A recent Nature paper from the Malinow lab in San Diego has attracted a lot of attention. Entitled "Engineering a memory with LTD and LTP", it describes how optogenetic stimulation of auditory inputs to the rat's amygdala were paired with mild foot shocks to produce a conditioned fear response.
Optogenetics is a genetic trick which allows the neuroscientist to activate a neuron with a flash of blue laser light. It's very trendy right now. The amygdala is a brain centre concerned with emotion, and particularly fear. You may have heard it mentioned a few times on "Firefly".
The clever thing is that the investigators were able to break this conditioned link between the blue flash and the fear response by applying a protocol called LTD to the auditory inputs, using the same optogenetic flashes but arranged in a very specific time sequence that has been known for many years to produce long term depression (LTD) at an individual synapse. Moreover, another protocol called long-term potentiation (LTP) could restore the erased fear conditioning. So it seems as though an acquired memory (in this case an association between a particular auditory input and fear) can be turned off and on at will by a bit of clever manipulation of synaptic strengths. This seems very close (at first sight) to claiming that we have discovered how to create or erase memories at will by strengthening or weakening individual synapses.
I have a couple of questions about this paper. First, why is fear conditioning so popular as an experimental model of memory? Surely there must be less unpleasant (for the rat) and just as robust protocols for producing associative memory. My friend Z tells me that fear conditioning is popular because the relevant circuitry is well understood, and the association can be produced with a single trial, so maybe not.
Second: does this experiment tell us much about how or even where this associative memory is encoded? After all, we can manipulate the strength of the synapse between an incoming auditory fibre and its postsynaptic target in the amygdala, but the actual site of the associative memory is not necessarily that synapse, it could be much deeper in the network, perhaps several synapses further along the chain, maybe not even in the amygdala at all. All that the LTP and LTD protocols are doing is "gating" the first synapse in the chain, allowing or preventing the incoming auditory signal to proceed further into the network. But the actual locus of memory formation is most likely elsewhere in the network, further down the chain of synaptic connections.
Nabavi S, Fox R, Proulx CD, Lin JY, Tsien RY, Malinow R (2014) Engineering a memory with LTD and LTP. Nature 511:348–352.
Showing posts with label Memory. Show all posts
Showing posts with label Memory. Show all posts
Monday, 28 July 2014
Thursday, 26 April 2012
Towards an artificial engram
Liu X, Ramirez S, Pang PT, Puryear CB, Govindarajan A, Deisseroth K, Tonegawa S (2012) Optogenetic stimulation of a hippocampal engram activates fear memory recall. Nature.
Tim Bliss and friends recently wrote about the possibilities of using fancy genetic tools to test the hypothesis (quite an old hypothesis too) that memories are encoded in networks of neurons whose synapses have been modified by LTP-like plasticity (Neves, Cooke & Bliss, Nat. Rev. Neurosci. 2008). This week's Nature has a remarkable paper from the Tonegawa group, where they have made a big step forward along this path. They used a clever combination of transgenic mice and virus transfection techniques to produce a mouse where activated neurons in the dentate gyrus (DG) of the hippocampus are not only labelled with YFP but also (and here's the powerful bit) express light activated sodium channels (ChR2). They found that fear conditioning labelled a subset of neurons in the DG, and what's more, that activating this set of neurons a few days later (by shining blue light into the hippocampus), even in a different context to the original conditioning, produced the freezing reaction that you see in fear conditioning.
So, a significant step forward in the quest for the engram. It's interesting to ponder how accurately this result reflects how memory operates in the intact animal. You might expect that more than the DG is involved, for example. Would it be possible to do similar experiments with similar results in CA1 or CA3? And what about other forms of memory, such as episodic memory, would they work in a similar way?
Tim Bliss and friends recently wrote about the possibilities of using fancy genetic tools to test the hypothesis (quite an old hypothesis too) that memories are encoded in networks of neurons whose synapses have been modified by LTP-like plasticity (Neves, Cooke & Bliss, Nat. Rev. Neurosci. 2008). This week's Nature has a remarkable paper from the Tonegawa group, where they have made a big step forward along this path. They used a clever combination of transgenic mice and virus transfection techniques to produce a mouse where activated neurons in the dentate gyrus (DG) of the hippocampus are not only labelled with YFP but also (and here's the powerful bit) express light activated sodium channels (ChR2). They found that fear conditioning labelled a subset of neurons in the DG, and what's more, that activating this set of neurons a few days later (by shining blue light into the hippocampus), even in a different context to the original conditioning, produced the freezing reaction that you see in fear conditioning.
So, a significant step forward in the quest for the engram. It's interesting to ponder how accurately this result reflects how memory operates in the intact animal. You might expect that more than the DG is involved, for example. Would it be possible to do similar experiments with similar results in CA1 or CA3? And what about other forms of memory, such as episodic memory, would they work in a similar way?
Subscribe to:
Posts (Atom)