THE USE OF VIRTUAL REALITY IN THE ASSESSMENT OF SPATIAL SKILLS

CONFERENCES
TOPIC: NEUROPSYCHOLOGICAL TESTING

---

Abstract

The assessment of visuospatial ability has received increasing attention.  The impetus for this is two-fold:  1) aspects of visuospatial ability show robust gender differences;  and 2)  spatial performance may provide crucial information in the diagnosis of specific pathological conditions common in aging.  The area of visuospatial performance that shows the most consistent and sizable gender difference is mental rotation (MR), or the ability to turn something over in oneís mind.  Traditional measures used for the assessment of this spatial ability leave uncertainty as to the exact cognitive processes that underlie this ability, germane to both the study of gender differences and the neuropathology of aging.  The most common test uses two-dimensional stimuli that portray three-dimensional objects and requires complete mental processing of the stimuli without any motoric involvement.  We propose that tests of spatial rotation ability that are administered in a virtual reality (VR) environment may prove to be a superior method of assessing spatial cognition.  The use of VR in the assessment of cognitive abilities allows for better standardization of stimulus presentation as well as quantification of multiple characteristics of the stimuli.   Further, responses of the subjects can be quantified on a range of characteristics that cannot be evaluated using traditional psychometric instruments.  The combination of greater control and description of the stimuli along with more precise measurement of responses will allow for characterization of the cognitive processes involved in spatial skills in a more discrete fashion than is possible with standard measures.  Comparison of performance in VR with performance on standard measures offers the potential to better understand this crucial cognitive ability.
 

Background

Virtual Reality in the Assessment of Cognitive Performance: VR technology has undergone a transition in the past few years that has taken it out of the realm of expensive toy, into that of functional technology.  While many VR applications have emerged in the areas of entertainment, education, military training, physical rehabilitation, and medicine, only recently has the considerable potential of VR for the study of human cognitive processes been recognized (Psotka, 1995;  Pugnetti, et al, 1995;  Rizzo, et al, in press).  VR has the potential to create human testing and training environments which allow for precise control of complex stimulus presentations as well as provide accurate records of multiple responses.

Virtual reality can be defined as "...a way for humans to visualize, manipulate, and interact with computers and extremely complex data." (Aukstakalnis & Blatner, 1992).  In the context of cognitive assessment, VR can be viewed as an advanced form of computer interface that allows the user to interact with a computer generated environment in which they are immersed.  This can be accomplished with a variety of existing technologies including head-mounted displays and stereo display screens, which can include auditory inputs. Tracking devices, variously using electromagnetic, ultrasonic, or optical techniques, transmit six degrees of freedom locational information to the computer which updates the images and sounds presented to the user in an effort to create the illusion that she is immersed within a virtually real environment. The user may navigate and interact with objects in the virtual environment through the use of various devices including data gloves, joysticks, wands and voice controllers.  Recent developments have reduced the costs of systems that would have been prohibitive only 5 years ago.  Effective PC-based systems can be built for under $10,000 (Blackburn, 1996) and as with most computing technology, the costs are expected to continue to decrease over time.

Virtual reality technology offers the potential to develop human performance testing environments that could supplement traditional assessment procedures and improve accepted psychometric standards.  The assessment of cognitive functioning hinges on the ability to reliably and validly measure behaviors.  Reliability is the capacity of an instrument to consistently obtain the same results. Validity is concerned with how well an instrument actually measures what it purports to measure. Traditional assessment methodology presents the neuropsychologist with both reliability and validity problems.  Virtual reality offers opportunities for these assessment issues to be addressed. The state-of-the-art in neuropsychological assessment involves, primarily, the use of paper and pencil tests.  The reliability of such tests can be negatively impacted by the fact that the test is administered by different examiners, by differences in the testing environment such as lighting and background noise, and by differences in the quality of the stimuli presented to the subject.  A VR approach could limit these sources of error thus providing for better reliability.

