From a contemporary perspective, there are at least two reasons why the early studies with experimental animals did not succeed. First, beginning in about 1980, it became clear that memory is not a single entity and that only one kind of memory is impaired following damage to the medial temporal lobe. The kind of memory impaired in H.M. and other amnesic patients is termed declarative memory. Declarative memory supports our capacity to recollect facts and events and can be contrasted with a collection of nondeclarative memory abilities, including the capacity for acquiring habits and skills, simple forms of conditioning, and other abilities that allow us to change through experience how we interact with the world. The discovery of multiple memory systems in the brain meant that, if one’s goal is to reproduce in an animal the memory impairment exhibited by H.M., only certain behavioral tasks will be appropriate for detecting memory impairment.

The second reason that early studies did not achieve an animal model of amnesia is that experimental animals can use nondeclarative strategies to learn many tasks (especially tasks that are learned gradually over many trials), even when the same tasks are learned declaratively by humans and are performed poorly by amnesic patients.

As it turned out, the most appropriate tasks for studying memory impairment in animals were developed first for the monkey. These were tasks of one-trial learning that assessed memory at some interval after the presentation of a single learning trial. Accordingly, in the early 1980s the first animal models of human amnesia were developed in the monkey. With the animal model in place, it became possible through a program of systematic, cumulative experiments to identify the structures that when damaged produce memory impairment. This series of studies came to completion in 1991 with identification of the hippocampus (including the dentate gyrus and subicular complex) and the adjacent perirhinal, entorhinal, and parahippocampal cortices as the important structures comprising the medial temporal lobe memory system.

So it developed that by about 1990, new information from humans, monkeys, and rats made it possible to suppose that the three major species involved in memory research (at the level of brain systems and behavior) were telling a consistent story about how the brain has organized its memory functions and in particular about the role of the hippocampus and adjacent cortex. With this idea in mind, in 1992 I wrote a review article that tried to tell this story. The article was in three parts. The first part considered the evidence that the hippocampus itself was important for memory. The second part discussed the concept of multiple memory systems, showing how the characteristics of these systems could help make sense of the pattern of sparing and loss observed after hippocampal lesions. The third part discussed the phenomenon of retrograde amnesia (the loss of memory that was acquired before the onset of amnesia), again illustrating the continuity of findings across species. This article, then, did not propose a new theory or describe a new finding. Rather, if the article has been useful, it is because it brought together several different threads of empirical work, at a time when some of the issues were in a certain amount of disarray, and it argued for a common way to understand a large amount of data.

Since 1992, the study of memory has moved forward in a number of interesting and exciting ways. Most notable perhaps is the possibility of studying the cellular and molecular basis of synaptic plasticity and memory in simple organisms like Aplysia and Drosophila. Work on memory with these organisms began in the 1960s and 1970s and has exploded in the 1990s with the development of techniques for manipulating single genes and studying their effects on learning and memory. Such studies are now being carried out in the mouse as well, with methods that restrict gene expression to specific brain regions and that allow gene expression to be turned on and off.

At the level of brain systems, an enormous amount is also being accomplished. Using the new techniques of functional neuroimaging, single-unit recording with multi-electrode arrays, manipulations of gene expression—in conjunction with traditional methods—work is proceeding to characterize how the different structures within the medial temporal lobe contribute to memory, where memory is stored, and how the several memory systems of the brain operate to record and retrieve the effects of experience. The rat and the mouse will be especially important in these efforts.

Dr. Larry R. Squire
University of California, San Diego
School of Medicine
San Diego, CA, USA