(New August 2006)
A page from: Eric Hargreaves'Page O'Neuroplasticity

No website on the neural basis of learning and memory would be complete without some mention or discussion of H.M., one of the true pioneers in the memory field. There is probably no single person, who has done as much to further our understanding of the neural basis of learning and memory then H.M., or as those who work have worked with him on occasion let slip, "Henry". So much so that Suzanne Corkin, one of those who has long worked with H.M., in her opening comments of a 2002 review article actively praises him. Henry Gustav Molaison or H.M.

"We all understand the rare opportunity we have to work with him, and we are grateful for his dedication to research. He has taught us a great deal about the cognitive and neural organization of memory. We are in his debt." (Corkin, 2002 p153).
Conversely, no individual has done as much to narrow the focus of memory research upon a few neural structures to the exclusion of all else, a narrowness that only in recent years has been broadened beyond the hippocampus and its closely allied cortices. Yet, all appearances to the contrary, H.M. is not a brilliant scholar, nor an astute clinician, but an unfortunate case of someone forever locked into the present with no recollection of the past subsequent to the late 1940s, and as such was a true child of the moment. There were many such moments, up until the morning of December 2nd 2008, when H.M. or Henry Gustav Molaison died quietly in a nursing home in Connecticut at age 82.

H.M. as a young man in the early 1950s, suffered from intractable epilepsy and as a radical treatment underwent a bilateral resection of the medial temporal lobes. Although, usually reserved for unmanageable psychotics, William Beecher Scoville, director of the Hartford Hospital's neurosurgery department, thought that this operation would potentially reduce the severity of H.M.'s seizures. Reterograde and anterograde amnesia Ultimately successful in reducing the seizures, the surgical resection however, left the young man with a reterograde amnesia that preceded the surgical lesion by a year or so. More importantly the surgery also left H.M. with a florid and profound anterograde amnesia that persists to this day some 5 decades later. The figure at left schematically depicts these two types of amnesia. Retrograde amnesia, recedes into the past from the time of the lesion, such that memories and recall for a time leading up to the lesion are gone or inaccesible. Often retrograde amnesia is graded, such that the further back one goes the more one may begin to remember until one is far enough back in the distant past that all is remembered as clearly as though there had been no incident generating the brain lesion. Retrograde amnesias, can span a time of minutes, from small concussions, to days, weeks and months in the case of H.M. Anterograde amnesia, on the other hand, proceeds subsequent to the lesion itself, is sharply delineated and in the case of H.M., devastatingly complete. Once I attended a talk by a british researcher, who shall go nameless, but he described H.M. as having a "Magnificent Deficit", and all I could think of at the time was "Bloody Brit..." (Of course they're not all like that, but for pomposity gimme someone from Cambridge or Oxford, or as they're sometimes referred to as Camford and Oxbridge). Anyway, essentially since the surgical lesion, H.M. has been unable to successfully form any new memories of the events and details making up his life and for that reason is locked into the immediate present, and the distant past of his childhood and young adulthood.

Brenda Milner Back in the day First mentioned by the surgeon William Beecher Scoville at the 1953 Harvey Cushing Society, the proceedings of which were published in 1954. H.M. was 1 (non-psychotic) of 2 cases out of 230 (unmanageable psychotic cases) that developed memory problems after removals involving the orbital area of the frontal lobes and varying degrees of the medial temporal lobe (Scoville, 1954). Independently, Brenda Milner and Wilder Penfield (1955) soon after reported two cases, where a similar impariment was found after uniliateral temporal lobectomy. In both of these latter unilateral cases it was suspected that there was also damage to the temporal lobe contralateral to the removal, resulting in a bilateral lesion. Evidence from an abnormal EEG in one case led to this conclusion (Feindel and Penfield, 1954), whereas ultimately a later autopsy confirmed the diagnosis in the remaining case (Penfield and Mathieson, 1974). Brenda Milner (originally another Brit, and portrayed above and to the right, back in the day) had been Donald O. Hebb's Ph.D. student at McGill University, whom he had passed onto Wilder Penfiled at the Montreal Neurological Institute or (MNI), which is sometimes referred to as "the house that Wilder built". Recognizing a connection when he saw one Scoville, invited Penfield and Milner to examine H.M.

