I really appreciated Professor McNamara's clear presentation of some basic ideas on spatial orientation (examples of egocentric vs. allocentric frames of reference). I also thought that his comments about the correlation between the ability to orient in space and success in various profession was very interesting. I'd also like to learn more about spatial orientation in parteints with Alzheimers disease and Schizophrenia, as well as about cases of infant hypoxia.
Overall, McNamara gave a very clear presentation of the differentail contributions of long-term and short-term memory on spatial cognition and explained very well that different processes are at work depending on the task that one is performing.

I really liked how McNamara clearly identified different uses of spatial representations (steering, path integration, wayfinding, place recognition). With regard to his spatial memory and navigation model, I'd like to know more about the viewpoint dependent subsystem and its relation to the egocentric subsystem. Is the viewpoint dependent subsystem simply a longer-term memory store for egocentric representations--that is, does it contain the same kind of information as the egocentric subsystem and make it available for similar purposes, or is it different in form and use? Also, I'd like to know more about how the subsystems are thought to interact (or not) during particular tasks.

I agree with Naja and Mitch that McNamara’s overview of spatial cognition was really helpful. Just to get from point A to point B, we need to contend with numerous environmental factors and engage multiple perceptual and memory systems. When you really look at it, the complexity is staggering. The experiments used intelligent methods to isolate variables and yield clear results. I’d like to hear someone talk about how these findings might apply to real world contexts, such as the design of libraries, parks and classrooms. McNamara’s communication was so clear in lecture format, I was puzzled why his paper was so hard to understand. The tyranny of publication deadlines perhaps?

Immersive VR seems to be an excellent paradigm for allowing real world motion while providing virtual world stimuli. The results are certainly far more compelling than armchair explorations and easier to control than real world manipulations. (Although if anyone does create an actuated shape changing room, I would love to see it!) I am still having a bit of trouble visualizing the vanishing corners, perhaps he merely meant an alpha transparency gradient from the center of each side to the edge of the side? Or why the break itself would not be a cue even if not as salient a cue as an actual corner?

It seems to me that it's remains difficult to characterize the cleanly the results from the study. One interpretation, indeed, is that there is a reference axis that is established, similar to the kind of axes we have imagined for doing math. But this seems just a little too clean to be biological. A more likely explanation might go along the following lines.
1) Clearly, there are more and less complicated ways of encoding information. If we see an array of objects on the floor, it would presumably be less efficient to have the memory of it be as though it was from underneath the floor, or from within the set of objects, or from so far away that the spatial locations of the objects are difficult to tell apart from each other.
2) So, if there are more or less complicated ways of encoding the spatial arrangement, there must be one way that uses the least amount of work/cognitive resources to encode the objects.
3) The question now becomes, what is the "cheapest" way to encode one's memory of the array of objects. This question likely depends on the developmental contingencies of the perceiver (whether they've seen many matrices, e.g.), but no longer do we presume the existence of a reference axis.
4) Lastly, we should not neglect the interpersonal component of these experiments. The experiments are run by experimenters, with the objects laid out by the experimenters. We should not presume that subjects are not thinking of the experimenter when doing the task. The subject sees the objects laid out in a non-random way, suggesting that the experimenter must have had some template for laying it out. In considering the experimenter, the subject might not adopt their own reference axis, but instead adopt the experimenter's. This leaves open the possibility that reference axes arise from an interpersonal understanding, rather than from an intrinsic cognitive reality.

I wonder how the results concerning reference directions translate to real world settings. The objects in these studies were all laid out in grid patterns that clearly supported one interpretation. Objects in the real world are rarely arranged so systematically. However, there are many other possible cues that can be used to establish and maintain a reference direction. I would guess that macro-level features such as coast lines or prominent landmarks may play a role, but also smaller features such the direction of streets or walkways. I also wonder how the concept of a reference direction may combine with different levels of spatial memory. For example, if the reference direction for memory of the location of a house on a street is the direction of that street, but that street is not aligned with a larger city grid, might the conflict between reference direction contribute to disorientation?

In considering the potential differences between VR simulations and non-VR experiments, it might be helpful to look at instances in which similar studies have been performed using both methodologies and compare the results. McNamara mentioned a VR study by Kelly, Avraamides and Loomis (2007) in which participants learned the orientation of items in one room (Room 1) from a particular position, then got tested in an identical (except for texture), empty room (Room 2), in which they were asked to imagine standing at the center of Room 1 facing one item, from where they were supposed to point to another (specified) item. The finding that interested me most was that facing (for example) right in Room 2 didn't give subjects an advantage in trials where they imagined facing right in Room 1. But, if subjects went back to Room 1 and were tested there, they WERE noticeably better when their actual and imagined orientations were parallel.

This seems to me to be a classic example of contextual encoding, a principle of learning and memory which states that, for the most part, you're going to be better at recalling information when you're tested in conditions that are more similar to those in which you encoded that information. People have previously shown that if you study a list of words, you do better at recalling them when you're tested in the same room than when you're tested in a different one; in fact, even if you're tested in a different room, you do just as well as if you're tested in the same room if you just take a minute before the test to close your eyes and mentally reinstate your context at time of study (e.g., think about the room, what was on the walls, etc.) (Smith, 1979).

I'm not pointing to this as a way of saying that all VR experiments would yield the same results as their non-VR counterparts, but it does suggest that comparing VR and non-VR studies might be one way of resolving the question of ecological valdity.