So I spoke with Professor Tan who is visiting from Purdue last week about the different testing options that are available to get the information I need.

Mainly, I'm concentrating on the torque vs. displacement theory...which one is better to apply? Although initially I thought that doing some threshold tests and matching magnitude tests would be helpful, I realized that doing those tests would provide data, but would not really help me to distinguish which one was better. While talking with Hong, I realized that what I really wanted to know was which method of skin stretch (applying a torque or applying a displacement) would be the most uniform. For example, I had noticed on the back of my hand, that when my fist was open and the skin was looser, I could apply a 20 degree displacement and not feel a very intense reaction, but if I made a fist and had tighter skin, a displacement approximately similar, produced a much more intense feeling. This made me think that the skin properties from person to person would affect the intensity felt. The same skin stretch displacement could be applied to two different people, and one person could perceive the stretch as a very low magnitude, while the other could feel it was much higher.

After this disucssion, Hong suggested doing magnitude estimation tests instead. The magnitude estimation tests would basically provide a curve, plotting torque applied and estimated magnitude. The hypothesis is that while there will be some variation in slope/location of curves plotted versus torque, there will be much more variation versus displacement. This test will help us see how the perceived intensity differs with the different type of application of skin stretch. This seems to be the easiest, and quickest test that can be done to give us the information we need. The drawbacks of the magnitude estimation test is that it relies heavily on the human subject accuracy/attentiveness. There can be occaisions where the subject gets bored and starts randomly assigning magnitudes, or there are heavy biases...one person may tend estimate larger numbers than another, etc.

However, it would be quick, and apparently there are methods to try and reduce the bias. Hopefully we can get this running soon, because it's a bit of a hump in our project flow. I'd like to be able to get this information quickly so that we can move on to more relavent experiments.