The case of H.M. did not have its full impact upon the memory field until formal testing of H.M. by Brenda Milner and the resulting paper, which is still one of the most highly cited papers in Neuroscience with over 1700 citations, 300 of which have occurred in the last decade. Scoville and Milner (1957) found evidence of amnesia in 8 of the testable pyschotics, who had also received this operation. Ultimately, Scoville and Milner (1957) attributed H.M.'s severe amnesia to the removal of both hippocampi and amygdali, and so began the investigation of the hippocampus as the source for all memory and learning.

H.M.'s Precipitating epileptic aetiology

The only child of middle class parents, the father an electrician originally from Louisiana, and the mother a local girl, H.M. was born in Manchester, Connecticut on february 26, 1926 and grew up in Hartford and the surrounding towns (Hilts, 1995). There were no complications at birth and H.M. developed normally as a child until age 9 (Scoville and Milner, 1957; Corkin et al., 1997) or alternately reported as age 7 (Scoville, 1968; Corkin, 1984), when he was knocked down by a bicycle. Apparently, sometime during an afternoon he was crossing the street and was clipped by a boy riding downhill too fast to stop (Hilts, 1995). From this accident H.M. received a supra orbital laceration on his left side and lost consciousness for approximately 5 minutes (Scoville and Milner, 1957; Corkin et al., 1997; Corkin, 1984). How or when, or even if he was taken to the hospital (as opposed to the family doctor) is not present in the record, but it is known that he required 17 stitches to lace to gether the lacerations received from the incident (Hilts, 1995).

For most of the academic articles that deal with the aetiology of H.M.'s epilepsy it is supposed that the blunt force trauma sustained by H.M.'s impact with the cyclist, may have resulted in his skull being knocked backwards rapidly enough to collide with his brain. Most people presume that the brain is safely and "evenly" encased and protected by the skull, but there are parts, or lobes that are closer to specific bone ridges on the skull's internal surface, which come in contact with the skull during sudden and severe impacts. It is these parts that are most vulnerable to contusions or abrasions resulting from blunt force trauma. In particular the "anterior poles" of the temporal lobes and their "mesial" surface (underneath and closer to the middle) are commonly damaged during head on collisions. The facts that the bicycle knocked H.M. unconsciousness and gave him a cut above his left eye are indicative of a severity and location to suggest that the particularly vulnerable temporal lobe anterior poles and mesial surface may have been damaged.

Epileptic progression

At age 10 H.M. experienced his first minor seizure an "atypical petit mal", and on his 16th birthday his first "grand mal" or full seizure (Corkin, 1984). Because of the interval between the incident and the initial seizure, it is clear that the brain injuries H.M. may have sustained from the bicycle incident did not themselves directly lead to seizures. However, it is not unusual for the brain to attempt to heal itself, or at least fill in the gaps with "scar" tissue that in itself can become epileptogenic, and the focal source of subsequent seizures. Although the accident with the bicycle has been largely attributed as the precipitating event to H.M.'s seizures (Corkin et al., 1997), an additional propensity may also be accorded to the presence of epilepsy in three cousins on his father's side (Corkin, 1984). Thus, there could also be a heredity or genetic component to the development of H.M.'s seizures.

Regardless, with the increasing severity of the epileptic seizures, H.M. dropped out of high school soon after turning 16, largely due to teasing about the side effects of his seizures, which could result in urinary incontinence, tongue biting, and loss of consciousness followed by somnolence (Scoville and Milner, 1957; Hilts, 1995 ;just what every 16 year old wishes to be known for... bleeding at the mouth, pissing himself, and sitting in a stupor for long periods of time). After a two year hiatus, and under anticonvulsant medication, H.M. returned to a different high school, where he took the "practical course", instead of the "college course", and claimed to have enjoyed the Science Club, and roller skating, but avoided math because he did not care for it (Corkin, 1984). Graduating at age 21, in 1947 H.M. does remember being upset by not being allowed to march across the stage with the others, his teachers fearing a fit at the ceremony (Hilts, 1995). Before the seizures, it was expected that H.M. would become an electricion like his father, but with the increasing severity of the fits these plans ended and his father became more distant. After graduation H.M. worked on an assembly line in the underwood typewriter plant inserting pieces and later for several years held a job as a motor-winder in a small electric motor shop (someone who literally wound the copper wire on the armature of electric motors and as such, would more than likely be classified as a semi-skilled labourer).