The validity of paper and pencil tests could also be improved with VR technology. The primary means for improving validity would be the ability to measure responses to stimuli more discretely and with greater precision.  The quantification of more discrete behavioral responses will allow for the identification of more specific cognitive domains. The validity of paper and pencil tests is attenuated by the fact that performance requires multiple cognitive functions while the quantification systems used to measure responses does not allow for separation of the different cognitive functions needed to perform the task.  With VR technology, multiple aspects of the subjectsí responses can be quantified.  With appropriate analytic techniques, global cognitive constructs will be refined to more specific cognitive components.

An area of concern in the use of any VR application is the presence of side effects associated with VR exposure.  A commonly reported VR side effect is a form of motion sickness that has been termed "cybersickness".  Symptoms of cybersickness are reported to include nausea, vomiting, eyestrain, disorientation, ataxia, and vertigo (Kennedy, et al, 1994).  Cybersickness is believed to be related to sensory cue incongruity. This is thought to occur when there is a conflict between perceptions in different sense modalities (auditory, visual, vestibular, proprioceptive) or when sensory cue information in the VR environment is incongruent with what is felt by the body or with what is expected based on the userís history of "real world" sensory experience (DiZio & Lackner, 1992). The reported occurrence of cybersickness in virtual environments varies across studies depending upon such factors as:  the type of VR program used, technical drivers (vection, response lag, field of view), the length of exposure time, the person's prior experience using VR, active vs. passive movement, gender, and the construction of the actual self-report rating scale used to assess occurrence (Regan & Price, 1994;  Kennedy, et al, 1992).  In one study, 61% of healthy subjects reported "symptoms of malaise" (i.e., dizziness, stomach awareness, headaches, eyestrain, lightheadedness, and severe nausea) at some point during a 20-min. immersion and 10-minute postimmersion period (Regan & Price, 1994), while VR training for the Hubble telescope repair ground crew suggested low incidence rates, (5%-40%) (Loftin & Kenny, 1995).  Also, the presence of maladaptive after-effects, such as postural ataxia, eye-hand coordination difficulties, and flashbacks, have been reported with VR exposure (Regan & Price, 1994).

Gender and Spatial Rotation
Gender differences in cognitive performance have been identified for several decades and have recently been recognized as crucial in understanding the effects of hormones on the brain.  A case-in-point has been the identification of verbal deficits in women with Alzheimerís disease (AD), which in light of verbal advantages in nondemented women--advantages attributed to the effects of estrogen on the brain--spurred the study of estrogen deficiency as a risk factor for AD among women.  While the study of gender differences in verbal performance has been the focus of much attention, gender differences in cognitive performance are largest on spatial tasks, specifically in the ability to mentally rotate objects (see Figure 1 for an example).  A recent meta-analysis (Voyer, et al, 1995) reports that the average difference (d) between men and women, for the most frequently used psychometric test of mental rotation (MR) was .94, indicating that men perform nearly one standard deviation above the average performance of women.

Figure 1.  Mental rotation stimuli are 3-D block structures shown from different viewpoints

There is considerable debate as to factors that contribute to this difference.  It is argued that Western cultures perceive spatial tasks as masculine in nature, and that differences in spatial ability can be minimized by engendering the perception that spatial tasks are appropriate for female participants (Newcombe, et al, 1983).  The cultural bias against women on spatial tasks has been suggested to be manifested in performance factors that are part of standard tests, namely in poor performance on timed tasks and in a reluctance to guess  (Goldstein, et al, 1990).  The influence of performance factors on MR ability is supported by findings that MR can be modified by computer training.  Groups of college-age men and women both improved on an MR test after performance of a computer game that involved MR.  Interestingly, womenís performance also improved after performance of a computer game that did not involve MR, while mensí performance was unchanged (De Lisi & Cammarano, 1996).  However, several studies suggest that while womenís performance is adversely impacted by performance factors, differences in spatial ability remain (Prinzel & Freeman, 1995;  Delgado & Prieto, 1996;  Stumpf, 1993).  In a compelling study of academically talented students, a spatial battery of tests yielded a strong general spatial factor.  When the effects of this factor were removed in considering the gender effect for each test, gender differences remained significant on mental rotation.  This strongly suggests that the gender difference on mental rotation is not explained by any performance biases inherent in tests of spatial abilities (Stumpf & Eliot, 1995).

