Purpose

There seems to be a region of skin stretch where it is undetectable. From our proprioception experiments, there appeared to be a "deadband" of rotational skin stretch, where subjects complained that they could not detect any movement. From qualitative observations and earlier tests, I would guess that this detection threshold is less than 10 degrees of stretch. In other reaction time experiments, subjects were able to detect 10 degrees of stretch consistently, yet it was an extremely "weak" signal and subjects observed that it would take them a little longer to register that a stimulus had been applied.

Research Questions

What is it that we detect first? i.e, are we detecting pure displacement of skin?

If velocity of skin stretch changes, does the detection threshold change? (hypothesis: yes) If so, then what is the dynamic detection threshold?

How do different skin properties effect the detection threshold?

How does cognitive load effect detection threshold? In a concentrated, secure environment, are the thresholds lower than in real-life scenarios? (i.e. moving user experiment- while a subject is on the treadmill, can they detect signals that are detectable in static environments?)

Is the detection threshold the same in increasing/decreasing skin stretch? i.e. detect change in 4-5 degrees, vs. change from 5-4 degrees?

References

Hong Tan

Hong Tan has several papers/lectures regarding perception and psychophysics as used in haptic applications. I used her papers/online resources as a starting point.

• Some lecture notes about adaptive psychophysical procedures. Adaptive procedures seem to be used more frequently due to their efficiency (less time and less number of trials) and http://cobweb.ecn.purdue.edu/~ece511/lectureNotes2006/pp12.pdf and http://cobweb.ecn.purdue.edu/~ece511/lectureNotes2006/pp13.pdf. The goal is to find a stimulus level where subjects can identify the stimulus correctly at least 70% of the time. Several adaptive methods have been developed to converge on a level that is detectable 71% of the time. This corresponds to the 71% level on the psychometric function...

• "On a psychophysical transformed-rule up and down method converging on a 75% level of correct responses" - http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=31916- simply gives some more examples. " Transformed-rule up and down (UDTR) methods seem to constitute the most popular group of psychophysical methods for stimulus detection experiment"

• The original paper that details the UDTR method. psychoacoustics.pdf: