Effects of Tachistoscopy on Mental Rotation

Previous studies by Shepard and Metzler (1970) demonstrated peoples ability to perform mental rotation of perceived shapes in order to determine if the shape is the same as another or different. An earlier study by Sperling (1960) demonstrated the capability of iconic memory to retain visual data (in this case a letter grid) which can be recalled after the image is removed. The present study combines elements of both experiments. The effects and interaction of the two is, to our knowledge, a rather new idea. The goal was to determine if there would be a significant difference between mentally rotated images presented in static or tachistoscopic formats. It is believed that either tachistoscopy will aid mental rotation by removing the burden of constant iconic processing, or that it will hinder rotation due to iconic decay and processing. It was proposed by Ruthruff and Miller (1995) that mental rotation may begin before the perception of the image is completed. However, if letters are being used, rotation might not proceed until a lexical identification of the letter is made. If this lexical processing is being used to identify the letters, mental rotation might be delayed until a complete identification is made. This study may help to determine when mental rotation occurs (before or after identification of a letter or image) and if mental rotation can be performed on an image stored in iconic memory without degradation of performance.

The experiment was presented to 31 subjects, 18 male and 13 female Alma College students, using the Superlab program on a Macintosh computer. Subjects were shown pairs of shapes with the shape on the right side of the screen rotated at varying angles. Some of the pairs consisted of two identical figures while others consisted of a figure and its mirror image. These figures included the letters F, R, and P, a dollar sign, an S with lines extending from the top and bottom, and an asymmetric quadrilateral figure as shown in Figure 1.

Fig 1. Images used as stimuli.

For each pair, three rotation measures were randomly chosen. Three trials were run with identical copies of each shape at those rotations, and three were run with mirror images at the same rotations. Rotations in all cases were made using a perceptual vertical axis at random selections from 0, 30, 60, 90, 120, 150, and 180 degrees. Subjects were made aware of the nature and identity of the six figures used prior to the first set. Each subject first performed a set of 36 trials in which the images were presented tachistoscopically for 80 milliseconds. Then the same 36 trials were presented again non-tachistoscopically (static) so that the results could be compared. The trials were randomized within each set. The subjects were instructed to identify whether the rotated image was the same by striking the "F" key (forward) or a mirror image by striking the "M" key (mirror). If an incorrect response was made, the subject was asked to respond with the correct response in order to continue. All times used for data were the times necessary for correct response. Incorrect response times were discarded.

Reaction times were on the average higher for the pairs presented tachistoscopically, with the average reaction time for these trials being 1509.57 ms., compared to 1491.03 ms. for static trials (see fig. 2).

Fig. 2: Average Reaction Times for Static and Tachistoscopic Trials Note there is not a significant difference between s-ave and t-ave.

However, the difference is not significant as shown by the results of a two-tail T-test {T(1115)=-.396, P=.6925}. As the subjects progressed through the experiment, their reaction times decreased for both tachistoscopic and static trials, with the reaction times for tachistoscopic actually falling below those for static around the 20th trial (see fig. 3).

Fig. 3: Mean Response Times for Trials in Sequential Order Note tachistoscopic trials have higher reaction times initially but have lower reaction times at the end, compared to static trials.

The T-test for the first 12 trials (fig. 3) was {T(371)=2.991, P=0.003}; for trials 13-24 {T(371)=0.067, P=0.9468}; and for trials 25-36 {T(371)=-5.217, P=0.0001}. When considering individual images, reaction times for shapes were lower than those for letters on the average. The letter P was consistently the most difficult image to identify. This was the case both in tachistoscopic and static trials (see fig. 4).

Fig. 4: Mean Response Times for Individual Images Note the slightly higher response times for letters.

The difference found between tachistoscopic and static trials on the whole, while not significant, does demonstrate a slight advantage for static trials. This slight difference between the two types of trials would initially seem to indicate that the mental rotation process was accelerated by a constant image. Consequently, it would seem that the iconic image was fading as the process of mental rotation commenced.

When the results were broken down to see what effect exposure to the experiment had on response times, a learning curve became apparent in both static and tachistoscopic trials. The learning curve for tachistoscopic trials had a much more accentuated slope in the early trials and proceeded to decline below the response times for the static trials near the end of the testing. The eventual advantage of tachistoscopic over static trials near the end of the experiment was pronounced with an approximate mean difference of 400 ms. over the final ten trials. This is contradictory to the overall average reaction times. A reason for this may be that once the subjects adjusted to the way the experiment was run, it was easier for them to mentally rotate the tachistoscopic images, possibly because the continued presence of the image in static trials interfered with the rotation process.

Another interesting finding was the lower reaction times of shapes as opposed to letters. The significance of this finding was established by a two-tail T-test, {T(557)=7.128, P=0.0001}. Reaction times for shapes may be lower than letters because, when shown letters, subjects spend a certain amount of time determining which letter the rotated image is before performing mental rotation. This might deal with the brains lexical processing of known symbols. In his 1980 study, White found that identification did not require mental rotation. However, mental rotation would be necessary to determine a mirror image character. The difference then, between letter processing and symbol processing, might come from the fact that the brain is lexically processing a letter, for purposes of mental comparison. A shape would not be lexically processed, and would be a simple pattern / attribute comparison.

Abboud, Hisham A. (1994) Superlab 1.68. Macintosh Testing Software. Cedrus Corporation.

Ruthruff, Eric and Jeff Miller. (1995) Can Mental Rotation Begin Before Perception Finishes? Memory and Cognition 23, 408-424.

Shepard, Roger N., and Jacqueline Metzler. (1970) Mental Rotation of Three-Dimensional Objects. Science 171, 701-703.

Sperling, George A. (1960) The Information is Available in Brief Visual Presentation. Psychological Monographs 74, 498.

White, Murray J. (1980) Naming and categorization of tilted alphanumeric characters do not require mental rotation. Bulletin of the Psychonomic Society 15, 153-156.

This page last modified on 24.5.1996 by Brian Martin

More Posters

Web Posters from other courses and projects

Alma College Psychology Department