In their Nature Reviews|Neuroscience article, Finding the engram, Sheena Josselyn, Stefan Köhler and Paul Frankland discuss the recent developments, mainly in circuit neuroscience in mice, that contributed to finding memories on the cellular level, the so-called engram. They accumulate the evidence from past years mostly based on studies using fear- or reward-learning that show that one can identify, modify, disturb and also cross-link cellular ensembles that are necessary and sufficient for the recall of memories. For example, it is fascinating that one can express a stimulator like channelrhodopsin specifically in neurons that have been activated during a defined memory-task.

One caveat the authors mention is that this has only been thoroughly studied for avoidance- and reward-learning, therefore working with a binary behavioral task. This might mask imprecisions and problems that would be obvious when trying to write or recall more difficult memory-tasks.

However, one point of view could be, looking at this corpus of research, that the memory problem is solved. Not only can the memory-forming cells be found, but they can also be manipulated in order to show their causal involvement. Like every solved problem (given it is really solved), it immediately becomes boring; or, at least I’m tempted to look at the weak points or missing links, in order to be able to say, ok, this problem is not solved at all.

The main weak points that I can see:

1. The temporal sequence of those ‘engram neurons’ during activation is typically lost when expressing labels or stimulators like opsins in the respective neurons. Therefore, recall results in a tattered re-generation of a once temporally ordered pattern, and it is unlikely that nothing is lost during such a recall.
2. The molecular mechanisms of memory also remain to be elucidated.
3. It is not understood how the memory recall and associated processes like pattern completion work en detail. This is what happens on a timescale of maybe 100-500 ms. The ‘engram’ is sometimes treated like something static and stable, like a binary pattern on a hard drive. But in reality, memory recall is a process, and this process is still not understood and is indeed difficult to observe, because it is not known how many neurons and brain regions have to be observed at which timescale.

But despite these remaining open questions, it has to be acknowledged that the current state of research has already answered some interesting questions. Maybe the authors have the same opinion. The title of their review, ‘Finding the engram’, reminded me of the last chapter of ‘In search of the lost time’, called ‘Finding time again’, where the main protagonist of the novel indeed finds a way to access and work with his precious memories, about the loss of which he has written this huge tome.