Understanding of the role gender and mental rotation has been advanced by studies of how mental rotation mediates performance on another gender sensitive cognitive domain--mathematical ability.  While the literature on a male advantage in math has been equivocal, strong differences emerge when samples are limited to high ability students (Hedges & Nowell, 1995).  In a study of high ability college-bound students, a significant difference was found on the Mathematics Scholastic Aptitude Test (SAT-M).  Using path analysis, 36% of the math difference was mediated by math self-confidence, yet 64% was mediated mental rotation (Casey, et al, 1997).  Using samples of high and low ability students, males outperformed females on both mental rotation and SAT-M in the high ability samples only.  However, when  mental rotation ability was adjusted for, gender differences in SAT-M were eliminated, again suggesting that spatial ability may be needed for high-level mathematical functioning (Casey, et al, 1995).  A recent study of elderly same-sex twin pairs found that estimates of heritability were lower for spatial ability (32%) than for any of the cognitive functions evaluated (general cognitive ability-62%, verbal ability-55%, speed of processing-62%, memory-52%) (McClearn, 1997).  This suggests that environmental and idiosyncratic developmental factors are the primary factors influencing this ability.  Biological explorations of the gender difference include studies of cerebral lateralization.  In a group of young males, regional cerebral blood flow measured with positron emission tomography, while subjects were performing a mental rotation task, found specific activity within the left inferior parietal region and the right head of the caudate nucleus (Alivisatos & Petrides, 1997).  In an attempt to determine if the processing of mental rotation is lateralized differently for men and women, a mental rotation task was centrally presented and tested for lateralized processing by simultaneously presenting a dichotic listening task (Freeman, et al, 1995).  Females did worse when instructed to attend to their left ear, suggesting that the right hemisphere is more involved in mental rotation for women than for men.

The most compelling explanation offered for the gender difference on MR is that it relates to some aspect of the distinct hormonal environments of men and women.  That differences do not emerge until early adolescence is supportive of this argument (Voyer, et al, 1995).  Similar effect sizes for gender are also reported across cultures (Japanese and Canadian) which further argues for a biological explanation that is constant and uniform (Silverman, et al, 1996).  Mann et al., 1990 argue that early exposure to testosterone slows development of the left hemisphere in males and enlarges corresponding areas of the right hemisphere.  This argument is inconsistent with the lateralization study (Freeman, et al, 1995) that reports greater right hemisphere involvement for women.  Levels of testosterone do not show a linear relationship with spatial abilities.  A curvilinear relationship may exist with high testosterone men and low testosterone women performing worse than high testosterone women and low testosterone men (Gouchie & Kimura, 1991).  It is also possible that estrogen acts to suppress MR performance.  An inverse relationship between estrogen levels and MR ability has been found in women based on phases of the menstrual cycle phase (Silvrman & Phillips, 1993).  Levels of estrogen in men are also inversely related to MR performance (Silverman, et al, 1995).

Spatial Ability in Aging and the Diagnosis of  Neuropathological Diseases of Aging
Gender differences in MR ability are reported to be more apparent in elderly subjects than in younger subjects (Dobson, et al, 1995).  (see preliminary studies for supporting data).   The hormonal implications of this may argue against a suppressive effect by estrogen given the loss of estrogen associated with menopause.  Given the gradual decline in testosterone that occurs in both men and women, a greater decline in MR among women remains to be explained.

The diagnosis of age-associated dementia has relied on the assessment of verbal memory.   The criteria for the diagnosis of dementia established by DSM-IV requires the presence of explicit verbal memory deficits.  However, recent studies suggest that spatial deficits may prove useful in the diagnosis of dementias.  In a study of frontotemporal dementia (FTD) and AD, preserved spatial orientation was present in 96.6% of the FTD subjects but only 16% of the AD subjects (Miller, et al, 1997).  Spatial orientation is also reported to be more impaired in AD subjects when compared to similar subjects with vascular dementia (VaD) (Signorino et al, 1996).  In a separate study of AD and VaD subjects, spatial deficits were also found to be more common in the AD group, while spatial deficits in both groups distinguished subjects from nondemented individuals (Gainotti, et al, 1992).  Spatial deficits have been frequently suggested to be a component of Parkinsonís disease (PD).  A matched case-control study of nondemented PD patients reports spatial deficits of PD include MR (Doyon et al, 1996).
 

