Lucia Jacobs’ “Parallel Map Model” elegantly describes animals’ behavior when facing a spatial cognition challenge, e.g. navigation or foraging. The existence of a grid-like metric “bearing map” and a relational transient “sketch map” seems well-supported with evidence that addresses Tinbergen’s four critical questions related to mechanism (process), function (adaptation), ontogeny (development) and phylogeny (evolutionary history and constraints). The PMM does not negate other models we have learned about, such as egocentric and allocentric reference frames, and orienting with geometry and features. Rather, it provides a larger organizing principle within which to understand spatial cognition grounded in biology and behavior.

I was really impressed with Lucia Jacobs' talk. In particular, I think it's great that someone is actually trying to make a connection between biology and cognition in their research. I liked the part where she talked about how the anterior hippocampus is dominant in males, and the posterior hippocampus in females. As a result, males are better at using distal cues and females are better with positional cues. I thought that her explanation of the gender differences in spatial cognition was pretty convincing.

I completely agree with Jacobs about the value of a comparative, evolutionary approach for identifying the cognitive functions performed by brain structures. I found the parallel maps model appealing, but had a hard time following all of the data she provided in support of it (for example, the many correlations she mentioned between hippocampal size in mammals and other traits--such as the wing size of bats being correlated with habitat complexity (how cluttered) and with hippocampal size).
I was especially intrigued by something that came up in the Q&A: that CA1 represents not just spatial topology, but can handle abstract topologies--I think list order was the example mentioned. I wasn't clear if she was saying CA1 in all species represents topology in this way, or if this is something unique in the human CA1--that the human brain has co-opted this mechanism of spatial cognition for other cognitive tasks. Either way, I think this would be especially interesting to cognitive linguists like Teenie Matlock, who argue that much of our abstract cognition involves spatial metaphors.

I agree with Mitch that the notion of hippocampus supporting topological relationships is one of the particularly compelling ideas Jacobs aired in her talk. I also enjoyed the clarity she employed is her use of Tinbergen's questions related to mechanism. Another important feature of her work was its decidedly cross-species tilt. This novel (to me) approach in the study of cognition seems promising, as it could lead us to discovering fundamental cognitive commonalities that might underlie much of animal cognition.
Though it might go without saying, her use of experimentation in the "real world" (with an all-volunteer subject pool of squirrels) is an excellent technique. Researchers will always be limited so long as they are studying the cognition of animals which were raised in an impoverished environment that could (and likely does) stunt cognitive development ...

I agree with Leo that the use of wild squirrels makes for interesting methodology. So many arguments on laboratory studies hinge on the unrealistic nature of the experimental spaces and the experimental tasks. To see a study run with wild animals in a free and open environment makes for a refreshing change and hopefully addresses some of the concerns of ecological validity. (Perhaps even given the tupperware, since these are city squirrels.)

While I like the idea of creating, maintaining and integrating different types of mental maps, I'm still not totally sold on some of the finer details. Jacobs claims in her 2003 paper that "…a sketch map requires significant spatial memory to encode the features of individual landmarks in an array. Sketch maps may be temporary or long-lasting; they are maintained in the hippocampus as an intermediate memory store during the period of consolidation, until they are overwritten. Should the information prove to be stable, such sketch maps are consolidated relative to the bearing map." From the sound of it, once a sketch map has been integrated into a bearing map, there is no need to maintain a separate representation of it, so it's overwritten. However, it's not clear to me how someone could later access the integrated sketch map in a different context – that is, if they later encounter a highly similar or identical local location in a different global location.
Someone in class gave the restaurant T.G.I.F. as an example, so I’ll run with that. I’m familiar with the T.G.I.F. near my home in Santa Clara, so I might have a sketch map for the restaurant which is integrated into a larger bearing map that includes sketch maps for surroundings (the mall) and landmarks (the hills to the east). When I came to La Jolla, I discovered that the T.G.I.F. on Nobel has EXACTLY the same layout as the one near my home (fact). However, if my sketch map for the other T.G.I.F. exists only as part of a larger representation, it’s not clear to me how I can access that information independently of the other sketch maps and the bearing map with which I associate it. Clearly I can; but if I can just pull it out of the integrated map, it's not clear to me what is really meant by "integration". I don’t think this is a major sticking point, but a better account of the discreteness/modularity of representations might have been helpful.