In 350 B.C., Aristotle, in his seminal treatise on the subject, described it as having two forms: familiarity and recollection. Of these, he considered recollection to be a purely human condition. That tenet is now being challenged by researchers at Boston University.

Neurobiologists at Boston University's Center for Memory and Brain have provided the first evidence that rats use recollection when recognizing items they have recently experienced. In addition, the researchers show that rodents' capacity for recollection-like memory retrieval depends on the brain structure known as the hippocampus, the same structure believed to be involved in recollection in humans. Their findings are published in the September 9 issue of the journal Nature.

Although neuroimaging studies of hippocampal activity in normal individuals as well as studies of amnesia indicate the hippocampus could be crucial to recollection, definitive methods for assessing hippocampal activity in memory have largely remained out of reach.

The BU research team, led by Norbert Fortin, a research associate in the Laboratory of Cognitive Neurobiology at the Center for Memory and Brain, and including Howard Eichenbaum, Center director and professor and chairman of BU's Department of Psychology, and Sean Wright, a former BU undergraduate, set out to better define the role of the hippocampus in the human recollection process. They approached this in a novel way -- by investigating the activity of the hippocampus in the rat brain. Their approach also meant that they had to think outside the conventions of the discipline and ask, "Do rats have a capacity for recollection?"

In humans, signal detection techniques have been used to distinguish memory responses triggered by familiarity, the general sense that a person or thing has been previously perceived, from those triggered by recollection, the sudden rush of detailed memory. The BU team chose to determine whether analyses related to this technique, known as receiver operating characteristic (ROC) curves, could be used to assess rat memory processes.

With familiarity, previous encounters are not recalled. For instance, if you encounter someone you know you've met, your sense of "knowing" that person may be weak or strong, but it does not include details of previous encounters. ROC analyses show that familiarity manifests itself as a continuous function reflecting the strength of a perceptual impression in memory.

Recollection, by comparison, is expressed in an all-or-none fashion, triggered when a certain threshold of associative and contextual information has been achieved. It is a sudden, overwhelming rush of detailed memory, such as that which is experienced when you eventually recall the prior encounter with that familiar person.

The researchers devised a memory test that capitalized on rats' highly developed sense of smell as well as instinctual foraging behavior. Initially, the rats explored a "list" of 10 common household odors (for example, cinnamon, oregano, coffee, chocolate), each mixed in ordinary sand that hid a buried food reward.

Following a 30-minute retention period, they were presented with a series of 20 odors, the 10 "old" odors plus 10 "new" ones. In addition, the animal's decisions were intentionally biased by varying the difficulty of responding to each odor and by varying the food "payoffs" for correct identification of old and new odors.

The test design allowed the researchers to measure the ratio of correct and false odor identifications across a range of bias conditions, generating the ROC curve. The shape of this curve indicated the existence of both an all-or-none threshold component and a continuous-strength component. The curves produced are very similar to those observed in the ROCs of humans, indicating the existence of recollection and familiarity in support of recognition performance.

The researchers divided the rats into two comparable groups: a group in which the hippocampus was removed and a control group. When the rats were again tested, the results were compelling -- recollection was lost in rats without a hippocampus but familiarity remained intact. A further test of the controls showed that, in the normal forgetting process, familiarity fades quickly while recollection persists, precisely the opposite pattern observed in the animals without a hippocampus.

This combination of results provides strong evidence that animals, contrary to Aristotle's contention, exhibit recollection as well as familiarity. Furthermore, the findings point to the hippocampus as critical to the process of recollection.

Bringing together faculty with expertise in psychology, neuroscience, and biomedical engineering, BU's Center for Memory and Brain provides a focal point for research on how the human brain stores and retrieves information. Boston University, with an enrollment of more than 29,000 in its 17 schools and colleges, is the fourth-largest independent university in the nation.