 
The VR Spatial Rotation (VRSR) Assessment System: The VRSR system uses an ImmersaDesk drafting-table format virtual prototyping device. The Pyramid Systems ImmersaDesk employs stereo glasses and magnetic head and hand tracking. This rear-projection system offers a type of VR that is semi-immersive. It features a 4 X 5-foot rear-projected screen positioned at a 45-degree angle. The size and position of the screen give a wide-angle view and the ability to look down as well as forward.

The VRSR assessment and training system was designed to present a target stimulus that consists of a specific configuration of 3D blocks within a virtual environment. The stimuli appear as "hologram-like" three-dimensional objects floating above the projection screen. After presentation of a target stimuli, the participant is presented with the same set of blocks (working stimuli) that needs to be rotated to the orientation of the target and then superimposed within it. The participant manipulates the working stimuli by grasping and moving a sphere shaped "cyberprop" which contains a tracking device. The motion of the sphere is imparted upon the working stimuli. Upon successful superimposition of the working stimuli and target objects a "correct" feedback tone is presented and the next trial begins. The new control object appears attached to the sphere (userís hand), and a new target appears. In this mode of interaction, users do not need to press any buttons or select any objects. Control objects simply appear attached to the sphere for users to manipulate.
We calculate the following information on the stimuli. The orientation of a stimulus can be represented by a single linear rotation around a three-dimensional vector. We specify orientation for each stimulus as a group of four values: three defining a three-dimensional vector of unit length, and an angle in degrees. We define magnitude as the angular difference between the two orientations. The stimulus orientation can be aligned to the target orientation by a single minimal rotation around some fixed three-dimensional vector. The degrees required to do this is the magnitude of the rotational task, ranging from 0 to 360 degrees. If the rotational axis is nearly parallel to the viewerís line of sight, rotations of the object will not reveal new faces to the viewer and the task is equivalent to a 2D rotational task. As the axis is moved to become parallel to the view-plane, the task requires a fully 3D understanding of the objectís appearance. This is calculated as the sine of the angle between the rotational axis and the line of sight, and so varies from 0 to 1.

We presently have assessed rotational ability by recording the amount of time to complete the rotation as well as the efficiency of the solution. The most efficient execution of a rotational task follows the shortest angular path from stimulus to target, about a fixed axis. Samples are taken at regular intervals during task execution, defining an angular path as the subject searches for a solution. Summing the angular differences between sequential samples gives the angular length of that path, and calculating the ratio of the shortest possible path to this summed length give the efficiency of the task execution. This value varies from 0 (poor) to 1 (ideal).

Pilot Study:
Subjects: Sixty subjects (26 males and 34 females) between the ages of 18-40 were tested.  Subjects included employees recruited at the Information Sciences Institute of the University of Southern California, graduate students from the Fuller Graduate School of  Psychology, and undergraduate students from the University of Southern California and California State University at Los Angeles.

The experimental sessions took place over a two hour period.  After informed consent was obtained, basic demographic information, computer experience and usage, and spatial activities history were recorded.  Next, a baseline measure of mental rotation ability is assessed using a redrawn version of the Mental Rotation Test (MRT-A) of Vandenberg & Kuse (1978), a twenty item, 2-dimensional paper and pencil task.  Subjects then completed a comprehensive neuropsychological battery administered under standard conditions.  Following the completion of the neuropsychological battery, subjects completed the Motion History Questionnaire (Kennedy and McCauley, 1984) and Simulator Sickness Questionnaire (Kennedy et al, 1993), which includes a pre-VR exposure symptom checklist.  Experimental subjects then participated in the fifteen minute VR task that both assesses and trains mental rotation abilities.  After five non-rotational practice trials, each subjectís VR spatial rotation baseline performance is assessed over twenty trials using a VR version of the items from the pencil and paper MRT.  Next, one hundred training trials of increasing stimulus complexity are administered.  After a one-minute break, the original twenty VR MRT trials are administered again to measure changes in VR spatial rotation ability.  Control subjects are given a filler task (crossword puzzle) of matching duration instead of the VR exposure.  The Simulator Sickness Questionnaire, which contains a post-VR exposure symptom-checklist, is then given to each subject.  Finally, an alternate form of the paper and pencil MRT is administered to assess changes in mental rotation performance.

