5.23.2008

switching to thesis mode

I'm in the beginning stages of writing my thesis. It will be on the topic of action selection (specifically, grip selection).
This means that if I write anything at all on this blog in the next few months, it will probably be a review of some paper that I'm going to cite in my thesis.

If you're interested in this kind of stuff, feel free to check up periodically. If not . . . well, consider yourself normal.

6 comments:

Lord John Marbury said...

Remember, Dan, that a thesis is "a statement or theory that is put forward as a premise to be maintained or proved," and you should be alright. And remember to do an outline. And remember to support your thesis. And remember to cite your sources. I before E except after C. Now get to work.

Russell said...

Don't take this the wrong way but what exactly is so groundbreaking about the grip-posture stuff from the previous post? To summarize what I took away:
* If you have to do something one of two ways, people pick the obviously better way.
* If the two ways are close to being equal, people think about it before deciding.
* If they're used to doing something a certain way, they develop a tendency to continue without hesitation.

I'm not too impressed with the findings. I already know that if I'm thirsty and holding a milk jug, I'm hardly likely to walk to the cupboard for a cup but if I happen to have a cup in my hand, I might deliberate for a second on whether I ought to use it or not. Same concept for using the remote control to push play. If you have a remote, you will, but if it's farther to the remote than to the DVD player, you're likely to just reach over and press play on the machine itself.

What are you trying to show that's not alreayd pretty common sense? (I trust that there is something new and sensational about what you are going to say or I would try to be politely silent and only think these thoughts to myself.)

daniel said...

Posters are merely skeletons that are meant to be fleshed out by a brief oral presentation. So it's forgivable that there are some fine-grained but important differences between what I'm studying and what you think I'm studying.
All of the conclusions you took away are in some way or another not supported by my data. I'll list those conclusions and respond to them.

1. If you have to do something one of two ways, people pick the obviously better way.
Well, this is the common sense position. And for the most part, this assumption formed the background for the more interesting aspect of my research, which was to see how people behave when the "obviously better way" isn't so obvious. However, in testing I found that many participants violated this commonsense assumption -- they consistently chose grip postures that were overtly awkward for them.

2. If the two ways are close to being equal, people think about it before deciding.
You gave some examples where this might be true, but I suspect that in most cases, this isn't a good description of our behavior. When you happen to encounter an object that is at an ambiguous orientation, you don't "decide" how to pick it up, and you certainly don't think about it under normal conditions. The object just looks like it should be picked up a certain way, and so you pick it up that way. In these cases, all of the work of "deciding" and "thinking" is already done in the act of perception, mediated by devoted areas of your brain, and when you do take up the activity of deciding and thinking in order to plan such actions, the resulting movements are demonstrably different in their kinematics than when they are done spontaneously (as they are in natural conditions). From this perspective, it would actually be surprising to see a significant difference between reaches toward non-ambiguous orientations and those toward ambiguous orientations. What is commonly called sense in some spheres isn't common sense in others.

3. If they're used to doing something a certain way, they develop a tendency to continue without hesitation.
Being "used to doing something a certain way" isn't quite the same as the idea of switching costs. For example, you'd probably agree that you've developed some refined skills for picking up a pencil. In a certain sense (your sense) it is just one skill. There is the pencil -- will this person automatically know how to pick it up, or will they have to figure it out because it is a new experience for them? Once they've done it enough, bam -- they've got the skill. But when I talk about the cost of switching, I'm talking about evidence that there may be two basic skills involved in picking up a pencil, and that we tend to be slower in picking the pencil up when we switch between these two basic ways of picking it up rather than using one way consistently over a number of instances. Presumably, you are used to picking up a pencil in both of these ways, and one way isn't inherently more difficult or strenuous than the other. So the surprising thing -- the un-common sense thing -- about this is that we have two basic skills for picking up something like a pencil. Not just one. That is what the data suggest, at least.

Have I addressed your questions? Let me know if there's anything unclear.

Russell said...

Ok, so that makes more sense--well, it makes more sense why it wouldn't be common sense and why you would find some value in studying it. I am curious though--do you think you will find that we are limited to 2 basic skills or are you extrapolating that there may be events that drive the creation or honing of multiple skillsets? In social settings, for example, we may have more than 2 for handling various situations? Is there any chance you are going to jump that far with your research? From there you could advance to why certain people seem to develop skillsets in larger multiples than others? Like the guy who always falls back on the same lame pickup line when confronted with a hottie who might be interested in him vs the fellow who seems to be original but in actuality has merely developed multiple skills for the situation.

daniel said...

That's a good question. There is a crucial difference between the task I gave the participants and a similar everyday task like picking up a pencil. In the experiment, the participants don't have any plans to do anything with the object once they pick it up. They just put it back down. But it is well established (by a researcher named David Rosenbaum at Penn State) that what we plan to do with an object constrains how we are going to pick it up.

The artificiality of my task works against me in some ways, and this is one of them. Participants were only allowed to use one of two possible grip postures, so it might seem strange for me to say that I didn't expect to see evidence that the brain treats them as two separate grip postures. Of course it does, right? But what if the brain only cares about where the thumb and fingers go on an object, and doesn't care that much about where they are in relation to each other (for the purposes of the grasp)? The two available grip postures would hardly be distinct for a brain like that. The point is, there are a few competing theories that attempt to explain how grip posture is formed during reach-to-grasp movements, and any evidence that the brain has separate (and possibly competing) representations for these two different grip postures is going to pose a problem for at least one of the dominant theories.

In no way do I intend to extrapolate that we are limited to two basic skills for every situation. That was just an artifact of how I designed my experiment. There are more than two ways to pick up a pencil, if only for the reason that there are more than two uses for a pencil (e.g. writing, drumming, relieving stress by twirling it, protecting oneself, etc.).

Debra Chambers said...

So, what happened to your thesis project? I hope everything went well with it, Danile. I would love to read about action selection, so I would be glad if you can give me a copy of your thesis abstract. Or better yet, the whole paper! Anyway, just beep me up when you can share it with me.