Sex differences in spatial memory: Object placement and recollection

The concept of spatial ability is a very broad and complex topic, which encompasses a wide variety of tasks. Maze learning, tracking objects, and map-reading are a few examples of activities that have been categorized under this heading. Studies consistently show that tasks involving spatial memory are accomplished to a better degree by males (Banich, 1996). Until recently, it was thought that object location memory could be placed in the same class as these other spatial activities. However, male superiority for tasks involving specific memory for object positions is now being called into question by many researchers (Postma, 1998).

The importance of object location memory is indisputable. Remembering where you put the car keys, or which cupboard the peanut butter is found in, all rely on this complex process. It is also fundamental for training situations, including hostage rescues, fire fighting, and building searches (Guest, 1997). Individuals involved in these types of scenarios are required to navigate through an unfamiliar environment by following instructions based on landmark location (for example, turn left at the fire door, etc.). Therefore, it is absolutely necessary for both everyday and life-saving situations to understand how objects are remembered and placed into positions, and consequently, how routes are learned in this way. Since sex differences have been consistently observed in the domain of spatial memory (Banich, 1996), it is also required that both the male and female cognitive processes involved in these tasks be fully comprehended.

Postma et al. (1998) have been studying object location memory for many years. They have broken down the process into two or three components. The first component involves encoding the precise position of an object (referred to in this paper as a positions-only process). This is followed by assignment of particular objects to their proper positions (defined as an objects-to-positions process). They also propose a possible third step, which consists of integration of information from the previous two steps. Evidence for this dissociation of spatial memory has been observed in amnesics (Postma, 1998) and older children, both of whom perform better on positions-only tasks (Schuman-Hengsteler, 1992).

It has been argued that this dissociation in object location memory can be based on the idea of coordinate and categorical representation proposed by Kosslyn (1987). Coordinate spatial encoding, preferentially handled by the right hemisphere, involves the use of a metric coordinate system to remember precise positions. In contrast, categorical processing, which is primarily a left hemisphere function, encodes objects according to their relative positions. For example, object A is to the right and below object B (Kosslyn, 1987;Postma 1998). It is proposed that positional encoding primarily employs a coordinate system, while categorical processing is necessary for object-to-positions assignment (Postma, 1998). In addition, it has been hypothesized that women are superior at categorical representations whereas men are better at coordinate encoding (Kosslyn, 1987).

A recent study preformed by Sandstrom et al. (1998) utilized a computer-generated version of the Morris Water Maze (MWM) to study gender differences in human spatial navigation. By altering distal cues such as landmarks and room geometry, it was shown that females rely heavily on landmark information whereas males use both landmark and geometrical cues (Sandstrom, 1998). It is hypothesized that because females use landmarks as an orienting strategy, they have a greater ability to recall landmarks in comparison to males (Kimura, 1992).

Postma et al. (1998) attempted to study gender differences in object location memory (specifically positions-only and objects-to-positions assignment) with the use of a two- dimensional computer task. The subjects, which consisted of twenty males and twenty females, were shown ten objects at various positions on a PC screen. The objects were displayed for 30 seconds before they disappeared. In the objects-to-positions test the objects reappeared at the top of the screen and subjects were asked to relocate them to their original positions, which were indicated with a marker. In the second, positions-only task, subjects were required to place ten identical objects into their previous positions with no markers indicating the previous location of the object. In a third task subjects had to both correctly identify the object as well as its position on the screen. Postma et al. observed male superiority in the positions-only and the combined conditions tasks. However, no gender difference was found in the objects-to-position task. Therefore, it can be implied that the male superiority typically observed in spatial tasks is not a generalized phenomenon, but applies solely to positions-only memory.

One of the shortcomings of Postma’s research was the restriction of the two dimensional spatial environment. It has been questioned if various studies, including that of Postma, will hold true when they are tested in a more life like setting, such as a computer generated virtual environment. Virtual environments and table-top computer screens differ in fundamental ways. Virtual reality is able to alter perspective and frame of reference (allocentric vs. egocentric) so as to simulate reality more closely (Maguire, 1999).