Testing Instruments: The neuropsychological battery included a diverse collection of instruments.  Mental rotation ability is assessed using the Mental Rotation Test.  This test uses line drawings of block stimuli and consists of two ten-item sections in which the subject is required to match two of the four choices to a target figure.  Incorrect choices are mirror images of the target or alternative block configurations.  Standard administration provides for a five-minute time limit.  The alternate form of the MRT uses the same drawings but reorders their presentation and switches position of the target stimuli.  Verbal attention and mental control is assessed with the Digit Span Forward and Backward test from the Wechsler Adult Intelligence Scale-Revised.  Visuoconstruction abilities are measured by the Block Design subtest of the WAIS-R.  The Trail-Making Tests A and B are used to evaluate executive control processes and attention.  The Judgment of Line Orientation test is used to evaluate visuoperceptual skills.  The California Verbal Learning Test is employed to assess verbal learning and memory.  Nonverbal memory is evaluated by the Visual Reproduction subtest of the Wechsler Memory Scale-Revised.  These tests are all commonly used for neuropsychological assessment of these cognitive processes and as such have widely used normative information available.  Finally, surveys of simulator sickness and motion sickness history are administered.

Data Analysis and Research Questions: We collected data from a variety of domains.  These include: 1)  Neuropsychological performance on tests of cognitive functioning (attention, verbal and visual memory, visuospatial abilities, etc.); 2)  Demographic factors (education, gender, reproductive history, etc.); 3)  Spatial Activity History (a self-report scale of participation in everyday activities that contain spatial components); 4)  Computer Usage History Questionnaire (a self-report measure that we developed to assess computer use, programming activities, use of computer games, etc.); 5)  Side effects assessment; and 6)  VE data including all movement digitized in real time and allowing for playback of each response. While we anticipate developing more sophisticated analytic techniques, we are currently analyzing time to completion per trial, path efficiency, and various compiled measures, such as the total time for the first twenty VRSR items vs. the last twenty.
From this data, we attempt to answer the following research questions:
1. Are there negative side effects associated with the use of the VRSR?
2. Does performance on the VRSR system demonstrate adequate psychometric properties?
  a. Is the VRSR reliable in terms of internal consistency (coefficient a)?
b. Does the VRSR demonstrate concurrent validity in the pattern of associations observed with standard neuropsychological tests?
3. Do the same sex differences that appear on the pencil and paper MRT appear on VRSR performance?
4. Does VRSR performance improve with practice (100 training trials) as seen by comparing twenty pre-training VR MRT items with twenty identical post-training VR MRT items (intra-method generalization)?
5. Does VRSR training improve post-training pencil and paper MRT performance.

Results:

Side Effects: A split plot factorial ANOVA indicated that the interaction between group (VRSR and control) and occasion (pre and post testing) was significant for the amount of side effects reported.  While the trend was for the VRSR group to report increased side effects, the trend for the control group to report fewer side effects also contributed to this interaction.  Post hoc analyses of the VRSR group found that the only item where there was a significant increase at post testing was blurred vision.

Reliability: The time to complete items showed good internal consistency, as analyzed by coefficient a, at both pre-testing (a= .65) and post-testing (a = .83).  The efficiency ratio showed good reliability at pre-testing (a = .65) and acceptable reliability at post-testing (a = ..68).  The reliability of the time to complete the VRSR meets standard psychometric criteria.

