Hippocampus has its cake and eats it too

The ability to recall our experiences as they evolved over time is truly an impressive feat accomplished in large part through the working of a thumb-sized portion of the brain called the hippocampus.  How the brain encodes memories is a difficult, but exciting and burgeoning area of neuroscientific research.  Several distinct types of memory are known to exist, each of which plays a necessary role, integrating with other forms of memory and learning to form cognition.  One type, episodic memory, refers to the ability to recall our experiences in the world and the order with which they take place.  A simple trip to the grocery store is an episode packed with images, events (e.g., an encounter with a friend), and actions (e.g., a left turn at an intersection). 

Yet, memories for different experiences are valuable not only in how they differ from other experiences, but also in the details they share.  Getting dressed and leaving the house to attend a wedding is a very different ‘episode’ from getting dressed and leaving the house to attend a football game, yet the paths taken through the house, the location of the car keys, and the roads taken to reach the highway may be exactly the same. So, then, does the hippocampus take account of the features common to two otherwise very different episodes?  In a word, yes.  Actually, our most recent work suggests that the hippocampus is at least as interested in what different episodes have in common as in making sure they are remembered as different. 

We consider the content of episodic memories by examining the activity of single hippocampal neurons in rats taught to make traversals of pathways through an environment.  As navigational skills go, rats are geniuses and remember path-running episodes through careful attention to the positions encountered during a run along a path.  This makes a great deal of sense as the different experiences which together compose an episode are usually associated with particular places in the world. 

The accounting of positions encountered takes the form of what are called ‘place fields’, a classic example of which is shown in the rightmost panel of the accompanying figure.  The figure depicts the activity rate of a single hippocampal neuron (among hundreds of thousands) as the rat ran along a squared spiral path like the one schematized at the left of the figure.  Activity rates are given along a color scale where brown corresponds to low rates (near zero) and high rates are color-mapped as yellow and white.  This neuron has the remarkable property of firing over only one distinct portion of the full track.  Neighboring neurons, not shown, also exhibit place fields, but across different, similarly-sized portions of the track.  In this way, each position on the track and the sensory experience associated with it is mapped in a different way by the hippocampus.  By generating a record, or memory, of the sequence of such mappings, the hippocampus creates episodic memories.

However, most of the hippocampal neurons we examined, fully 75%, exhibited not single firing fields, but rather several firing fields.  Firing ratemaps for neurons of this type are shown in the middle two panels.  There were two types of multi-field neurons which, within the laboratory, we call ‘stretchers’ and ‘corners’.   Across the five loops that make up the squared spiral track, both stretchers and corners fire across several sections that are analogous with respect to which way the animal is heading and what type of behavior he exhibits.  The stretcher neuron on the left fires whenever the animal runs straight and northward.   Its firing covers all those positions where what the animal does and what the animal sees are very similar.  As the animal reaches the northwest corner of the track, the corner neuron takes over.  Turning right across the northwest corner is an experience repeated five times during the overall path-running episode. 

In this way, the hippocampus seems to have its cake and eat it too.  Stretcher and corner neurons appear to be responsible for extracting features common to two sets of experiences.  Place-field neurons, surprisingly in the minority, are poised to help us disambiguate two experiences which are largely the same.  In future work, we hope to determine just what features of experience can cause the hippocampus to be biased toward seeing two experiences as the same versus different.  Perhaps then we’ll understand why all weddings are different to some people and all the same to others.