This study attempts to replicate portions of Postma’s paper with some modifications. A virtual reality device was used in this study to test Postma’s findings on gender difference in a more life-like setting. Using a stationary bicycle, both male and female subjects navigated through a virtual reality maze, which contained seven landmarks varying in shape and colour. In order to assess their ability to place objects to their correct positions, subjects were asked to identify the location of the objects they had seen in the maze by marking an 'X' on a scaled, 2-D representation. The subjects were then shown the correct positional information and in an attempt to measure their efficiency at an objects-to-positions task, they were asked to identify both the colour and shape of the object at each of the seven positions.

Based on male and female differential dependence on landmark cues (Sandstrom, 1998) and differential processing of spatial information (categorical vs. coordinate) (Kosslyn, 1987), it is expected that Postma et al.’s general findings should be replicable in a virtual environment. Therefore, females should be more efficient than males at an objects-to-positions task. In contrast, males should excel in a positions-only task involving a measure of accuracy in recollection of precise object position.

It will be shown that in a virtual environment, a female superiority effect exists for an objects-to-positions task. However, due to test sensitivity or difference in methodology, the greater efficiency of males compared to females in a positions-only task, which was previously observed by Postma et al., was found to be non-existent in this study.

Method

Subjects

Members of the McMaster University/Hamilton Community were selected to participate in the experiment. The subjects consisted of undergraduate students between the ages of nineteen and twenty-four (mean age = 20). All of the individuals volunteered to perform the task and answer questions following the completion of the experiment. Five males and eight females comprised our experimental group of thirteen subjects.

Apparatus

The experiment required the use of a bicycle, a virtual reality head set and a SGI 02 computer. The bicycle was located in the middle of a room and was fixed to the ground. Since the bicycle was stationary, it could not be made to move through real space. Peddling of the wheels, however, would create the perception of movement under a virtual-reality setting. A virtual reality head set (HMD, V8, Virtual Research) was used to create the three-dimensional atmosphere.

The visual scenes, viewed through the HMD, were updated by an SGI 02 computer - through inputs mounted on the bicycle. The use of two sensors was required in our experiment. The first sensor was located on the steering column (to signal direction) and the other sensor on the rear wheel (to signal the magnitude of velocity in the forward direction). Peddling of the stationary bicycle allowed the subject to move through the virtual maze. The visual perception created by this device included a practice maze (without landmarks) and a test maze with seven objects positioned at fixed locations throughout the maze.

Procedure

All subjects, that volunteered to take part in the study, were required to complete a waver prior to the experiment. This waiver informed the reader of the purpose of the study and the consequences that may arise from performing the experiment (such as nausea or dizziness). Prior to the spatial navigation task, in the test maze, subjects were informed that they would be given time to familiarize themselves with the equipment in a practice maze (Phase one). Following navigation through the test maze (Phase two), the volunteers were brought into a separate room and asked questions (Phase three). The questions required the subjects to recall the locations of the various objects. The exercise that followed required the subjects to recall the colours and shapes of the objects themselves, after they were given the correct locations in the maze (the locations they had attempted to identify in the first set of questions).

As stated previously, the purpose of the first phase of the experiment was to allow the subjects to adapt to a virtual reality environment and become familiar with the equipment. At the beginning of the study each subject was informed that they would have as much time as they needed to explore the practice maze. They were also told that they could stop the study at any time they deemed appropriate. The subjects were then instructed on how to use the equipment (i.e. peddling allows you to move forward in the maze) and perform the task (i.e. try not to go through walls). Finally, after the instructions, the subjects explored the maze until they felt comfortable and notified the experimenter that they were ready to try the ‘test’ maze. The time subjects allocated themselves in the practice maze was recorded without their knowledge.