At this time H.M. was experiencing about 10 petite mals a day and one major seizure a week (Corkin, 1984). Although living at home decisions about H.M.'s epilepsy and general care had fallen upon his mother. The father unable to cope with his son's affliction, would frequently head off to the neighbourhood bar after a major seizure, commenting that it was shameful to have "a mental" in the family (Hilts, 1995). In spite of increasingly heavy doses of various anticonvulsants like dilantin, phenobarbital, tridione, and mesantoin, H.M.'s seizures continued to increase in frequency and severity, eventually forcing him to leave his job at age 27 in 1953.

Radical Proposal: Bilateral medial temporal lobe resection

Portrait William Beecher Scoville Soon after leaving his work H.M. and his mother were referred to the neurological team at the Hartford Hospital, where William Beecher Scoville, founder and Director of the Department of Neursosurgery, depicted on the left during his prime, proposed a "frankly experimental operation" that would involve the bilateral resection of the medial temporal lobes (Scoville and Milner, 1957). Neither x-ray series nor pneumoencephalograms exhibited any abnormalities. In the latter diagnostic technique, the ventricular cerebro-spinal fluid is literally blown out of the ventricles and replaced with air to improve the contrast (and therefore clarity) of subsequent x-rays. Both simple x-rays and pneumoencephalograms were fairly crude techniques for diagnosing brain abnormalities, and have easily been surpassed by more modern computer enhanced imaging techniques, such as CTscan, PETscan, MRI, and fMRI. In addition and as a matter of course in pre-operative preparation for neurosurgery an EEG session was undertaken on August 17 of 1953, during which "a short clnical attack was said to be accompanied by generalized 2 to 3 per second spike-and-wave discharge with a slight asymmetry in the central leads" (Scoville and Milner, 1957; p.16). Thus, none of the diagnostic techniques of the time indicated any obvious neural malformation, and the EEG presented no localizing signs of epileptigenesis even during the briefly observed epileptic episode. Despite the absence of evidence indicating focal damage or abnormalities of any kind, the reasons for doing the surgery remained fairly compelling: first, the obvious disrupting force that H.M.s epileptic seizures had on his daily life, second, the anticonvulsant medications' inability to bring the seizures under control, or prevent them from progressing further, third, the probable liklihood that the seizures did have an origin in the anterior temporal lobes, and fourth, and finally, the lack of seizures in a similarly resected bi-lateral fractional lobotomy group of psychotics, in comparison to other bi-lateral orbital undercutting fractional lobotomy groups previously performed by Scoville and colleagues (Scoville et al., 1953). As such, H.M. and his mother decided to permit the surgery. A final note to the attempts to localize any epileptigenic tissue; during the actual operation after the initial retraction of the frontal lobes and exposure of the "mesial" portion of the temporal lobe, surface and depth recordings were made one last time, again without revealing any epileptogenic focus (Scoville and Milner, 1957).

Surgical Description

What exactly was removed? Well, that depends upon who you talk to... ...errr read, well OK, not really, but for the longest time only Scoville's surgical notes provided the answer as to what was removed. According to these notes, Scoville estimated that all tissue 8cm back from temporal poles and medial to the temporal horns of the lateral ventricles was removed. In the original Scoville and Milner (1957) report, the surgical removal was briefly described as including the uncus and hippocampal gyrus, and the underlying amygdaloid nucleus and "presumeably two-thirds of the hippocampal complex bilaterally" (Scoville and Milner, 1957; p.11). A more detailed description had been given earlier, when Scoville et al., (1953) first described the same operational suite as it pertained to the psychotic population for which it had originally been developed. This same description was re-itterated by Suzanne Corkin (1984) as it directly applied to H.M. In addition, she provided a succinct description of Scoville et al.'s (1953) operational procedure itself. H.M.'s surgery frontal approach

"[Scoville] approached the brain through two 1.5inch supraorbital trephine holes. By inserting a flatbrain spatula through each hole, he was able to elevate both frontal libes thereby exposing the tips of the temporal lobes. They in turn were retracted laterally in order to permit access to the medial surfaces, where electrocorticography was carried out in order to assess the activity of the uncus amygdala and hippocampus -- structures that are often implicated in epilepsy. There was no clear-cut evidence of an epileptic focus in this region. An incision was then mde that bisected the tips of the temporal lobes, and he resected the medial half of the tip of each temporal lobe. Next, Dr. Scoville removed by suction all of the gray and white matter medial to the temporal horns of the lateral ventricles, sparing the temporal neocortex almost entirely. The removal was bilateral, and is said to have extended 8cm back from the tips of the temporal lobes. It included the prepyriform gyrus, uncus, amygdala, hippocampus and parahippocampal gyrus, and must have produced an interruption of some of the white matter leading to and from the temporal lobes. H.M. was awake and talking during the operation.(Corkin, 1984 p249-250).

