While the memory of a graduation party or first kiss may last a lifetime, other memories, such as an overheard phone number or person’s name, may last only seconds.

Those fleeting memories are arguably more important in our daily lives, as they allow us to track the flow of a conversation or remember directions long enough to carry them out. This type of short-term memory has been called “working memory,” because it is used to solve a problem or perform a task.

Our brains are remarkably adept at filtering what goes into working memory. Without this ability, we would become lost in a dazzling confusion of old, new and largely irrelevant information.

I am interested in how our brains store temporary memories. The prevailing hypothesis is that working memories are stored in the activities of neurons in the frontal cortex, a region of the brain behind the forehead. This makes sense, because damage to the frontal cortex from a stroke or other brain trauma results in a profound loss of ability to recall recent information.

In addition, neurons in the frontal cortex respond to specific images, actions and objects stored in memory. These “memory neurons” were discovered in the early 1970s.

Despite more than 40 years of research, many mysteries regarding how these neurons maintain memories remain unanswered. For example, scientists don’t yet know whether these neurons are influenced by body movements, by the expectation of rewards, or, like Morse code, store memories in the precise timing of their activity.

To investigate these questions, my students and I designed experiments in which rats were trained to remember sounds over short delays. We taught the rats to keep their noses inside a “nose-poke hole” for some sounds but to pull out of the hole for others. If the animal made the correct choice, it received applesauce. We recorded the activity of the rats’ brains during this experiment.

As in the 1970s experiments, we found neurons in the frontal cortex that appeared to respond only to the memory of specific sounds. However, our analysis of the movement of the rats revealed that — to our surprise — these neurons were not just responding to the memory of the sound but were also responding to unique body movements the rats adopted for each memory.

Our results suggest memory neurons are also “body neurons”: The same neurons involved in storing memory are involved in sensing movement.

We don’t yet know why this is the case, but now think the frontal cortex uses “embodied” strategies for storing memories. For example, the rats may have adopted particular movements to help them store the different sounds. Humans do something similar when rehearsing a phone number out loud or tying string around a finger to remember to pick up milk on the way home.

Our finding suggests the metaphor of memory being like a computer’s RAM is incorrect. Instead, our memory neurons appear to be far more flexible, capable of enhancing memory by using information from our bodies and perhaps other senses.