Amygdala and Hippocampus

 I am applying for a PhD position in the group of Andreas Luethi at the Friedrich Miescher Institute for Biomedical Research in Basel. The application is in November and the interviews are in January. I decided to thoroughly prepare for the interview by reading literature and blogging about it to help consolidate the memories. The plan is to freely write about research to help me consolidate in memory as many research papers as I can read and trying to make connections between multiple research papers. I have found that writing about research in my own voice really helps with memory consolidation and fluid recall.

I have never performed animal experiments. I have worked extensively with other people's data. I lack the practical skills to assess whether my ideas are experimentally possible. I am interested in amygdala-hippocampus interactions, but I don't know whether it would be possible to record from both structures at professor Luethi's lab. So I decided to read as many papers about these two brain structures as I can, to familiarize myself with what other people have done.

One of the first papers I stumbled on and that is right on the mark is the paper by Ramirez, Liu, Lin, Suh, Pignatelli, Redondo, Ryan and Tonegawa (RIKEN–Massachusetts Institute of Technology (MIT) Center for Neural Circuit Genetics at the Picower Institute for Learning and Memory, Department of Biology and Department of Brain and Cognitive Sciences, MIT and Howard Hughes Medical Institute, MIT), "Creating a False Memory in the Hippocampus", Science, 2013. Here's a summary of my understanding of the paper.

It's an optogenetics paper. The experiment starts with injecting a channelrhodopsin virus into either the dentate gyrus (DG) or C1 area of the hippocampus. Channelrhodopsins are a key tool in optogenetics. Because it's a light-gated ion channel, cells that express this channel can be photo-activated. Mice on a particular diet, doxycycline diet, didn't express this channel in the DG. If doxycycline is withdrawn for a time period and the animal explores a new context, channelorodopsin channels are expressed in the cells forming the memory engram for that context, furthermore mCherry fluorescently labels these cells. Thus, at the end of the first phase, the cells that are activated in the context A are labelled and can be activated using light.

In the next phase of the experiment, the mice were fear conditioned in the context B. Fear conditioning means exposing the mice to a noxious unconditioned stimulus, in this case an electric shock. Now, a different ensemble of dentate gyrus cells gets activated by the context B endogenously. But this memory engram is artificially expanded by in addition optogenetically activating the tagged cells from context A. This creates a false memory which you can test by later placing the mice in context A and seeing if the animal freezes as though it had experienced the painful unconditioned stimulus in this context. If you place the mouse in a novel context C that was never conditioned, the mice should not freeze beyond baseline levels (the animals are placed on doxycycline before being put into context C to prevent labeling these cells with mCherry and the expression of channelrhodopsin). That's the test for the false memory. For further rigor, the experimenters tested for the overlap between the cells activating in context A and context C, by looking at cFOS expression in mice that were euthanized after the experiment. They found that context A and C activate non-overlapping cell ensembles.

At the end of the paper, the authors mention that this effect can be explained by updated Rescorla-Wagner componential model for two independent conditioned stimuli. This is highly interesting! Next, I'm listening to a MIT lecture on "Reinforcement Learning" by Sam Gershman to learn more about this model:-) https://www.youtube.com/watch?v=K5RVbXeDE5A&ab_channel=MITCBMM