Concurrent Validity: Pearson Product-Moment correlations  between the VRSR time to complete and all standard measures of neuropsychological functioning yielded a number of statistically significant effects.  It was highly correlated with the Efficiency Index (r = -.76, p < .001) and with the paper and pencil MRT (r = -.45, p = .01).  It also correlated significantly with tests of visual memory under both immediate (r = -.50, p = .006) and delayed (r = -.48,  = .008) conditions.  There was a significant association with visual attention as measured by Trails A (r = .38, p = .04).  There were also strong correlations with two measures of executive functioning one that includes a strong visuoconstructional component (Trails B; r = .46, p = .01, Block Design; r = -.64, p < .001).  Surprisingly, time on VRSR also was associated with aspects of verbal learning notably the consistency of items recalled over the 5 trials of the CVLT (r = -.52, p = .005) and the number of perseverations, or times when they repeated the same word.  These findings may relate more to the ability to maintain concentration when presented with a large amount of new information (working memory) than to verbal memory per se.  (Note that the direction of all correlations is such that slower completion time is associated with worse performance on each test) Correlations between the Efficiency Index and other neuropsychological tests were minimal.  As reported above, Efficiency Index  did correlate significantly with time to complete the VRSR.  It also correlated significantly with one executive functioning test (Block Design) which requires manipulation of physical blocks.

A comparison of associations between the paper and pencil MRT with the other neuropsychological tests provides a useful reference point for interpreting the above correlations.  The tests that correlated with the MRT are generally very consistent with the tests that correlated with the VRSR time to complete with one notable exception.  While performance on the JLO was not associated with the VRSR, it was strongly correlated with the MRT.  The JLO is a two-dimensional task that evaluates the ability to perceive spatial orientation.  That it would be associated with the ability to mentally rotate two dimensional portrayals of 3D objects and not with the ability to physically manipulate 3D virtual objects suggests that one of the major cognitive components underling ability on the MRT may relate to the personís ability to construct and manipulate 3D images from two-dimensional perception. The concurrent validity of the time to complete the VRSR again meets standard psychometric criteria.  It significantly correlates with other neuropsychological tests that assess domains of performance that should be related to the ability to perform the VRSR, namely visual memory and working memory/executive (Note that the executive function test where we see the largest involves physical manipulation of blocks).  The concurrent validity of the Efficiency Index is questionable, however there are again suggestions that it warrants further exploration. The Efficiency Index allows for precision of measurement of a response to a cognitive stimuli that far exceeds any of the neuropsychological tests with which it is compared.  Thus it may be evaluating a more discrete cognitive domain than the other tests. The fact that it did correlate significantly with time to completion and Block Design, which is clearly the most similar task, does suggest that it may be measuring some valid cognitive domain.

Sex Differences on Rotational Tasks: Men scored significantly better on the MRT given before the VRSR (p < .04).  There were no differences between men and women on either the time to complete or the Efficiency Index of the VRSR (pís > .8).  Interestingly, the difference between men and women on the MRT after completing the VRSR was no longer significant. The existence of gender differences on the MRT is well established but the mechanism for this difference is not identified.  That women can manipulate three-dimensional objects as efficiently as men while they cannot visualize the same process as well has potentially profound implications for understanding brain functioning as well as designing effective computer interfaces.

VRSR and Training/Transfer Issues: Subjects showed significant improvement on the VRSR after completing 100 training trials for both time to complete (p < .001) and Efficiency (p = .03).  Subjects in the VRSR group showed a nonsignificant trend toward improved performance on the MRT (p < .06).  However, when the changes in MRT performance between the VRSR and control group were compared by utilizing a split plot factorial ANOVA, the interaction between group and change over the two testing occasions was nonsignificant.  This indicates that there is a general practice effect for the MRT and that VRSR exposure does not have a specific effect on improving performance among all subjects. However, of particular interest, is how individuals who have relatively poor initial MRT scores perform after VRSR exposure.  Our sample of subjects is notable for performing much better on the MRT than do broader populations.  If rotational skills can be trained, it would seem likely that individuals with high existing levels of rotational ability would be less likely to show improvement than individuals with less ability.  To directly test this, we divided subjects into groups, based on the MRT scores at the pre-testing. We used a cutpoint of 20 (out of a possible 40) to create a group of subjects with scores closer to those reported in other studies.  Again using a split-plot factorial design, we found a significant (p <.02) interaction between group (VRSR and control), MRT group (£ 20, > 20) and occasion (pre and post MRT) such that those low scorers on the MRT who were in the VRSR group improved significantly more than other groups.  This leads to the intriguing suggestion that rotational skills can be trained when they are relatively poor to begin
with. This could have implications for cognitive rehabilitation strategies aimed at impaired populations.