Much of this description is based on the visible landmarks of sulci and gyri, and uses surface terms typically employed by neurosurgeons, which makes sense when you're a neurosurgeon working almost exclusively with external landmarks and/or responses to electrical stimulation. Regardless, above and to the right is the surgical diagram which appeared in the original report, with the thatched areas indicating the tissue that was removed from H.M. (Scoville and Milner, 1957).

Many of the details Corkin (1984) discusses, such as the flatbrain spatula lifting the frontal lobe, and the aspiration device, wired with an electrocoagulation tip for cauterizing ruptured blood vessels, can be seen in the above figure. In addition, it was noted for the first time in 1984 that some of the excised tissue had undergone a gross neuropathologic study, including an examination of tissue from the uncus, and amygdalae of both sides under a microscope. All the tissue samples were found to be without inflammation or scarring (Corkin, 1984). This is an important result, since scarring was one of the proposed aeitiological mechanisms underlying H.M.'s epilepsy, as discussed above. Yet, an examination of the raw tissue beneath a microscope is not the same as a thinly sliced, and stained histological examination underneath the microscope. Therefore, once again, even though no gross abnormalities presented themselves in the tissue studied, a finer histological examination may have revealed differences in cell population densities that may have been indicative of seizure based damage.

Finally, at the time of Corkin's (1984) review, H.M. underwent for the first time one of the more modern imaging techniques, a CTscan, or contrast enhanced computed tomography performed by a GE 9800 (yes, thats General Electric, same as your light bulb). The CTscan results did not obviously contradict what had now been accepted for some 30years, based on Scoville's 1953 surgical notes (Corkin, 1984).

H.M.'s MRI

Leap forward to the early 90s, specifically May of 1992 and Aug of 1993, at 66 and 67 years of age, H.M. finally underwent the gold standard of imaging technques the MRI. Since it had been commercially available since 1980 (albeit not to the same degree of resolution), what took so long? Well, often during neurosurgery, metal vascular clips are left intentionally inside the wound to stablize blood flow. Scoville had placed three of these clips on the dura at the margin of the wound to coagulate a number of veins. Thus, there was metal in H.M.'s head. Metal and magnets do not go well together in this case. In the worse case scenario the clips could actually have become dislodged and spun around inside H.M.'s head. Another bad thing that could have happened is that these clips may have been vibrated at frequencies fast enough to heat them up (imagine taking a lighter to a paper clip and then popping it in your mouth). Regardless, "Venous structures thrombose after a number of days" (Corkin et al., 1997; p3975), meaning the clipped veins would have collapsed and fused shut after only a few days, unlike clipped arteries, which could tear and start an intracranial bleed. Also, the surgery having been performed some 40 years earlier the clips would be also be overgrown with fibrous material securing them in place. Further, since they were at the margin of the wound, and therefore not central, they also would not be in close contact with the brain, nor were they near any of the identified arteries on the earlier CTscan. After some detective work tracing down the exact make of the clips, and contacting the company and consulting their records Corkin et al., (1997) were able to determine that the clips were not ferromagnetic, being made out either silver or tantalum in 1953. Based on the materials they also were able to estimate that the clips may only heat up to the temperature of the brain during a "febrile state" (catch phrase meaning... heat up to the level of a fever, from 98.6F to maybe 102-103F? The feel of a warm coffee mug on your lips, before the coffee hits). Finally, they consulted an expert in the field, Dr. D. Piegrass, "Chair of the Task Force on Clips and Clip Appliers, a subcommittee of the Joint Committee on Devices and Drugs of the American Association of Neurological Surgeons and the Congress of Neurological Surgeons" (Dontcha' love the medical field and its beurocracy?). Anyway, it was a long and cautious process to verify that it was relatively safe for H.M. to be scanned in the MR, which in all liklihood was probably a good thing.

