Contrast Sensitivity Function
Comparing the use and accuracy of an arcade game for Measuring the Contrast Sensitivity Function.
A new method of assessing vision known as the Contrast Sensitivity Function (CSF) has emerged. It is important to measure the sensitivity of the visual system for humans and other animals as it offers information regarding individual visual function capabilities.
Contrast Sensitivity testing complements and extends the assessment of visual function performed in traditional acuity tests.
A computer program "Gabori Attack" was used among 10 participants to test the use and accuracy of an arcade game to measure the CSF. Methodical limitations prompt the expectation that the method will not be as accurate as previous methods.
Practical applications include the changes in vision with age and reasons for this to occur.
The most widely used measure of visual resolution is visual acuity. Acuity is based on the size of the smallest detail in a visual target. The value of visual acuity measurements is well proven for correcting refractive errors. However a limitation is that individual variation in standard measurements of visual acuity is rarely able to predict individual variation in performance, such as target detection and identification (Ginsburg, 1983). The letter ‘E’ acuity chart, developed by Snellen in 1862, has long been accepted as the standard for vision screening (Block, D., and Bosworth, M., 1992).
In the past two decades a new method of assessing vision has emerged, the measurement of the contrast sensitivity function (CSF). Computerised formats also available that measure contrast sensitivity functions include Nicolet CS 2000 vision Tester and wall chart VISTECH Contrast Sensitivity test System (Long and Penn, 1987). These methods expected the peak sensitivity to be 100.
Today, the CSF is typically measured using sine-wave gratings as targets instead of the letter optotypes typically used in tests of acuity. A contrast sensitivity assessment consists of presenting the observer with a sine wave grating target of a given spatial frequency. The contrast of the target grating is then varied while the observers contrast detection threshold is determined. This use of sine wave gratings was first introduced in vision by Schade (1956) and was subsequently used by early investigators to measure basic visual sensitivity (Campbell and Robson, 1968).
Contrast sensitivity testing complements and extends the assessment of visual function provided by simple acuity tests. Contrast sensitivity measurements offer information about an individuals ability to see low contrast targets over an extended range of target size and orientation.
Modern vision research has clearly demonstrated that the capacity to detect and identify spatial form varies widely as a function of target size, contrast, and spatial orientation (Brad*censored*, Campbell & Atkinson, 1978).
Unlike traditional methods of measuring the CSF, a computer program Gabori Attack was proposed to test CSF. This program is presented as a video game with clear instructions. The task was to quickly squish "Gaboris" that slowly emerge, gradually increasing in contrast at a random position from the grey screen using mouse clicks.
Due to uncertainty of the position, spatial frequency, and reaction time delay, the computer game tends to underestimate contrast sensitivity http://vsoc.benkely.edu/vsoc.
It is expected from details http://vision.psy.mq.edu.au/%7Epeterw/csfl.htm that the function will be shifted downward compared to other more rigorous methods. Therefore, it is hypothesised that it is possible to use an arcade game "Gabori Attack" to measure CSF however, we expect to reject the null hypothesis that the peak is less than 100. Additionally, due to methodic limitations the computer game is expected to be less accurate.
The human visual system is able to detect spatial frequencies up to about 60. However, spatial vision changes with age, an infants window of visibility is very different from an adults. An infant can’t see fine spatial details visible to an adult. In addition, even for spatial frequencies visible to both, infants requires more contrast than adults do (Sekular, R. 1994). This is because infants have an immature visual nervous system that fails to encode high frequencies. The CSF remains stable through young adulthood; but after age30, systematic changes in the CSF begin to appear. There is a steady loss in high spatial frequency sensitivity, much of the loss results from optical changes in elderly eyes.
The study used a computer game called Gabori Attack. This program was used for testing contrast sensitivity function. The peak sensitivity average result was measured at each spatial frequency. This was a repeated measure design as there was more than one measurement from each subject. The independent variable was spatial frequency as this could be controlled and manipulated. As the contrast increased, vision became more visible. The dependent variable was threshold contrast as there was no control over this.
The participants for the study were 10 Macquarie University Academic Staff (mean age = 42 years, sd = 7).