Summary and Future Plans: We interpret these findings as an indication that the VRSR has good potential as a test of cognitive ability.  Side effects do not appear to be a major concern.  We have identified two variables (time to completion and efficiency) that are generally reliable and valid.  Our experimental results have broad ranging implications, from better defining the nature of sex differences in cognition to suggesting the possibility of efficient training of fundamental cognitive abilities. Our future work will focus on two areas.  We will continue to explore the nature of gender differences in rotational skills.  To determine if womenís disadvantage on paper and pencil tests of rotation stem from the two-dimensional nature of the presentation, we will present a mental rotation task in a virtual environment, both in two-dimensions and three dimensions. We are also completing an NIA funded project to evaluate non demented elderly individuals on the VRSR.  This will be one of the first studies to assess performance and address the feasibility concerns of exposing elderly subjects (to these types of virtual environments) to an immersive environment.  Our longer range goal is to develop a comprehensive cognitive assessment system to be administered in immersive virtual environments.  We argue that these environments have the potential to provide unparalleled improvements in the assessment of cognition and may also prove useful in the training or rehabilitation of cognitive abilities.  Additonally, information gained through the systematic evaluation of cognitive abilities in immersive virtual environments should prove invaluable in designing more effective human computer interfaces.
 

Alivisatos B, Petrides M.  Functional activation of the human brain during mental rotation.  Neuropsychologia  1997;35:111-118.

Aukstakalnis S, Blatner D.  Silicon Mirage: The Art and Science of Virtual Reality. Peachpit Press:Berkeley CA. 1992, p.7

Blackburn D. The six thousand dollar system. Virtual Reality: Special Reports  1996;3:22-24.

Casey MB, Nuttall RL, Pezaris E.  Mediators of gender differences in mathematics college entrance test scores:  a comparison of spatial skills with internalized beliefs and anxieties.  Developmental Psychol  1997;33:669-680.

Casey MB, Nuttall R, Pezaris E, Benbow CP.  The influence of spatial ability on gender differences in mathematics college entrance test scores across diverse samples.  Develop Psychol  1995;31:697-705.

De Lisi R, Cammarano DM.  Computer experience and gender differences in undergraduate mental rotation performance.  Computers H um Behav  1996;12:351-361.

Di Zio P, Lackner JR.  Spatial orientation, adaptation, and motion sickness in real and virtual environments. Presence: Teleoperators and Virtual Environments 1992;1: 323

Delgado AR, Prieto G.  Sex differences in visuospatial ability:  do performance factors play such an

Dobson SH, Kirasic KC, Allen G L.  Age-related differences in adultsí spatial task performance:  influences of task complexity and perceptual speed.  Aging Cognition  1995;2:19-38.

Doyon J, Bourgeois C, Bedard P.  Visual-spatial deficits of Parkinsonís disease  J Inter Psychologie  1996;31:161-175.

Freeman FG, Pearson G, Evans D.  Lateralized processing of centrally presented verbal and spatial tasks:  effects of attention, gender, and hand used.  Neuropsychiatry Neurosychol Behav Neurol  1995;8:88-98.

Gainotti G, Parlato V, Monteleone D, Carlomagno S.  Neuropsychological markers of dementia on visual-spatial tasks:  a comparison between Alzheimerís type and vascular forms of dementia.  J Clin Exper Neuropsychol  1992;14:239-252.

Goldstein D, Haldane D, Mitchell C.  Sex differences in visual-spatial ability:  the role of performance factors.  Memory Cognition  1990;18:546-550.

Gouchie C, Kimura D.  The relationship between testosterone levels and cognitive ability patterns.  Psychoneuroendocrin  1991;16:323-324.

Hampson E.  Estrogen-related variations in human spatial and articulatory-motor skills.  Psychoneuroendocrinol 1990;15:97-111

Hedges LV, Nowell A.  Sex differences in mental scores, variability, and numbers of high scoring individuals.  Science  1995;269:41-45.

