Tracing Permanent Memories

by Maria Schamis Turner

February 17, 2010

The story of memory research has shown that memory is complex, involving different brain mechanisms and structures that, while connected to one another, also function separately, preserving some types of memory when others are impaired. In addition, certain brain structures, such as the hippocampus, have been shown to serve multiple roles in memory storage and retrieval, researcher Brenda Milner and other speakers reported at a recent symposium in Montreal. 

Milner, a neuropsychologist at the Montreal Neurological Institute (MNI), is widely recognized for her studies with the famed patient known as H.M. Along with neuroscientist Morris Moscovitch and institute researcher Wayne Sossin, she presented a history of memory findings and set them in context during the session “Memory & Memories,” part of the institute’s 75th anniversary symposium.

From Wilder Penfield to H.M.

Milner began her research on memory in the 1950s while working with MNI’s Wilder Penfield. Penfield was performing unilateral temporal lobe surgery, which involves removing a portion of the temporal lobe of the right or left hemisphere of the brain to ease the symptoms of epilepsy. Initially interested in studying visual perception in these patients, Milner switched her focus after several began to complain of post-operative memory troubles.

“We found effects that you’re all familiar with now: defects in verbal memory from removals from the left hemisphere, anterior to the speech area,” Milner explained. “Patients with right temporal lesions had the sorts of deficits that were predictable from what was known in monkeys: They had deficits in memory for faces, places and tunes.” Still, “this was a good exchange for the control of epilepsy,” she added. “Remember that these are elective surgeries, to improve quality of life.”

The operation did not work for all patients, however, and Penfield started to remove more tissue, extending into the anterior and medial structures of the temporal lobe, including the hippocampus. After one such surgery, a patient known as P.B. suffered a devastating impairment to his memory.

“He had a profound anterograde amnesia,” Milner said. “He was forgetting his life as he lived it. As his attention was diverted, and life was always diverting it, he’d forget what happened before.” After another such patient reported this condition, Milner, Penfield and their colleague Herbert Jasper began to speculate about what might be happening. They postulated that these patients had atrophy in the hippocampal region in the right hemisphere, left untouched by surgery, so when the left hippocampus was removed, they were completely deprived of hippocampal function.

“Now why did we say hippocampus?” asked Milner. “We said hippocampus because it was only after the second operation in P.B.’s case that this memory impairment was seen. We never meant to imply that it was only the hippocampus, but we were sure that the hippocampus had a role to play in it.” When P.B. died 12 years later, results from pathology exams confirmed that he had more atrophy in the hippocampal region on the right side than damage from the surgery on the left.

After reporting these findings, Milner was invited to Hartford to study  H.M. (Henry Molaison), who had undergone a bilateral temporal lobectomy that included removal of major portions of the hippocampus. Milner discovered that like P.B., H.M. also suffered severe anterograde amnesia, adding to the evidence that the hippocampus was crucial in the formation of memories.

“H.M. never learned to recognize me,” she said. “As long as he was rehearsing something, he could retain it, but as soon as his memory shifted, what went before was lost.”

To further understand H.M.’s memory impairment, Milner decided to see what she could teach him. She spent three days training him to reproduce a drawing of a star by looking at it in a mirror. As with a normal subject, H.M.’s performance improved over the three days--with one major difference. “He had absolutely no memory of all these trials he had been through,” Milner said. “There was a total dissociation between his experience and his excellent performance.”

This was evidence that there was another memory system in the brain. “I postulated then that motor learning was quite independent of the hippocampal system,” Milner said. This early demonstration of multiple memory systems in the brain was “very, very exciting for me.”

The many traces of memory

The idea that memories become more stable over time, called memory consolidation, has a long history. Since the 1950s, the standard model of consolidation proposes that memories are initially dependent on the hippocampus but with time are transferred to other neocortical structures, where they are represented in long-term memory. Once the period of consolidation is complete, these memories can be retrieved directly from other structures without using the hippocampus.

This would suggest that patients with retrograde amnesia (an inability to recall events before the amnesia occurred) would show severe memory loss for information that was acquired around the time of the event that caused the amnesia, as these memories would not have had time to fully consolidate.

But when Morris Moscovitch studied the literature, he discovered that in some cases retrograde amnesia seemed to extend back three decades or more.

“I thought, my God, nobody who thought about systems consolidation really intended the consolidation period to last over three decades,” said Moscovitch, a professor of psychology at the University of Toronto and senior scientist at the Rotman Research Institute. Moscovitch and his colleagues then performed a study where they divided conscious memories into episodic or semantic memories. “Episodic memory involves recollection,” Moscovitch explained. “It’s a memory where we re-experience or relive a past event in the mind, and it’s characterized by recovering the context in which the memory occurred. That’s different from semantic memory, where we know facts about the world and about ourselves, but there’s little or no information about the context in which the knowledge was acquired.”

Moscovitch and his colleagues discovered that for people with retrograde amnesia, episodic memories were much more likely to be impaired. The amnesia patients differed little from the control group in their recall of semantic memory, but for episodic memories, their retrograde amnesia extended all the way to childhood.

The researchers then did a functional neuroimaging study in people with normal memories using family photos to probe memory. They found that as normal people recalled vivid episodic memories, either recent or long since past, the hippocampus was activated. “The hippocampus seems to be active specifically when re-experiencing,” Moscovitch said.

These results led him and his colleagues to propose an alternative hypothesis to the standard consolidation model, the multiple trace theory, suggesting that memories are stored jointly in the hippocampus and other brain structures. “The hippocampus is always needed to recover [episodic memories],” Moscovitch said. “Over time, some episodic memories are transformed into semantic memories, which reside in the neocortical structures and the hippocampus is not needed to recall them.”

Seeking the traces

Wayne Sossin closed the session by talking briefly about current memory research at the synaptic level.

“One of the things that my research has shown is that there are multiple molecular traces in the brain when memories are made,” he said. In June he and collaborators from the institute along with researchers from McGill University and the University of California, Los Angeles, captured an image of protein translation, the production of new proteins locally at a synapse, during memory formation. Researchers used a “translational reporter,” a fluorescent protein that can be easily traced, in a laboratory culture of the sensory-motor synapses of the marine mollusk Aplysia. They discovered increased protein synthesis in response to stimuli known to induce changes in behavior of the animal. Sossin and his colleagues concluded that their results “provide direct evidence that local translation occurs at synapses in response to stimuli that induce transcription-dependent, learning related synaptic plasticity.” This was the first time that this mechanism, which is one of the mechanisms thought to underlie long-term memory formation, has been recorded. These findings were published in the June 19 Science.

Sossin was optimistic about the future of memory research, both on the synaptic level and the neuropsychological level.

“My hope,” he said, “is that we will eventually fit the two together in some important way, where we can find the biochemical and molecular changes that serve different forms of memory.”