Development of Spatial Ability Tests

In our lab, we design, refine, and validate assessments of visualization abilities. In particular, we design and validate tests in immersive and non-immersive environments.

Our research has demonstrated individual differences in visualization abilities; i.e., dissociation between object and spatial visual abilities (Kozhevnikov, Kosslyn, & Shephard, 2005), and a further distinction between spatial allocentric and spatial egocentric abilities (Kozhevnikov & Hegarty, 2001). Although allocentric and egocentric spatial abilities are correlated, they were also found to have distinguishable characteristics and showed different relationships to real world performance (Kozhevnikov, Motes, Rasch, & Blajenkova, 2006; Kozhevnikov & Hegarty, 2001; Kozhevnikov, Blazhenkova, & Becker, 2010).


Our lab has developed unique tests for assessing egocentric mental transformation ability, the 2D and 3D Perspective-Taking test, which reliably predict spatial navigation performance (i.e., spatial navigation and orientation). This Test is currently used for testing the spatial abilities of navigators, pilots and, also, in Man-Vehicle Laboratory at MIT for predicting astronauts’ spatial orientation skills.

Perspective Taking Test:

Perspective Taking

Furthermore, experimental verification and comparative analysis with 2D non-immersive and immersive 3D Perspective-TakingAbility tests provided experimental evidence that the new 3-D PTA test is the best and unique instrument to measure spatial orientation and spatial navigation abilities.

In addition, we developed 3D immersive tests of allocentric visualization ability (Mental Rotation)
Mental Rotation

Science Learning in 3D immersive VE

This line of research focuses on exploring the strengths and limitations of virtual reality as a medium for learning scientific concepts (i.e., their potential to convey abstract scientific concepts). We investigate how various aspects of virtual realities (multisensory immersion, 3-D representation, shifting among various frames of reference
), when applied to scientific models, might facilitate students’ understanding of abstract phenomena and help in displacing intuitive misconceptions with more accurate mental models.  We also study the role of the interaction between virtual reality’s features and other factors (i.e., learners’ individual characteristics, domain-specific knowledge and interaction experience) in shaping the learning process and learning outcomes.

In particular, we investigate learning of the relative motion concept, using Immersive Immersive 3D Virtual Environment
simulations’ Relative Motion setting, which
explores an innovative instructional technology platform as a new media for learning concepts for introductory physics curriculum for K-12 and higher education. This approach supports the learning process by providing a unique possibility for students to interact with and explore their hypotheses in VR-generated worlds, thus making it possible for students to “experience” what they are learning in an entirely new way. The module includes educationally powerful dynamic visual representations (highly “realistic” objects, visualization of concepts such as forces and velocities, visualization of processes and things invisible to the naked eye, focusing on core-concepts [e.g. highlighting, magnifying, removing irrelevant aspects], a real-time graphing tool, etc.), and allows for real-time interaction. Students can move and look around, point and gesture, experience motion, etc. in a simulation, and these “first hand” experiences can significantly contribute to the sense of “presence” students can feel in a virtual environment.

Neural and Behavioral Correlates of 3D Visual-Spatial Transformations

Our research examines how different environments affect encoding strategies and choice of allocentric versus egocentric frames of reference
. We compared subjects’ performance on allocentric (e.g., Mental Rotation Task
Mental Rotation
) and egocentric (e.g., Perspective-Taking Task
) spatial tasks in non-immersive and immersive 2D vs. 3D environments. Our findings demonstrate a unique pattern of responses in the 3D immersive environment, and suggest that 3D immersive environments are different from 3D non-immersive and 2D environments, and that immersion is necessary to provide adequate information for building a spatial reference frame crucial for high-order motor planning and egocentric encoding. Furthermore, they suggest that non-immersive environments might encourage the use of more “artificial” encoding strategies in which the 3D image is encoded with respect to an environmental frame of reference, and in particular, to the computer screen. On the other hand, immersive environments
can provide the necessary feedback for an individual to use the same strategy and retinocentric frame of reference as he/she would use in a real-world situation.

We plan to further investigate the neural correlates of 3D visual-spatial processing in immersive virtual environments using EEG.

Training in Three-Dimensional Immersive Virtual Environments

Most studies on training imagery skills (either through a particular set of training exercises or indirectly through geometry, chemistry or physics courses) have produced at best small gains in spatial skills and limited transfer of training to a different stimulus set. We suggest that the reason for previous limitations of training visual-spatial abilities using conventional 2D tasks that is that encoding of spatial relations and cognitive strategies applied to perform visual-spatial transformations in 2D non-immersive and 3D immersive environments are different. Thus, in our research we particularly interested in investigating training in immersive 3D virtual environments , and the effects of individual differences in visual imagery ability on training efficacy.

Our research suggests that 3D immersive virtual environments
are critically important for effectively assessing and training large-scale spatial rotation and orientation abilities or any other tasks that might rely on the egocentric spatial system. The perceptual immersivity of an environment seems to be the most important factor in providing information for building an egocentric spatial reference frame needed in performance for real-world, large-scale spatial tasks and higher-order motor planning.

Immersive virtual environments are particularly relevant for training real-world egocentric spatial tasks, such as navigation, teleoperation, or medical surgery, which require visual-spatial processing, due to two major factors: immersion and feedback (Kozhevnikov & Garcia, in press).

In particular, we investigate how distinct visualization abilities
could be improved as a result of training in 3D immersive virtual environments. Our results demonstrate that 3D immersive environments appear to be significantly more efficient for training imagery skills than 2D or 3D non-immersive environments
. Our findings revealed that the 3D Perspective-Taking Test
facilitated a 200% increase in performance (i.e., the rate of error reduction), compared to the non-immersive 2D version of the test.

3D Visualization in Immersive Virtual Environments

Our research on 3D visualization in immersive virtual environments includes the following directions:

We are interested in the contribution of immersion to spatial processing and compare subjects’ performance in non-immersive and immersive 2D vs. 3D environments. Our 3D virtual environment
provides a sensation of immersivity and allowing the participant to move freely while his/her motion is tracked, and interact with the virtual world using a specially designed actuator device.


Recently, more realistic 3D displays have been designed as new, more ecologically valid alternatives to conventional 2D visual displays. However, research has thus far provided inconsistent evidence regarding their contribution to visual-spatial image encoding and transformation. The majority of experimental studies on 3D visual-spatial processing have been conducted using traditional 2D displays. Our research suggests that immersivity is a critical feature of 3D virtual environments for facilitating visual processing and the training of visual ability.