Kampen DL, Sherwin BB.  Estrogen use and verbal memory in healthy postmenopausal women.  Obstet Gynecol  1994;83:979-983

Kennedy RS, Berbaum KS, Drexler J.  Methodological and measurement issues for identification of engineering features contributing to virtual reality sickness. Paper presented at: Image 7 Conference. Tucson, AZ. 1994.

Kennedy RS, Lane NE, Lilienthal MG, Berbaum KS, Hettinger LJ. Profile analysis of simulator sickness symptoms: Application to virtual environment systems. Presence: Teleoperators and Virtual Environments. 1992;1:295-301.

Kimura D.  Sex differences in the brain.  Scientific American  1992;267(3):119-125.

Loftin RB, Kenny PJ.  Training the Hubble Space Telescope Flight Team. IEEE Computer Graphics and Applications. 1995:31-37.

Mann VA, Sasanuma S, Sakuma N, Masaki S.  Neuropsychologia  1990;28:1063-1077.

McClearn GE, Johansson B, Berg S, Pederson NL, Ahern F, Petrill SA, Plomin R.  Substantial genetic influence on cognitive abilities in twins 80 or more years old.  Science  1997;276:1560-1563.

Miller BL, Ikonte C, Ponton M, Levy M, Boone K, Darby A, Berman N, Mena I, Cummings JL.  A study of the Lund Manchester research criteria for frontotemporal dementia:  clinical and single-photon emission CT correlations.  Neurology  1997;48:937-942.

Newcombe N, Bandura MM, Taylor DG.  Sex differences in spatial ability and spatial activities.  Sex Roles  1983;9:377-386.

Prinzel LJ III, Freeman FG.  Sex differences in visuospatial ability:  task difficulty, speed-accuracy tradeoff, and other performance factors.  Canadian J Exper Psyhcol  1995;49:530-539.

Psotka J.  Immersive training systems: Virtual reality and education and training.  Instructional Science  1995; 23:405-431.

Pugnetti L, Mendozzi L, Motta A, Cattaneo A, Barbieri E, Brancotti S.  Evaluation and retraining of  adults' cognitive impairments:  which role for  virtual reality technology?  Computers in Biology and Medicine  1995;25:213-227.

Regan E, Price KR. The frequency of occurrence and severity of side-effects of immersion virtual reality.  Aviation, Space, & Environmental Medicine. 1994; 65:527-530.

Rizzo AA, Buckwalter JG, Neumann U, Kesselman C, Thiebaux M.  Basic issues in the application of virtual reality for the assessment and rehabilitation of cognitive impairments and functional disabilities. CyberPsychology and Behavior In press.

Sherwin BB.  Estrogen and/or androgen replacement therapy and cognitive functioning in surgically menopausal women.  Psychoneuroendocrinol  1988;13:345-357

Signorino M, Pucci E, Brizioli E, Cacchio G, Nolfe G, Angeleri F.  EEG power spectrum typical of vascular dementia in a subgroup of Alzheimer patients.  Arch Gerontol Geriatr  1996;23:139-151.

Silverman I, Phillips K, Silverman LK.  Homogeneity of effect sizes for sex across spatial tests and cultures:  implications for hormonal theories.  Brain Cognition  1996;31:90-94

Silverman I, Phillips K.  Effects of estrogen changes during the menstrual cycle on spatial performance.  Ethology Sociobiol  1993;14:257-270.

Silverman I, Choi J, Kastuk D, Phillips K, Hampson E.  Sex hormone levels and cognitive abilities in males.  1995, Presented at the meeting of the Human Behavior and Evolution Society, Santa Barbara, CA.

Stumpf H.  Performance factors and gender-related differences in spatial ability:  another asssessment.  Memory and Cognition  1993;21:828-836.

Stumpf H, Eliot J.  Gender-related differences in spatial ability and the k factor of general spatial ability in a population of academically talented students.  Person Indiv Diff  1995;19:33-45.

Vandenberg SG, Kuse AR.  Mental rotations, a group test of three-dimensional spatial visualization.  Perceptual Motor Skills  1978;47:599-604.

Voyer D, Voyer S, Bryden MP.  Magnitude of sex differences in spatial abilities:  a meta-analysis and consideration of critical variables.  Psychol Bull  1995;117:250-270.

 

 
 
 
 
 
 

---