H.M. had been knocked down by a bicycle at the age of 7, began to have minor seizures at age 10, and had major seizures after age 16. (The age of the bicycle accident is given as 9 in some reports; for clarification see Corkin, 1984.) He worked for a time on an assembly line but, finally, in 1953 at the age of 27 he had become so incapacitated by his seizures, despite high doses of anticonvulsant medication, that he could not work or lead a normal life. Scoville offered H.M. an experimental procedure that he had carried out previously in psychotic patients, and the surgery was then performed with the approval of the patient and his family.

When Milner first visited H.M., she saw that the epilepsy was now controlled but that his memory impairment was even more severe than in Penfield’s two patients, P.B. and F.C. What she observed was someone who forgot daily events nearly as fast as they occurred, apparently in the absence of any general intellectual loss or perceptual disorder. He underestimated his own age, apologized for forgetting the names of persons to whom he had just been introduced, and described his state as “like waking from a dream ... every day is alone in itself...” (Milner et al., 1968, p. 217).

The first observations of H.M., and the results of formal testing, were reported a few years later (Scoville and Milner, 1957). This publication became one of the most cited papers in neuroscience (nearly 2500 citations) and is still cited with high frequency. H.M. continued to be studied for five decades, principally by Brenda Milner, her former student Suzanne Corkin, and their colleagues (Corkin, 1984, 2002; Milner et al., 1968). He died on December 2, 2008, at the age of 82. It can be said that the early descriptions of H.M. inaugurated the modern era of memory research. Before H.M., due particularly to the influence of Karl Lashley, memory functions were thought to be widely distributed in the cortex and to be integrated with intellectual and perceptual functions. The findings from H.M. established the fundamental principle that memory is a distinct cerebral function, separable from other perceptual and cognitive abilities, and identified the medial aspect of the temporal lobe as important for memory. The implication was that the brain has to some extent separated its perceptual and intellectual functions from its capacity to lay down in memory the records that ordinarily result from engaging in perceptual and intellectual work.

The findings from H.M. were initially met with some resistance, especially because of the difficulty for many years of demonstrating anything resembling his impairment in the experimental animal. Efforts to establish an animal model in fact began almost immediately when Scoville himself came to Montreal and did the same surgery in monkeys that he had done with H.M. But these monkeys and others with medial temporal lesions seemed able to learn tasks that H.M. could not learn. Only much later did it become understood that apparently similar tasks can be learned in different ways by humans and monkeys. For example, the visual discrimination task, which is learned gradually by the monkey over hundreds of trials, proved to involve what one would now call habit learning. In the monkey, this kind of learning depends on the basal ganglia, not the medial temporal lobe. Eventually, tasks were developed for the monkey that were exquisitely sensitive to medial temporal lobe lesions (for example, the one-trial, delayed nonmatching to sample task), and an animal model of human memory impairment thereby became available (Mishkin, 1978).

Cumulative work with the animal model over the next decade, together with neuroanatomical studies, succeeded in identifying the anatomical components of what is now termed the medial temporal lobe memory system (Squire and Zola-Morgan, 1991): the hippocampus and the adjacent perirhinal, entorhinal, and parahippocampal cortices that make up much of the parahippocampal gyrus. This information showed which structures within H.M.’s large lesion were important for understanding his impairment and, more broadly, what structures are important for memory. A few years later, an improved description of H.M.’s lesion was obtained with magnetic resonance imaging (MRI) (Corkin et al., 1997). MRI had been delayed because of concerns that clips placed on the dura during surgery made H.M. ineligible for imaging. However, thorough inquiry revealed that the dural clips constituted no risk.

At this juncture, several points became clear. First, H.M.’s lesion was less extensive than described originally by the surgeon in that it extended a little more than 5 cm caudally from the temporal pole (not 8 cm). As a result the posterior parahippocampal gyrus was largely spared (specifically, the parahippocampal cortex or what in the monkey is termed area TH TF). Second, the reason that H.M.’s memory impairment was so severe was that the bilateral damage included the parahippocampal gyrus (anteriorly) and was not restricted to the hippocampus. Damage limited to the hippocampus causes significant memory impairment but considerably less impairment than in H.M. Third, memory impairment more severe than H.M.’s could now be understood, as when the damage includes the structures damaged in H.M. but also extends far enough posteriorly to involve the parahippocampal cortex (patients E.P. and G.P.; Kirwan et al., 2008).

