For example, we found that Global motion perception is poor in a child but how do we know that the poor perception is actually affecting him/her in real life situations. I am interested in correlating psychophysical tests to real life situations. Any ideas worth sharing will be highly appreciated.
This is tricky to answer. Psychophysical test are developed for lab work and are pretty long, as you know. During my PhD I would have loved to get a psychophysical measurement implemented as a routine examination in the clinic. I tested migraineurs with and without aura and controls regarding their shape detection and discrimination. We did a long lab-based test and we developed a short clinic based test (before used to test children). The results weren't as expected as both test did not show exactly the same result. Our aim was to find a short psychophysical test which may can distinguish between a migraineur and a migraineur with aura. This test needs to be modified but I think could work in the future. Generally speaking, I would always use a questionnaire as Michael Bowen suggested, in combination with psychophysical testing. But I also always wonder how related to "real life" can a psychophysical test be?
Hi Nabin,
Your question is a very sensible one, and one that the Bionics Institute psychophysics team is working on with patients using a Bionic Eye. My previous experience is with hearing and Bionic Ear patients over the last 30 years. For hearing it has never been easy to find strong correlations between psychophysical test results and performance on spoken language tasks in either children or postlinguistically deaf adults. I expect the same will be the case for Bionic Eye patients. The explanation is that real world tasks tap into very different perceptual mechanisms than the specialised psychophysics tasks that use a different skill set and are performed in a different context where the subject's attention can be focussed on very specific perceptual cues that may not be available or may be unreliable in everyday life contexts.
The reasons we work on psychophysics are to help fit and tune-up the bionic eye vision processor for individuals in a well-controlled manner, and to understand the fundamental limitations and characteristics of the device. Tuning up the device to optimise performance in the real-world is much more complex and involves decisions about how to use the fundamental psychophysical characteristics of artificial vision to code the most useful information about the environment.
In summary, there may not be a strong correlation because the tasks are so different and the perceptual mechanisms can also be very different. If you want to find a correlation you need to find psychophysics tasks and real-world tasks that are similar - eg recognising Snellen chart letters and reading words on signs with lettering of different sizes and distances.
Good luck with this!
Peter Blamey
It is required to select the specific domain scores within the psychophysical tests a priori that you think best reflect the child’s performance in real life situations. For example there is some published research papers in pediatric neurosurgery examining the correlation between the neuropsychological test performance and quality of life in a sample of school-aged children with long-standing hydrocephalus.
Dear Michael, Doreen, Peter and Aye,
I agree with all of you in this. Regarding the domains that might be specifically affected by the variables (vision, motion perception), we won't know which activity is more affected by these factors unless we do a cross-sectional analysis and find a correlation of visual parameters and other real life parameters which could open a way to generate simulation experiments. For example, if we found that people who had poor motion perception had frequent accidents, we can say that motion perception is somehow related to driving performance.Then we can generate simulation of driving tests and test in those with poor perception and normal perception whether there is an actual difference between them.
I totally agree with the idea of Peter that we need to find psychophysical tests that simulate the real life situations. For example, Instead of using random black or white dots for motion perception we can use stimulus ( like snow, or rain drops with clouds and sky on the background and grass on the ground or something like that) that resembles that objects or activities that we see in real life. I think that makes sense.
On the other note, to see how a particular psychophysical training has affected in different tasks (as that of bionic eye), vision specific QoL measures might be one option.
Thank you everyone for the active discussion, I hope to hearing from you more and more in the days to come. Your input is highly appreciated.
Regards,
Nabin
Thanks Michael, I am more interested to apply those methods in children so that they simulate real life situation. Let's see!
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence