Week 10 Summaries
Physiological Measures of Presence in Stressful Virtual Environments
This paper is mainly about the quality of effectiveness of a virtual environment. The authors would like to compare physiological responses evoked by VE and its corresponding real environment. They carry out three experiments. The physiological measures include heart rate, skin conductance and skin temperature. The results are used to prove that use of physiological reaction is reliable, valid, sensitive and objective.
The experiment environment is a training room and an adjacent Pit room with a virtual drop of 20 ft. The participants are asked to carry an object in the training room, take it into the next room and placed it at a designated spot. And the virtual drop is waiting for him to check their reaction.
Users report different feelings about it. The feelings can be used as proof for a measure of presence. After the experiment, they work out the user’ change in heart rate, skin conductance and skin temperature between the two room. Some Likert questions are designed for subjects to learn more about their feelings during the experiment.
Finally they found that change in heart rate satisfied their requirements for a measure of presence, change in skin conductance less, and change in skin temperature doesn’t. Then the reliability, validity, sensitivity and objectivity of physiological measures are discussed. They used the data of the experiment to prove the measure of presence satisfy the four requirements.
Question: Should the experimenters tell participants that the hole was virtual? If yes, the participants would be prepared for the frightening, which could affect the results. If no, how did they get the approval from IRB since they didn’t tell the potential risks to participants?
Variations in Physiological Responses of Participants
This paper is about analysis of differences in participant reactions at different stages of experiment of physiological. They also use the method of participants observation to better interpret subjective ratings. The main way is to use physiological responses to evaluate stress level of subjects during virtual experience and figure out how nervous system reacts to novel situations.
In their experiment, they provide the scenario of a street with virtual people walking around. The independent variables are texture quality of the environment and visual aspect of the characters. Three signals have been recorded as dependent variables, which are respiration, GSR and ECG. Some questionnaires are used to get some subjective measures of presence. There are some other variables contributed by questionnaires about participants’ age, gender, occupation, language, experience in computer games, programming and virtual experience.
Forty individuals are included into experiment and distributed into four groups, 5 males and 5 females each. The participant will stand in dark for 90 seconds so that experimenter can measure their physiological signal in a relaxed state. Later a VE appears to them and they are asked to practice navigation in the training room. After they are confident, they will stand in front of a virtual door leading to the virtual street. The virtual experience lasts for four minutes and signals are recorded.
In the next section, the authors compare the different values among four time segments:
1. in the baseline and in the training segment
2. in the training and in the first segment of experiment
3. between the two segments of the experiment
Their conclusions are there are four stress levels in different time segments. The mental stress increases at the beginning of the virtual experience, during the training, then decreases at the beginning of the actual experiment and increase again at the end of the experiment. They also do some work about evaluation of different reactions between the experiment factorials and how gender can affect the the results.
Question: How do you prove that the stress during the experiment is resulted from the immersive VE instead of other factors?
The Uncanny Valley: Effect of Realism on the Impression of Artificial Human Faces
This paper talks about how realism of artificial human faces can affect people’s impression on them. Roboticists warned that the robots should not be too similar to real humans, otherwise they will fall into the uncanny valley. So they authors do an experiment to test whether the uncanny valley will emerge in the realism-impression plot.
Their experiments are web-based. They design frames of image sequences in which artificial faces was gradually changed to real faces. And the participants run JavaScript on their web and are provided the images. They have five button to show their attitude to the image they see, ranging from -2(extremely unpleasant) to 2(extremely pleasant). Once a button is clicked, the system will wait for a second to change to the next trial.
After the experiment, the experimenters analyze the data with ANOVAs. The data analysis is completed in three experiments. Their conclusion is that the uncanny valley actually emerge as expected. But it only emerges when face images involved abnormal features. So the unpleasant impression may not be attributed solely to the degree of realism. There can be more factor that affect humans’ feeling. They inform that the participants’ responses can be interpreted as judgement of facial attractiveness. Realism is only one faction of attractiveness.
Question: Is there any culture difference here? I mean things may be different for people in different countries.
Comments are closed.