Suzanne Corkin My small amount of contact with this episode of the H.M. saga happens here. The scans were performed in 1992 and 1993, but the data were not fully published until 1997 (Corkin et al., 1997). During the interrim Dr. Corkin did what any other self respecting scientist with hot data would do... ..she went on tour with it. Well OK, not really like a full concert tour of your favourite rock band, but once it was known that the data from H.M.'s scans were out there, Suzanne Corkin got invited to give lot of talks, one of which naturally had to be at the MNI. I say naturally had to be at the MNI, because as discussed earlier this was where the detailed examination and study of H.M. was originated by Brenda Milner, which truly triggered all that is H.M. (Scoville and Milner, 1957). I happened to be doing my first postdoc at McGill University during this time, and wouldn't have missed this talk for anything. So on Novemeber 3rd 1994, I got myself there bright and early, took copious quantities of notes, and immediately thereafter transcribed them to a word processor (if you saw my handwriting, you know why this would be critical). Afterwards, did send these notes around to a few interested friends (actually a lot of interested friends) and since that time have come across smatterings of my transcriptions in the oddest places on the Web, and in numerous instructors' class notes. Of course once the article came out in 1997 it immediately surpassed anything I had transcribed, so any fragments of my notes have long since vanished. However, I still have a few "nuggets" from the talk that were never intended to be published, and it was of course here that I learned H.M.'s full name, but not because of any slip from Suzanne Corkin herself.

The MRI itself was performed by a 1.5Tesla GE Signa System (Yup, its General Electric again). T1 weighted coronal and sagittal series were obtained, with the scan details as follows [coronal: repetition time(TR)=550; echo time(TE)=16; 5mm slice thickness, 1mm interval between slices, (256X192) matrix; Number of excitations(NEX)=1.0; Field of View(FOV)=22; sagittal: TR=600; TE=19; 4mm thickness 1mm interval (256X192)matrix; NE=+1.0; FOV=22].

MRI Results

So after all this preamble, what did the MRI results reveal that were not already in Scoville's notes? Well for such a long build up the original description of the removal was not that inaccurate. However, the rostro-caudal extent of the removal (the front-to-back amount taken out) was grossly exaggerated. For some unknown reason, neurosurgeons tend to overestimate the removal in their reports, while at the same time being quite conservative in the actual removals. Scoville was no exception, and his originally reported 8cm rostro-caudal removal was revealed by the MR imaging as really only a 5cm removal. HM lesion diagram old and MRI base The 3cm difference is fairly substantial, a 38% change in amount, and at this scale a fairly visible difference, being a little more than the top joint of my index finger. At the same time, 3cm is still only 3cm, and the difference is the extra tissue was left in, and not the extra tissue that was taken out. Regardless, according to H.M.'s MRI, had Scoville actually removed the amount he reported, much of the optic radiations and tisse surrounding the calcarine fissure would also have been destroyed, and much of H.M.'s sight along with it. Since to this day H.M. still has no visual defects, it is safe to say these regions were spared. The figure at right is composed of two columns of drawings. The left column is Lamar Roberts M.D. 1957 artist's rendering of H.M.'s removal (note the small signature in the righthand corner of the bottom coronal section), which appeared as figure 2 in the original report (Scoville and Milner, 1957). The right column is the same figure altered to reflect the 1992 MRI based version of the removal. Top row of the figure are ventral views of the brain, with an intact hemisphere on the right and the mesially resected temporal lobe on the left. The full rostrocaudal extent of the removal is delineated by the drawn arrows and the dashed lines labelled A-D indicate the level at which the coronal sections below were taken. The coronal sections, as above, show the left hemisphere as respresenting the extent of the removal, and the right hemisphere showing the intact structures had they been left in place. Comparing across columns at the same level, one can see that at section A) the initial estimate and approach match the information derived from the MRI images. More caudal at section B)the removal of the surrounding neocortex is shown to be more extensive in the original estimate, and not as great in the updated version as indicated by the presence of the collateral sulcus. At the level of section C) the original estimate still indicates a full removal of the hippocampal tissue, while the updated version, shows that only minor damage was done to the hippocampus at this level. The final section D) again shows that the original estimate of the removal was quite extensive, while the MRI based version, now shows that the left hemisphere is as intact as the right hemisphere.

HM's Composite MRI Of course, whats the point of looking at artist's (or neurosurgeons, as was Lamar Roberts) depictions of missing brain parts, when you've got the actual MR imges available? To the left is a similar composite figure as before, the top row being the identical ventral view, with the left hemisphere depicting the rostrocaudal extent of H.M.'s MRI based resection. Again the drawn arrows and the dashed lines labelled A-C on the ventral view indicate the level at which the coronal sections below were taken. Of the coronal sections indicated, the left hemisphere is still the updated artist's rendering, but now over the right hemsphere of each coronal section I have superimposed the matching half of the MR image taken at the appropriate level.

Although some graying of the resected area appears to indicate that sparse material may still be present at the levels of A and B, one can clearly see the darkened region of the lesion sequeing into the lateral ventricles (which was one of the surgical margins used to limit the resection). At the level of B) one can clearly identify the collateral sulcus and the surrounding neocortical tissue that is present on both banks. Thus, unlike that which was originally estimated and thought for a number of decades, the margin of the resection had an oblique ventromedial orientation to it, sparing much of the neocortex wrapped around the collateral sulcus. This intact tissue becomes important later, when discussing Horel's temporal stem theory of H.M.'s amnesic syndrome, which also created a big "kafuffle", in the experimental monkey literature, in which attempts to recreate H.M.-like deficits were made. In contrast to the artistic rendering the MRIs appear to indicate a degree of enlargment of the ventricles. This was also noted with the earlier CTscan in 1984, but as was commented upon then, the enlargement is not out of range for H.M.'s age, as are the widening of the cortical sulci and concommittent shrinking of the gyri (Corkin, 1984).
Since the original 8cm estimate of the temporal lobe resection was reduced to 5cm based on the MRI evidence, what does that mean for the original Scoville comments suggesting that 2/3 of the hippocampus was removed? Sagittal sections best reveal the extent of the hippocampus removal, which was surprsingly symmetrical. The overall rostrocaudal removal extended 5.4cm on the left side, while extending 5.1cm on the right side. HM's left sagittal MRI
Above is one of the sagittal sections depicting the extent of the left hippocampal removal, marked by the asterix, as well as the remaining extent of the hippocampal residual marked by the arrow. Based on the MRI evidence 1.9-2.2cm of the hippocampus tissue remained visible. The normal rostrocaudal extent of the hippocampus in humans is approximately 4cm (Amaral and Insausti, 1990). As such, almost 50% instead of the originally estimated 33.3% remained. However, Corkin et al., (1997) note that what did remain appeared to somewhat atrophied (literally withered or wasted away), and therefore the "functional integrity could not be assured". So how much of H.M.'s intact hippocampus is fully functional remains at present still unanswered, and due to the atrophy, which could have any number of causes, including the original epilepsy for which H.M. was treated, Scoville's estimate may yet still stand. Finally, as was to be expected the fimbria and fornix, the efferent tracts of white matter accumulating the major output of the hippocampus, also appeared to be "somewhat thin" (Corkin et al., 1997).

Apart from the specific damage to the hippocampal system, the strong degree of cerebellar atrophy is also plainly visible in the sagittal section. There is a marked and "diffuse atrophy of the vermis and [cerebellar] hemispheres", evidenced by enlarged subarachnoid space surrounding the cerebellum (Corkin et al., 1997). The cerebellar atrophy is a side effect of the neuroleptic dilantin, which H.M. took at relatively high doses for a number of years prior to the removal, and for almost another 30 years after the removal, albeit at a much reduced dose. In 1984, H.M.'s Dilantin medication was replaced with Tegretol another neuroleptic. Regardless, H.M. was exhibiting neurological signs of dilantin toxicity as early as 1962 (Corkin et al., 1997). Outside these two areas, damage was also noted in the dorslateral region of the frontal lobes as "prominent sulcal spaces", located deep to the location of the trephine holes. As with the hippocampal atrophy, whether this damage occurred at the time of the surgery, or developed slowly in the years afterwards is unknown. However, there was no apparent damage to the orbital or ventromedial portion of the frontal lobes, which were lifted up and to the side to reveal the temporal lobes during the surgery. The mammillary bodies, were also noted to be "reduced in size" (Corkin et al., 1997), but given the connectivity of the fimbria fornix to the mammillary bodies, this result might be expected. Finally, there was no apparent damage to the mediodorsal nucleus of the thalamus, another region that has been implicated in memory and amnesic disorders, such as Korskoff's syndrome, and not unlike H.M. in the single unfortunate case of N.A., a young airforce radar technician who wound up with a fencing foil up his nose during some horseplay in his barracks (Teuber et al., 1968).