In the early years, the anatomy of the medial temporal lobe was poorly understood, and terms like hippocampal zone and hippocampal complex were often used to identify the area of damage. With the elucidation of the boundaries and connectivity of the structures adjacent to the hippocampus and the discovery that these structures are important for memory, vague terms like hippocampal complex became unnecessary (though one can still find them in contemporary writing). It is now possible to achieve careful descriptions based on anatomical measurement and modern terminology.

For a time, it was rather thought that motor skills were a special case and that all the rest of memory is impaired in patients like H.M. Later it became appreciated that motor skills are but a subset of a larger domain of skill-like abilities, all of which are preserved in amnesia. The demonstration of a fully preserved ability to learn the perceptual skill of mirror reading suggested a distinction between two broad classes of knowledge: declarative and procedural (Cohen and Squire, 1980). Declarative memory is what is meant when the term “memory” is used in everyday language, i.e., conscious knowledge of facts and events. Procedural memory refers to skill-based knowledge that develops gradually but with little ability to report what is being learned.

In the years that followed, other preserved learning abilities began to be reported for amnesic patients, and the perspective shifted to a framework that accommodated multiple (i.e., more than two) memory systems. As Endel Tulving wrote:

But even if we accept the broad division of memory into procedural and propositional forms ... there are phenomena that do not seem to fit readily into such a taxonomy (Tulving et al., 1982, p.336).

Subsequently, a particular interest developed in the status of autobiographical memories for unique events, which are specific to time and place, and methods were developed to assess the specificity and the detail with which such recollections could be reproduced. In the earliest efforts along these lines, as summarized by Suzanne Corkin (Corkin, 1984), H.M. produced well-formed autobiographical memories, from age 16 years or younger. It was concluded that H.M’s remote memory impairment now extended back to 11 years before his surgery. The situation seemed to change further as H.M. aged. In an update prepared nearly 20 years later (Corkin, 2002), H.M. (now 76 years old) was described as having memories of childhood, but his memories appeared more like remembered facts than like memories of specific episodes. It was also said that he could not narrate a single event that occurred at a specific time and place. Essentially the same conclusion was reached a few years later when new methods, intended to be particularly sensitive, were used to assess H.M.’s remote memory for autobiographical events (Steinvorth et al., 2005). These later findings led to the proposal that, whatever might be the case for fact memory, autobiographical memories, i.e., memories that are specific to time and place, depend on the medial temporal lobe so long as the memories persist.

There are reasons to be cautious about this idea. In 2002-2003, new MRI scans of H.M. were obtained (Salat et al., 2006). These scans documented a number of changes since his first MRI scans from 1992-1993 (Corkin et al., 1997), including cortical thinning, subcortical atrophy, large amounts of abnormal white matter, and subcortical infarcts. These findings were thought to have appeared during the past decade, and they complicate the interpretation of neuropsychological data collected during the same time period. Another consideration is that remote memories could have been intact in the early years after surgery but then have faded with time because they could not be strengthened through rehearsal and relearning. In any case, the optimal time to assess the status of past memory is soon after the onset of memory impairment.

Other work has tended to support the earlier estimates that H.M.’s remote memories were intact. First, Penfield’s two patients described above, P.B. and F.C., were reported after their surgeries to have memory loss extending back a few months and 4 years, respectively, and intact memory from before that time (Penfield and Milner, 1958). Second, methods like those used recently to assess H.M. have also been used to evaluate autobiographical memory in other patients, including patients like E.P. and G.P. who have very severe memory impairment (Kirwan et al., 2008). In these cases, autobiographical recollection was impaired when memories were drawn from the recent past but fully intact when memories were drawn from the remote past.