After the subjects explored the practice maze they were then immediately told that they would have to ride the bicycle through the test maze, from entrance to exit. (See Appendix I for a two-dimensional representation of the test maze). In compliance with the practice phase, all subjects were informed that they could take their time navigating through the maze because time would not a factor in our evaluations of their performance. They were also told that they could stop the experiment if they were beginning to feel nauseous or disorientated. Subjects were then instructed to begin the task. Recordings were taken to measure the amount of time it took each subject to traverse the “test” maze. Upon completion of the task, each subject was asked to remove the virtual reality headset and dismount slowly from the bicycle.

Finally, after performing the navigation task in the virtual reality test maze (phase two), each subject was brought into a separate room (phase three). It was here that they were presented with two different two-dimensional representations of the test maze. The first representation outlined the general shape of the maze, but did not include any information on the characteristics or locations of the previously encountered landmarks (see Appendix II). Subjects were asked to locate and write, on the map, the positions of all the objects that they remembered encountering in the test maze. After completing the positions-only task, subjects were then presented with a second representation that included the locations of all seven landmarks (see Appendix III). This objects-to-position task required the subject to recall the shape and color of the object located at the marked positions in the maze. Individual performances, on both tasks, were recorded.

Following completion of the experiment, by all subjects, all data collected was formally tabulated, modified and then analyzed using Microsoft Excel. Subjects that were not able to complete the experiment were not included in the data analysis.

Results

The results calculated were separated into two different parts. The first part dealt with the positions-only reporting where the subjects were asked to place an ‘X’ in a topographical map to represent the position of an object. Previous research has shown that males should show a significant advantage over females. The second part of the experiment dealt with object recall, namely the colors and shapes of the seven different objects in the maze. It has been established that females have more of a landmark oriented spatial memory and should therefore be more likely to remember the objects they saw.

The statistics calculated include a total of ten subjects (males = 5 females =5). Even though a total of thirteen subjects were tested, three females had to be excluded because they did not complete the maze due to feelings of nausea. Since maze completion was essential for the testing, these subjects were not tested and therefore yielded no data.

The positions-only data was calculated by measuring the distance between the marked ‘X’ by the subject versus the actual position of the object in the maze. Because the subjects were not told that there were seven objects the experimenter had to decide which object the subject was marking on the basis of proximity. Creating virtual compartments in which any guess within a section of the maze was assumed to refer to a particular object(See Appendix 4). The average number of marked ‘X’s for males and females were very similar. (Males: M_m=4.8, SD_m=2.38,SE_m=1.06 Females: M_f=4.6, SD_f=1.14, SE_f=0.51)

[INSERT FIGURE 1 & 2]

A two-factor analysis of variance was used to analyze the data for the number of marked X’s and found no significant difference by subject (F(9,54)=0.822 p = 0.60) nor by object number (F(9,54)=0.924 p=0.48). This would suggest that there was no sex difference in the number of marked X’s in the positions only task and that there was no primacy or recency effect for the memory of positions.

[INSERT FIGURE 3]

A t-test was used to examine the mean distance between the marked positions versus the actual positions of objects for male subjects versus female subjects and found no significant difference when the data was grouped in this manner. (T(4)=-1.74, p > 0.05). This points towards a non-existent sex difference in the precision of positions only task

[INSERT FIGURE 4]

Subjects were also timed while they were in the maze. Both males and females spent approximately the same amount of time in seconds navigating through the maze. (Males: M=175.8, SD=96, SE=43, Females: M=179.6, SD=54, SE=24)

[INSERT FIGURE 5]

Both gender groups had an outlier for the amount of time spent in the maze who took much longer than the rest of the group. However removing the outlier from each group and analyzing the data showed no significant difference between males and females in time spent in the maze (T(3)=-1.11 p =0.17). Once again, if we remove the two outliers and examine the number of correct responses per second spent in the maze we find no significant difference (T(3)=-0.75 p=0.25). We can therefore eliminate the possibility that those who spent longer in the maze had an advantage over those who went through the maze more quickly.

In the object recall task it was expected that females would perform better than males as stated above. In analysis of the correct reporting of shape only we find that females do indeed have a better memory than their male counterparts (T(4)=-3.77 p=0.009).