Anatomical Summary of H.M.'s Temporal Lobe Removal

The MRI results combined with Scoville's notes gives a much better idea of what was removed and can be described in anatomical terms that go beyond the surface landmarks commonly used by neurosurgeons. The finer and more modern anatomical detail that follows in the next section is based upon meticuluous histological differences in the cortical cell layers, and may not be readily apparent at the gross level. Brenda Milner 2005

(older terminology meaning "hook", and referring to anterior portion of parahippocampal gyrus), and portions of the hippocampal gyrus (older term referring to posterior portion of parahippocampal gyrus). (Milner, 1972; Corkin, 1984).

Hippocampus and memory

The now famous case of H.M. has had a major influence in the field of memory, focusing much of the work since, on the structure and function of the hippocampus. Since that time, the amount or research that has gone into determining the role of the hippocampus in learning and memory processes, if any at all, has filled many books (The Hippocampus as a Cognitive Map, 1978; The Hippocampus Vols. I & II 1975; III & IV 1986; The Neurobiology of The Hippocampus, 1983 The Hippocampus - New Vistas, 1989 Functions of the Septo-Hippocampal System, 1978), and countless articles, and continues to do so...

At first, the hippocampus was promoted as an all encompassing learning structure. However, it soon became clear that H.M.'s motor learning ability appeared to be intact, although he had no recollection of ever performing the tasks when tested in repeated sessions (Corkin, 1965; 1968; Cohen and Corkin, 1981). The cerebellum, a brain structure involved in the smoothing of motor movements was, at that time, being argued as the critical learning centre for classically conditioned motor responses (for a review at the time see Thompson, et al., 1983). Thus, the cerebellum became the potential source of H.M.'s motor learning ability, and relegated the hippocampus to one of at least two learning structures. Of course, the hippocampus still held preminance as the structure responsible for conscious learning, an idea was reinforced by the report of a stroke case, in which damage restricted to the CA1 subfield of the hippocampus, resulted in amnesia (Zola-Morgan, Squire, and Amaral, 1986; but see Horel, 1994).

However other clinical cases, in which portions of the hippocampus were destroyed for relief from intractable pain (Gol and Faibish, 1967), or destroyed to reduce seizures (Glaser, 1980) did not necessarily result in amnesia, and a growing body of evidence from animal research indicated that lesions restricted to the hippocampus were unable to produce the global, florid amnesia expressed by H.M. As a result, much paring down has occurred during the last few years regarding the role of the hippocampus in learning and memory processes (Zola-Morgan et al., 1989a; 1989b; Meunier et al., 1993).

Currently, it appears that the hippocampus is part of a larger learning system or systems, and with hindsight it is more than likely that the severe, global amnesia of H.M. is caused by the bilateral removal of the hippocampus, amygdala, in concert with additional damage to the cortical areas overlying the hippocampus (Horel, 1994; Petri and Mishkin, 1994). Consequently, ideas of the hippocampus have evolved to become ideas about the hippocampal system, which anatomically, now encompasses the overlying cortices feeding into the hippocampus (entorhinal, perirhinal, and parahippocampal), and encompasses, as hippocampal projection sites, subicular, infralimbic and other prefrontal cortical areas (Petri and Mishkin, 1994 Horel, 1994; Cohen and Eichenbaum, 1993). How much of the front end, and how much of the back end is included in the hippocampal system depends upon whose conception of the hippocampal system is being discussed, with some not wishing to label it as the hippocampal system at all (Horel, 1994).

Additionally, the hippocampus may work in parallel, interact, or compete with other forebrain learning systems (Packard et al., 1989; McDonald and White, 1995a; 1995b; White and McDonald, 1993a; 1993b). Further, ideas about the hippocampus itself are also becoming internally more modular, or heterogenous, such that the different regions or subfields may participate in different computational functions (Cohen and Eichenbaum, 1993). Finally, the whole idea of the hippocampus' involvement in learning and memory has been called into question (Vanderwolf and Cain, 1994). Yet, despite the ongoing debate, most researchers believe that the hippocampus, or hippocampal system has some role to play in learning and memory processes.

H.M. References...

Amaral, D.G. and Insausti,R. (1990). The hippocampal formation. In: The Human Nervous System (Paxinos G., ed) p711-755. San Diego: Academic Press.

Cohen N.J. and Corkin, S. (1981). The amnesic patient H.M.: learning and retention of a cognitive skill. Neuroscience Abstracts, 7, 235.

Cohen, N.J. and Eichenbaum, H. (1993). Memory, Amnesia, and the Hippocampal System. MIT Press, Cambridge.

