Joe Z Tsien in the recent Scientific American (free pdf here until today) talks about how he devised different techniques with mice in order to understand what a memory is. In the course of the analysis he used Multiple Discriminant Analysis (MDA) to analyze how cell activation location where related to specific memories of different events (such as falling or an earthquake). The video of the MDA in action is here. This is fascinating on two counts: a Machine learning technique producing dimension reduction is used to isolate specific elements of an actual learning process. In the detailed version of the article, Tsien remarks that the initial feature space is about 260 and the subspace of interest is 3:
After collecting the data, we first attempted to tease out any patterns that might encode memories of these startling events. Remus Osan--another postdoctoral fellow--and I analyzed the recordings using powerful pattern-recognition methods, especially multiple discriminant analysis, or MDA. This mathematical method collapses what would otherwise be a problem with a large number of dimensions (for instance, the activities of 260 neurons before and after an event, which would make 520 dimensions) into a graphical space with only three dimensions. Sadly, for classically trained biologists the axes no longer correspond to any tangible measure of neuronal activity but they do map out a mathematical subspace capable of discriminating distinct patterns generated by different events.
One cannot be but thinking how this tool (MDA) is in fact mimicking the biological process. Tsien then goes on to talk about the hierarchical structure of the memory process not unlike the primary visual cortex model from MIT whereby the visual process reduces the feature space from several hundred dimensions down to very few.
The second fascinating aspect of this experiment with mice is the similarity with an experiment I went through multiple times: flying in zero gravity. Besides the floating experience, I have always noted the inability of most flyers to clearly remembers their experiences. This is so bad, that in any of the free flying planes used by either NASA or ESA, there are large counters of parabolas informing people how many remaining zero-g periods they will have to experience over the course of two hours.
One could think that this is due to the fact that people are not accustomed to the conditions and that it is somehow traumatic. It so happens, that even after 50 flights, I still have the hardest time remembering precisely what happened during these flights. My point is : Is the paper by Tsien really locating a real memory or something else ?