J.F.K., L.B.J., and H.M.: the famous memories of a famous amnesic.

Since publication of the landmark article describing his severe memory impairment (Scoville and Milner, 1957), patient H.M. has been a touchstone for research on amnesia and memory systems (Corkin, 2002). After surgical resection of a majority of his medial temporal lobes, H.M. experienced profound forgetfulness. Yet it was the observation of spared cognitive abilities such as attention and language combined with the subsequent demonstration of intact skill learning (e.g., Milner, 1962) that made his case surprising and truly fascinating. The dramatic combination of severity and purity in his memory impairment spurred decades of research in humans and animals that has lead to our current understanding of memory systems. The medial temporal lobe structures that were damaged in patient H.M. support what one typically thinks of as memory, a capacity termed declarative memory, but are not critically involved in examples of nondeclarative memory such as priming or habit learning or in other nonmemory cognitive abilities (Squire, 1992; Gabrieli, 1998; Eichenbaum and Cohen, 2001). Almost as impressive as the initial breakthrough is the fact that H.M. continues to contribute to the study ofmemory. In this issue ofHippocampus,O’Kane, Kensinger, and Corkin present new findings showing that H.M. has learned knowledge about the world that could have been acquired only after his surgery. The authors asked H.M. to recall last names of famous individuals when cued either by the individual’s first name or by first name plus a brief description. H.M. recalled 12 of 35 names (e.g., Johnson) when cued by first name (e.g., Lyndon) and 11 of the remaining 23 (23 of 35 total) when additionally cued with a brief description of the individual. In a second experiment,H.M. recalled extra factual information for a number of famous names that he correctly recognized. For example, he described John F. Kennedy (J.F.K.) as a Catholic president whowas assassinated. That his learning is below normal is not surprising. What is surprising is that he learned at all. Why are the findings surprising? First, previous accounts of H.M.’s impairment reported almost no learning of factual information (e.g., Gabrieli et al., 1988). The current findings are the first in five decades of study to report substantive and unmistakable postmorbid declarative learning in patient H.M. Second, H.M.’s medial temporal lobe damage is extensive and includes partial damage to the perirhinal cortex, complete destruction of the entorhinal cortex, and a combination of damage and disconnection that has left the hippocampus nonfunctional. Thus, the learningwas supported either by the parahippocampal cortex and damaged perirhinal cortex or by structures outside the medial temporal lobe. Third, his ability to retrieve day-to-day memory is reportedly nonexistent. The latter two findings are consistent with the view that the hippocampus contributes crucially to episodic memory, but not to semantic memory (Vargha-Khadem et al., 1997; Tulving andMarkowitsch, 1998), an interpretation considered byO’Kane et al. (this issue). Episodicmemory refers to the ability to reexperience the first-hand specifics of an event. Semantic memory refers to knowledge about the world that is not necessarily tied to a particular incident. H.M. acquired knowledge about the world despite an inability to recollect details of specific instances. Thus, it is possible that damage to the hippocampus and other medial temporal lobe structures eliminated his capacity for episodic remembering but left partially intact his ability to acquire new semanticmemory. If so, the results of the study support the connection that has been argued to exist between the specialized anatomy and plasticity of the hippocampus and the uniqueness of episodic memory (Mishkin et al., 1998; Nadel et al., 2000). The authors also consider the possibility that the difference between the knowledge that H.M. was able to acquire and the episodic details that invariably escape him might have more to do with the opportunities for learning rather than the category of retrieval (Alvarez and Squire, 1994; McClelland et al., 1995;Holdstock et al., 2002). The information that H.M. has learned since his surgery could have been encountered numerous times over the decades. In contrast, episodic memory by definition can draw on only a single incident. Thus, it is possible that H.M.’s damage did not impair episodic memory per se but rather severely disrupted declarative memory in such a way that one-trial learning was abolished, allowing learning only through extended repetition. The results of the article by O’Kane et al. cannot distinguish between the two possibilities, and the authors do not consider the alternatives to be incompatible. Indeed, the similarity between the concepts of episodic memory and single-trial declarative memory would make resolving the alternatives challenging. Nevertheless, it is useful to consider whether episodic memory captures the entirety of the contribution made by the hippocampus to declarative memory or whether a more general, mechanistic description of its function is warranted. Several points are worth discussing. First,H.M. shows the opposite pattern of performance as compared to his controls. H.M.’s best performance comes from the decades immediately following his surgerywhereas Center for Memory and Brain, Boston University, Boston, Massachusetts Grant sponsor: National Institutes of Health; Grant number: MH068982. *Correspondence to: Joseph R. Manns, Center for Memory and Brain, Boston University, 2 Cummington Street, Boston, MA 02215. E-mail: [email protected] Accepted for publication 22 January 2004 DOI 10.1002/hipo.20010 HIPPOCAMPUS 00:000–000 (2004)