[INSERT FIGURE 6]

Females did not however show a significant advantage in recalling the colour of the objects (T(4)=-0.75 p=0.25). Nor did they show a significant difference in getting both colour and shape correct (T(4)=-1.39 p=0.11).

The combined data for both colour and object shape did show a definite recency effect for objects 6 and 7 versus objects 3 through 5 (T(9)=2.64 p=0.01). The primacy effect was not as apparent (T(9)=1.40 p=0.096).

[INSERT FIGURE 7]

If we examine the relationship between precise positions and correct object identification we find no significant correlation (X²(1)= 0.289 X²_crit=3.84). A precise position was defined as doing better than the average distance for the positions only task (M=3.5 SE=1.02). This suggests that being able to identify an object correctly is independent of assigning the correct position.

[INSERT FIGURE 8]

References

Banich, Marie T. (1997). Spatial Memory, pp. 902-956. In David C. Lee, Neuropsychology – The Neural Basis of Mental Function. Houghton Mifflin Co., New York.

Guest, M., Bliss, J., amd Lohmeier J. (1997). Landmark enhancement and strategic processing: An evaluation of strategies for spatial navigation training. Perceptual and Motor Skills, 85, pp. 1123 – 35.

Kimura, Doreen. (1992). Sex differences in the brain. Scientific American.

Kosslyn, S.M. (1987). Seeing and imagining in the cerebral hemispheres: A computational approach. Psychological reviews, 94, pp. 148 – 175.

Maguire, Eleanor A., Burgess, N., and O’Keefe, J. (1999). Human spatial navigation: cognitive maps, sexual dimorphism, and neural substrates. Current Opinion in Neurobiology, 9, pp. 171 – 77.

Postma, Albert, Izendoorn, R., and De Haan, E.H.F. (1998). Sex differences in object location memory. Brain and Cognition, 36, pp.334 – 45.

Sandstom N.J., Kaufman, J. and Huettel, S.A. (1998). Males and females use different distal cues in a virtual environment navigation task. Cognitive Brain Research, 6, pp. 351 – 60.

Schuman – Hengsteler, R. (1992). The development of visuo-spatial memory: How to remember location. International Journal of Behavioural Development, 15(4), pp. 455 – 471.

Figure Captions:

Appendix I. A two-dimensional representation of the maze subjects were asked to navigate through. An ‘X’ indicates that a particular marker was present in the virtual reality maze.

Appendix II. A two-dimensional representation of the maze subjects were asked to navigate through with no landmarks indicated. Subjects were required to place an ‘X’ where they thought they had seen a landmark in the virtual reality maze.

Appendix III. A two-dimensional representation of the maze subjects were asked to navigate through with landmarks indicated with an ‘X’. Subjects were required to name the colour and shape of each landmark they had seen in the maze and identify the associated ‘X’ on the map.

Appendix IV. The compartments used to analyze the positions-only data.

Figure 1. The mean number of marked positions in the positions only task for each sex. Error bars represent the standard error between subjects.

Figure 2. A. Marked positions for all subjects. Each circle represents one marked position. B. Marked positions by male subjects only. C. Marked positions by female subjects only.

Figure 3. The average distance between the marked positions versus the actual positions by object and sex. Objects 1 & 5 showed the largest difference between the sexes with males identifying the position of object 1 more accurate than females but the reverse for object 5. Overall differences were insignificant.

Figure 4. The mean distance between the marked positions versus the actual positions across objects. Men were slightly better but statistically insignificant.

Figure 5. The average amount of time to complete the maze by sex.

Figure 6. Mean number of correct responses of shape and colour by gender. Females showed significantly better abilities at identifying shape but not colour.

Figure 7. The percent number of correct responses per object for both males and females. Graph shows a definite recency effect and a slight primacy effect.

Figure 8. The number of shapes correctly identified by both males and females according to whether position of the landmark was accurately identified or not. There appears to be no relationship between ability to correctly remember position and to correctly remember shape.

Appendix I