Corkin S. (1965). Tactually-guided maze learning in man: effects on unilateral cortical excisions and bilateral hippocampal lesions. Neuropsychologia, 3, 339-351.

Corkin S. (1968). Acquisition of motor skill after bilateral medial temporal-lobe excision. Neuropsychologia, 6, 255-265.

Corkin, S. (1984). Lasting consequences of bilateral medial temporal lobectomy: clinical course and experimental findings in H.M. Seminars in Neurology, 4, 249-259.

Corkin, S. (2002). Whats new with amnesic patient H.M.? Nature reviews Neuroscience, 3, 153-160.

Corkin, S., Amaral, D.G., Gonzalez, Johnson,K.A., Hyman, B.T.(1997). H.M.'s Medial temporal lobe lesion: findings from magnetic resonance imaging. The Journal of Neuroscience, 17, 3964-3979.

Feindel W, and Penfield, W. (1954). Localization of discharge in temporal lobe automatism. Archives of Neurological Psychiatry, 72, 605-630.

Glaser G.H. (1980) Treatment of intractable temporal lobe-limbic epilepsy by temporal lobectomy. Annals of Neurology 455-459.

Gol A.M, Faibish G.M. (1967) Effects of human hippocampal ablation. J. Neurosurgery 26, 390-398.

Hilts, P.J. (1995) Memory's Ghost: The Strange Tale of Mr. M and the Nature of Memory, Simon and Schuster, New York.

Horel, J. (1978). The neuroanatomy of amnesia. a critique of the hippocampal memory hypothesis. Brain, 101, 403-445.

Horel J. (1994) Some comments on the special cognitive function claimed for the hippocampus. Cortex 30, 269-280.

R.L. Isaacson & K.H. Pribram (Eds.) The Hippocampus Vols. I 1975, II, III, 1988 IV

Milner, B., and W. Penfiueld (1955). The effects of hippocampal lesions on recent memory. Transactions of the American Neurological Association, 80, 42-48.

Milner, B. (1972). Disorders of learning and memory after temporal lobe lesions in man. Clinical Neurosurgery, 19, 421-446.

Penfield, W. and Mathieson G.M. (1974). Memory: autopsy findings and comments on the role of the hippocampus in experiential recall. Archives of Neurology, 31, 145-154.

Scoville, W.B., Dunsmore, W.T., Liberson, W.T, Henry, C.E. and Pepe A. (1951). Observations on medial temporal lobotomy and uncotomy in the treatment of psychotic states. Research Publications-Association for Reseach in the Nervous and Mental Diseases, 31, 347-369.

Scoville, W.B. (1954). The limbic lobe in man. Journal of Neurosurgery, 11, 64-66.

Scoville, W.B. & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20, 11-21.

Scoville, W.B. (1968). Amnesia after bilateral mesial temporal-lobe excision: introduction to case H.M. Neuropsychologica, 6, 211-213.

Teuber, H.L., Milner, B., and Vaughan, H.G. Jr. (1968). Persistent anterograde amnesia after stab wound of the basal brain. Neuropsychologia, 6, 267-282.

Zola-Morgan, S., Squire, L.R. and Mishkin, M. (1982). Neuroanatomy of amnesia: amygdala-hippocampus versus temporal stem. Science, 218, 1337-1339.

Zola-Morgan, S. and Squire, L.R. (1985). Medial temporal lesions in monkeys impair memory in a variety of tasks sensitive to human amnesia. Behavioral Neuroscience, 100, 165-170.

Zola-Morgan, S., Squire, L.R. and Amaral, D.G. (1986). Human amnesia in the medial temporal region: enduring memory impairment following a bilateral lesion limited to field CA1 of the hippocampus. Journal of Neuroscience, 6, 2950-2967.

Zola-Morgan et al., 1989a; 1989b Zola-Morgan, S., Squire, L.R. and Amaral, D.G. (1989a). Lesions of amygdala that spare the adjacent cortical regions do not impair memory or exacerbate the impairment following lesions of the hippocampal formation. Journal of Neuroscience, 9, 1922-1936.

Zola-Morgan, S., Squire, L.R., Amaral, D.G. and Suzuki, W. (1989b). Lesions of perirhinal and parahippocampal cortex that spare the amygdala and hippocampal formation produce severe memory impairment. Journal of Neuroscience, 9, 4355-4370.