the word “carriage” in a recently presented list) is more impaired than is memory for single items (e.g., remembering that the word “carriage” appeared in a recently presented list) (Giovanello, Verfaellie, & Keane, 2003). This behavioral finding accords well with theories that assign to the hippocampus a critical role in binding together unrelated items in memory.
Another differential impairment in amnesia is revealed when one considers two qualitatively distinct processes that support explicit memory: recollection and familiarity. Recollection refers to intentional retrieval of the content of a prior experience, whereas familiarity refers to an ease of stimulus processing that creates the sense that one has encountered that stimulus previously. By way of illustration, if one has recently made a new acquaintance, one might recollect (i.e., intentionally retrieve) that person’s name upon encountering that person again, or one might experience a sense of familiarity (ease of recognition) that confirms that one has indeed met this person before. There is some evidence to suggest that recollection is more impaired than familiarity in amnesia (Yonelinas et al., 2002). For example, some patients with damage limited to the hippocampus show impaired performance when asked to recall a recent event (a task that requires recollection), but show intact performance when asked to recognize whether a particular item was present in that event (a task that may be based on a sense of familiarity with that item).
Patients with more extensive MTL lesions show explicit memory impairments that are evident both in recall and in recognition tasks, but that are more severe in the former than in the latter instance, consistent with a greater deficit in recollection than in familiarity. Findings such as these have led to the view that recollection is mediated by the hippocampus (which is typically severely compromised in amnesia), and familiarity depends upon extra-hippocampal MTL regions such as the perirhinal cortex that are more variably and less severely compromised in amnesia (Brown & Aggleton, 2001). This issue is controversial, however, because some amnesic patients—even those with damage restricted to the hippocampus—have shown equivalent impairments in recollection and familiarity (Manns, Hopkins, Reed, Kitchener, & Squire, 2003) supporting the alternative view that all MTL regions contribute equally to these two explicit memory processes.
A classic finding in amnesia is that the retrograde memory impairment is temporally limited, such that very remote memories are spared, whereas memories from the time period closer to the onset of amnesia are compromised. This observation gave rise to an influential theory positing that the MTL is engaged in the consolidation (stabilization) of long-term memories (Squire, Cohen, & Nadel, 1984). Once the consolidation process is complete, memories no longer depend on the hippocampus, but are supported by neocortical brain regions that are not damaged in amnesia. This view provides a straightforward account of the pattern of retrograde memory loss in amnesia: Very remote memories are intact because they had already been consolidated in neocortex at the onset of amnesia, and more recent memories are lost because they were in the process of consolidation at the time of onset. By extension, the formation of new memories after the onset of amnesia is impaired due to the absence of a consolidation mechanism.
Recent studies, however, have called into question the view that very remote memories are entirely spared in amnesia. Although it seems clear that remote memory for semantic information (facts and concepts) is intact in amnesia, there is growing evidence that remote memory for episodic (autobiographical) information may be compromised. Detailed examinations of the quality and number of remote autobiographical memories in amnesia have revealed that there may be impairments that extend back in time for decades prior to the onset of amnesia. Such temporally extensive remote memory impairments cannot be explained easily by a consolidation view, because consolidation is unlikely to operate over such a lengthy time frame.
An alternative to consolidation theory known as the multiple memory trace view proposes that the hippocampus is engaged each time an autobiographical memory is retrieved (regardless of the age of that memory), and that each such retrieval establishes a novel memory trace in the hippocampus (Moscovitch, Nadel, Winocur, Gilboa, & Rosenbaum, 2006). By this view, autobiographical memories never become independent of the hippocampus and consequently are compromised in amnesia regardless of how remote they are. The degree of impairment may be smaller for very remote memories, because such memories have presumably been retrieved on numerous occasions over time, and are therefore associated with larger number of hippocampal traces and more resistant to partial hippocampal damage.
In summary, amnesia is a circumscribed memory impairment characterized by severe impairments in anterograde memory and more variable or limited impairments in retrograde memory. The functional impact of such deficits on activities of daily life is profound. In this context, the relative preservation of some forms of long-term memory (implicit memory, familiarity) takes on clinical importance, as these areas of preserved function may form the basis for the development of compensatory strategies or rehabilitative approaches in amnesia.
Declarative Memory; Hippocampus; Memory Functions.
Related Credo Articles
This article provides a comprehensive survey of amnesia as it occurs in a variety of situations, including brain damage and the effects of...
Loss of memory resulting from such causes as head injuries, drugs, hysteria, senility, or psychological illness. The memory loss may be for...
Memory is the ability to store and subsequently retrieve past experience, and is fundamental to many cognitive functions. Amnesia can be defined...