Episodic memory, or memory for autobiographical experiences, has the remarkable feature that each remembered experience happened exactly once, and yet exists in memory for decades. How does the brain capture fleeting experiences and transform them into patterns of neural activity that are stable over our lives?
The hippocampus has long been recognized as a neural hub for memory formation. Synapses in this part of the brain are highly plastic—that is, capable of rapidly turning their connection strength up or down to reconfigure the network of neurons that are connected to each other. Many neurons in the hippocampus also have the intriguing property of becoming active when the animal is at a specific location in its environment, and are hence known as “place cells.” This place-code is thought to combine with information about time within the hippocampus to link events to a particular time and place, a key step in the formation of episodic memory.
A recent paper from Jeff Magee’s lab at the Janelia Farm Research Institute, published in Nature Neuroscience in August, attempts to address several important unanswered questions about the cellular mechanisms underlying place cell activity. How do place cells become selective for their preferred locations, known as place fields?
The answer to this question must have something to do with the unique combinations of inputs place cells receive from other cells. The inputs to an individual place cell come from several different brain regions, and inputs that come from the same brain region tend to be grouped on the same part of their target cell. Do these input groups play distinct roles in shaping place cell activity? Do these inputs carry information about the spatial location of the animal that might contribute to place cell formation?
To approach these questions, the authors made electrical recordings from individual place cells in awake, behaving mice. In contrast to most place cell recordings, which are extracellular and pick up on the output of a cell, Bittner and colleagues achieved technically challenging intracellular recordings, which carry information about the cells’ inputs. To allow for these delicate recordings, the researchers fixed the animals’ heads in place and trained them to walk on a mini treadmill. The treadmill was covered with various objects, such as stickers and posts, so that place cells in the hippocampus would learn to become active when the mouse walked past a particular landmark.
The authors were interested in changes in the intracellular electrical activity called dendritic plateau potentials, which cause changes in the strength of input signals that arrive simultaneously. They hypothesized that these moments when the connection strength between neurons is malleable are the key to the formation of new place fields. Sure enough, whenever a cell spontaneously formed a new place field, the authors observed a dendritic plateau potential had occurred when the mouse walked past the same location on the preceding treadmill lap. Even more convincingly, when the authors created plateau potentials by injecting current into the cell, they were able to cause new place fields to form in 15 out of 16 cells tested.
These results showed that dendritic plateau potentials can cause new place fields to form, a significant contribution in its own right. But Bittner and colleagues did not stop there. They next asked how plateau potentials relate to the different inputs received by the cell. One organizing feature of the hippocampal network is the theta oscillation, an electrical signal arising from the combined activity of many cells that oscillates roughly 10 times a second. Inputs from different brain regions occur at different times relative to this oscillation, and the authors leveraged this fact to study the relationship between plateau potentials and the different input classes.
Using the theta oscillation as a reference, the authors found that plateau potentials occurred in lockstep with input from a brain region called the entorhinal cortex. When they inhibited input from this region using optogenetic techniques, they , observed fewer, and weaker, plateau potentials. These results strongly suggest an essential role for entorhinal input in the generation of dendritic plateau potentials, and therefore in the creation of new place fields. This finding is exciting for the field because the entorhinal cortex houses a number of cell types, such as the grid cell, that become active in specialized patterns relative to the animal’s environment.
One of the grand challenges in neuroscience is to connect the various levels at which the brain operates, from molecules and synapses, to cells and networks, to cognition and behavior. Place cells in the hippocampus are an essential node in the brain’s memory network, and yet have response properties that are concrete enough to be dissected at the molecular and cellular level. This paper combined several technically challenging methods to make novel contributions to our understanding of how place cells rapidly form new place fields, and as such is an excellent example of the types of insights that are now possible in the field and can be expected in years to come.
Reference (open-access link):
Bittner KC, et al (2015). Conjunctive input processing drives feature selectivity in hippocampal CA1 neurons. Nature Neuroscience, 18 (8), pp. 1133-42.