Mental rotation is the cognitive ability of the human brain to view a stimulus of unknown rotation, and then internally rotate a mental image of that object until it is upright, and comparable with a mental perception of what it should look like (Yule, 1997). It was Burrhus Skinner (1953) that was first to come up with a primitive mental rotation test (at the time termed operant seeing) (Cohen & Blair, 1998).
This involved timing a subject’s response time in reorganising small, painted cubes, organised to form what is most simply described as, an unbound Rubik’s cube. By viewing the subject’s behaviour, Skinner was able to interpret the subject’s mental behaviour, and concluded that the subject would mentally rotate the cube through its faces, and if the pattern in the subject’s mind was equal to that which was desired, the subject would then manually move the block to that position. He pointed to the subject’s response time increasing proportionally to the number of block stimuli present, as evidence to his theory (Cohen & Blair, 1998).
This was a conclusion shared by Shepard and Metzler (1971) (as cited in Cooper and Shepard, 1973) based on their initial mental rotation experiments. They reported that Reaction Time (RT), when comparing two, same shape, three-dimensional line drawings, increased linearly with the angular difference between the two (Koriat & Norman, 1985). Shepard and Metzler took this to mean that the increasing delay in responding was due to the time taken to rotate the stimulus in the subjects mind increasing in relation to the “complexity” of the stimulus (Cooper & Shepard, 1973).
Cooper and Shepard’s (1973) data conflicted with that of Shepard and Metzler’s. They found that their results, when graphed, showed a quadratic (non-linear) function for rotated, normally presented alphanumeric stimuli, rather than the linear plot found by Shepard and Metzler (1971). However for reversed stimuli, the plot was a straight line (linear) function (Cooper & Shepard, 1973).
It was suggested that a possible reason for the conflicting data was that as a subject became more familiar with the stimulus, small departures from normal would not cause an increase in response time, while a novel stimulus, when rotated would not be as easily identified, leading to larger response times (Koriat and Norman, 1985). However, Cooper and Shepard did concur with Shepard and Metzler’s (1971) mental rotation hypothesis, saying that a subject will mentally rotate a stimulus and compare it to a representation of the figure stored permanently in their brain (Cooper & Shepard, 1973).
It wasn’t until Koriat and Norman (1985) that another reasonable hypothesis was suggested. For their initial experimentation, they tested a broad orientation tuning hypothesis, whereby it was suggested subjects would become so familiar with a stimulus, that small deviations in the angle in-which it was presented would result in little difference in the response time when asked whether it was presented normally or reversed (Koriat & Norman, 1985).
In further experiments, Koriat and Norman attempted to further support their broad tuning hypothesis. They first trained subjects on novel, nonsense stimuli, where the subjects were to identify whether the stimulus was normal or a mirror image (though always upright). This was followed by testing where the stimuli were rotated away from normal. Results showed that the more a stimulus was practised, the more the plot of the data moved toward a non-linear function for both normal and reversed stimuli (Koriat & Norman, 1985). The aim of the present experiment was first, to replicate Koriat & Norman’s (1985) experiment, demonstrating the mental rotation phenomenon. The second aim was to determine whether there is a different function for reversed versus normal stimuli, providing support for the broad tuning hypothesis.
Subjects for the experiment were 23 students from the University of Otago, and it was hoped that this would give a valid representation of society. The experiment design was within subjects as all subjects were put through the same computer aided testing, where they were presented with one of three different letters in different orientations (normal or reversed), and one of six different rotations from upright (see appendix 1 and 2). These were the manipulated (independent) variables. The measured or dependant variables were the reaction time in milliseconds (msecs) for each orientation and each rotation.
The hypothesis for this experiment was that normal letters would provide a non-linear function and that reversed letters would provide a linear function. Method Participants Participants for this experiment were 23 (5 males and 18 females) undergraduate students from an Introductory Psychology class from the University of Otago. The subjects were required to participate in this experiment as part of the terms requirements for the class. The median age for the subjects was 18 years. It was assumed that no members of the group had any disabilities that may have affected the outcome of the experiment. It was also assumed that no subject had previous experience in mental rotation activities. It is worth pointing out, that one of the 23 subjects did not speak English as a first language. However, G, R and F, are all characters are part of his native language.
The equipment used for the experiment was an IBM computer with keyboard, connected via cabling to the University of Otagos’ local area network (LAN). The connection to this network allowed the use of software specifically designed to run the Mental Rotation experiment, and this program was preloaded on the computers. The keyboard allowed for the subject to respond to the presented stimuli via the “Z” and “M” keys (“Z” for normal and “M” for reversed). At the completion of testing, results were collected and collated by the network.