Measuring the Angle(s) of Color Aftereffects in the McCollough Effect

Stuart Bachner, Rachel Eaton, and Meghan Rector, Alma College

     The purpose of this study was to measure the angle(s) of color aftereffects due to orientation in the McCollough Effect using a rotating grid test pattern. In 1965, Celeste McCollough "demonstrated a contingent color aftereffect by pairing complementary colors with orthogonally orientated grids." McCollough further pointed out that the appearance of color aftereffects "indicate that edge-detector mechanisms in the visual system are subject to color adaptation, responding with decreased sensitivity to those wavelengths with which they have recently been most strongly stimulated." To further our study, a comparison was done between the two tests, Vertical/Horizontal (V/H) and Diagonal/Diagonal (D). The measurements of the angles of color aftereffects were compared to discover any difference in the strength and breadth of the angles between V/H lines and D lines.

Methods:
    Ten subjects were used in each representative study. They were randomly chosen from the available campus population. The studies were run on Eyelines (Mac) software. Each study was done using complementary colors (green and magenta), one study used vertical/horizontal stimuli; the other study used diagonal/diagonal stimuli.
     Vertical/Horizontal: A grating of vertical black and magenta stripes was viewed for several seconds on a screen, alternating with an identical grating of horizontal black and green stripes (See Figure 1). Subjects, after watching this alternation for 4-5 minutes, reported a green aftereffect in the left half of Figure 2 and a red aftereffect in the right half. The grid test pattern in Figure 2 was then rotated until the color was gone, indicating the span of the color adaptation contingent on the orientation of the stimulus.

    Diagonal/Diagonal: A grating of diagonal black and red stripes was viewed for several seconds on a screen, alternating with an identical grating of opposite diagonal black and green stripes (See Figure 3). Each subjects viewed this alternation for 4-5 minutes. With the grid pattern turned 45 degrees from original, subjects reported a green aftereffect in the upper left half and a red aftereffect in the lower right half. The grid test pattern in Figure 2 was then rotated until the color was gone, indicating the span of the color adaptation contingent on the orientation of the stimulus.

Figure 2. Test pattern used for measuring angles of color aftereffect in the Vertical/Horizontal and Diagonal/Diagonal tests. The test pattern can be rotated to measure the angles of color aftereffects in both tests.

Results:
     Eighteen of the 20 subjects reported seeing both color aftereffects. Of the two who reported seeing only one, one subject reported seeing only green and one only red, both subjects were those exposed to the vertical/horizontal test. Figure 4 shows the means of each aftereffect along with the range of degrees.
     Vertical/Horizontal: The mean for the red aftereffect was 15.028 degrees, 14.877 degrees for the green aftereffect.
     Diagonal/Diagonal: The mean for the red aftereffect was 17.495 degrees, 14.709 degrees for the green aftereffect.

Discusion:
     Subjects 1-4 for the Vertical/Horizontal and Subject 1 for the Diagonal/Diagonal test were measured on a different test grid than the rest of the subjects. These subjects were measured on a rotating grid of horizontal lines, with no comparison of vertical lines beside it. It was found that the color aftereffects were more difficult to see and therefore measure using this testing pattern, so it was changed into the grid seen in Figure 2. As the original test pattern was more difficult to use for measurement, it is possible that the data for those subjects is not as accurate as the other subjects. The original grid could be a cause that led to the two subjects reporting only one color aftereffect.
     The graph in Figure 4 allows for easy comparisons between the Vertical/Horizontal and Diagonal tests. As stated in the results above, the range of angles for color aftereffect are larger in the Vertical/Horizontal test than in the Diagonal/Diagonal test. This shows lower orientation specificity for the Vertical/Horizontal tests, and respectively, higher orientation specificity for the Diagonal/Diagonal tests. Also, as seen in Figure 4, the means for the V/H-Red, V/H-Green, and D-Green are close, with D-Red being just slightly higher.
     The data from this study show interesting comparisons between the two tests. The idea that vertical and horizontal lines should be less orientation specific is not so out of the ordinary. People look at vertical and horizontal edges many times everyday, constantly stimulated the visual system. And as stated by Celeste McCollough in the introduction, the visual system responds "with decreased sensitivity to those wavelengths with which they have recently been most strongly stimulated." It would make sense for diagonal lines be more orientation specific as our visual system is not so often strongly stimulated by them.

More Posters

Web Posters from other courses and projects

Alma College Psychology Department