Contrast sensitivity was measured using and Imac computer with a screen refresh rate of 80 hertz with settings to millions of colours. Details of the program at http://vsoc.benkely.edu/vsoc
Data was obtained by using the program Gabori Attack; it is presented as a video game. The task is to squish "Gaboris" that slowly emerge from the grey background as quickly as possible with mouse clicks. Each of the 10 participants had exposure to each of the eight spatial frequencies. The size of the room was 6 metres long, four metres wide and three metres in height. The height of the desks that the computers were on measured 80cm. The peak sensitivity average result was recorded at each spatial frequency.
The mean sensitivity values for the 10 participants at each of the 8 spatial frequencies are shown. As shown in the fig. 1. The peak sensitivity, 69.359 is lower when compared to the hypothesis that the peak sensitivity is 100. The graph shows the relation between sensitivity and spatial frequency at varying levels of spatial frequency.
In the one sample test comparing the peak sensitivity 100 the results show a significant difference (t= 105.94287, where p * 0.0005). From these results we can reject the null hypothesis u=100.
Secondly, in the dependent sample test either side of the peak was measured there significant differences seen in fig 2 & 3 respectively (t= 76.939279 where p * 0.0005, and t = 165.1436 where p * 0.0005). These results are consistent with the characteristic inverse U shaped curve.
Fig 2. Fig 3.
Results of the tests indicated support the first hypothesis that the peak sensitivities was less than 100. The measure of 69.359 was significantly lower than that of 100. The CSF was shifted downward and slightly left.
Our second hypothesis was also supported, as the computer game did allow a reasonably good measure of the CSF. This was evident from results comparing sensitivity either side of the peak. There was a characteristic inverse U shape curve representing the CSF.
In tests that are more rigorous there would be differences between the methods used to this computer game. These differences would account for the movement of the CSF.
Firstly, the lighting of the room would be controlled to minimal lighting as opposed to the fluorescent classroom lights. The subjects of this test had no constant fixation points, the distance from the screen was not controlled and the use of the mouse was slow to identify the Gabori immediately. All these factors should be taken into consideration when results are obtained.
Considering the three classical behavioural methods proposed by Fechner, (Sekular, 1994) the method used in our arcade game experiment does resemble similarities to two of these psychophysical methods. Firstly, the method of constant stimuli as the stimuli is presented multiple times in a random order and the responses plotted, similar to the inverted U shaped curve developed in the gabori game.
The method of limits (Sekular, 1994) offers a shortcut to the previous method, where each of the stimuli is changed gradually until the observer response changes. The absolute method is very similar where by the light are increased in steps until the observer can see it and the observers response is recorded. This method closely resembles the Gabori arcade game, as the task was to squish the Gaboris as soon as they became visible that is, increasing the intensity of the Gabor patch.
The method of adjustment did not resemble the Gabori game as the observer was given control over the intensity of the stimulus, which was not the case in our experiment.
From Sekular (1994), it is also known that different species have different CSFs and therefore are most sensitive to different spatial frequencies. This is due to the different nature of each species and their role in the environment. Sekular (1994) suggests that the aquatic environment a fish lives in prevent high spatial frequencies from ever reaching the fish’s eye. Therefore, a human has very different spatial frequencies than a goldfish because of their different natural environment.
Block, D.J., and Bosworth, M.F., (1992). Contrast sensitivity vision testing: new screening technology for family physicians. American Family Physician, 45, n2 pp655 (5).
Brad*censored*, O., Campbell, F.W. & Atkinson, J. (1978). Channels in vision: Basic aspects. In R.Held, H. W. Leibowitz & H. Teuber (Eds.) Perception. Berlin: Springer- Verlag. pp3-38.
Campbell, F. W., and Robson, J. O. (1968). Application of' Fourier analysis to the visibility of' gratings. Journal of' Physiology, 197, 551-566.
Ginsburg, A. P. (1983). Contrast Sensitivity: Relating visual capability to performance. USAF Medical Service Digest (Summer): 15-19.
Long, G.M., & Penn, D.L. (1987). Normative contrast sensitivity functions: The problem of comparison. American Journal of Optometry & Pysiological Optics, 64, 131-135.
Sekular, R. (1994). Perception. (3rd edition). New York: McGraw-Hill (pp.92-94; 153-167;490-493).
Details found at http://vsoc.benkely.edu/vsoc
Details found at http://vision.psy.mq.edu.au/%7Epeterw/csfl.htm