From Neural Wiki

Psychophysics experiments test individual responses to stimulus and are designed to help reveal how the brain works. Typical psychophysics experiments consist of a subject being presented with a stimulus and being asked to respond to it. There are many different flavors and this page seeks to explore visual system experiments in detail. I intend for this to be a detailed overview, if you notice any missing procedures please feel free to email me mike(at)neuralwiki.org.

[edit] Types of Visual System Psychophysics Experiments

The mammalian visual system is a fascinating bit of evolutionary engineering which is currently unmatched by artificial systems. Therefore there is great interest in understanding how the mammalian system works so that artificial systems can perform tasks which we take for granted like object recognition. The visual system is comprised of many neuronal layers, starting with the retina.

One way to think about psychophysics is the search for where human abilities break down..... i.e. what can the human visual system accomplish, and what can it not?

[edit] Open Source PsychoPhysics

A neat program called PsychoPy makes it possible to create PsychoPhysics experiments using open source Python software. My Psychopy page.

[edit] Definitions:

Stimulus Onset Asynchrony - (SOA) This is the time period between when a stimulus image is shown, and a mask is presented. Longer SOA's allow a subject to have more time to digest and identify what they saw in the stimulus image. If a stimulus image is presented for the same amount of time, but the SOAs are varied, then the subject will be better at identifying images with longer SOAs.......up to a point. For example, results obtained with an SOA of 1 minute is likely similar to results obtained with an SOA of 1 day, but an SOA of 200 ms will produce more accurate results than and SOA of 20 ms

Stimulus Image - Also referred to as the target image, this is the image that participants are supposed to focus on.

[edit] Rapid Series Visualization Presentation (RSVP)

In this type of test, several images are presented in a row with a short amount of time between them. A subject is asked whether they saw a particular image in the set. This test can be conducted where the subject knows what image to look for ahead of time, or where the subject is asked after the images are presented whether they saw a particular image.

[edit] Masking

In masking experiments a masking image is presented either before or after a test image. The subject is asked to indicate whether they saw a target in the test image. Masking works by fusing the stimulus image with an unrelated image and obscuring the target. Masks must be presented in a timely manner. I cannot show you an image in the morning and then mask it in the evening and expect it to be effective. Typically the time range for effective masking is in the range of milliseconds.

[edit] Backward Masking

In this test, a mask is applied after a stimulus image is shown. Between the stimulus image and the mask is a time period called the Stimulus Onset Asychrony (SOA). This time period is adjusted based on what researchers are looking for.

[edit] Forward Masking

In forward masking, the mask is applied before the stimulus image.

[edit] EEG Psychophysics

Electroencephalography (EEG) data allows researchers to visualize neural activity and does not depend on a participant providing subjective data. This is an important tool in performing psychophysics because it provides a view of what the visual system is doing with the images. Even if a subject cannot identify an image, there is some visual activity.

[edit] References