Brain mechanisms of ego- and allocentric representations in VR-based imagined and actual navigation
Poster No:
961
Submission Type:
Abstract Submission
Authors:
Jie Song1, Emmanuel Badier1, Ilaria Sani1, Patrik Vuilleumier1
Institutions:
1University of Geneva, Geneva, Switzerland
First Author:
Co-Author(s):
Introduction:
Every day we need to navigate and be aware of our location in space. Previous studies suggested that brain mechanisms of actual navigation operate using two different reference frames - egocentric (orienting based on a frame of reference centered on the self) and allocentric (orienting based on the spatial relationship between landmarks) [1]. However, it is still unclear whether and how we rely on different strategies in imagined navigation during route planning, and how brain mechanisms differ between imagined and actual navigation in virtual reality (VR) environments.
Methods:
Participants: We recruited 19 healthy participants (7M/12F, age 23-30; 9 for behavior only; 10 for fMRI; inclusion criteria: no history of psychological or neurological disease and with normal or corrected vision; ethics approval # 2023-00325).
fMRI: MRI data were obtained on a 3T Siemens Prisma MRI scanner with a 32-channel head coil. The functional data were obtained using GE-EPI sequence with TR=1300 ms, TE= 30 ms, FA = 64°, FoV=210 × 224 mm2, image matrix=84 × 84, slice thickness = 2.5 mm, voxel size = 2.5 mm, and 84 sagittal slices covering the whole brain.
VR tasks: We leveraged two VR navigational tasks [2] with encoding, imagination, and retrieval phases to assess participants' ability to repeat/retrace a route in the same/reverse directions.
In encoding phase, participants started from a black car and moved passively through 3 street intersections until a red telephone box (Fig. 1). Different intersections featured different houses.
In imagination phase, participants were positioned at one intersection/car and asked to imagine the route to the phone box (repetition task) or at one intersection/phone box to imagine the route back to the car (retracing task).
In retrieval phase, participants passively moved to one intersection where they chose to turn left, turn right, or proceed straight to reach their goal. After their response, participants were always moved to the correct direction and received feedback for wrong responses.
Behavioral data were collected with (in scanner) and without the imagination phase.
Data analysis: Behavioral data were analyzed in terms of accuracy and reaction times. Functional MRI data were preprocessed using fMRIprep 20.2.2 with slice-timing correction, co-registration, normalization, and 6 mm FWHM smoothing. We used a generalized linear model (GLM) to analyze main effects of egocentric (navigate from car/phone box) and allocentric (navigate from intermediate intersection) conditions, for both imagination (GLM1) and retrieval (GLM2) phases. Motion parameters, global signal, and top 10 anatomical components were included as nuisance variables. Group-level statistics were estimated from individual z maps.
fMRI: MRI data were obtained on a 3T Siemens Prisma MRI scanner with a 32-channel head coil. The functional data were obtained using GE-EPI sequence with TR=1300 ms, TE= 30 ms, FA = 64°, FoV=210 × 224 mm2, image matrix=84 × 84, slice thickness = 2.5 mm, voxel size = 2.5 mm, and 84 sagittal slices covering the whole brain.
VR tasks: We leveraged two VR navigational tasks [2] with encoding, imagination, and retrieval phases to assess participants' ability to repeat/retrace a route in the same/reverse directions.
In encoding phase, participants started from a black car and moved passively through 3 street intersections until a red telephone box (Fig. 1). Different intersections featured different houses.
In imagination phase, participants were positioned at one intersection/car and asked to imagine the route to the phone box (repetition task) or at one intersection/phone box to imagine the route back to the car (retracing task).
In retrieval phase, participants passively moved to one intersection where they chose to turn left, turn right, or proceed straight to reach their goal. After their response, participants were always moved to the correct direction and received feedback for wrong responses.
Behavioral data were collected with (in scanner) and without the imagination phase.
Data analysis: Behavioral data were analyzed in terms of accuracy and reaction times. Functional MRI data were preprocessed using fMRIprep 20.2.2 with slice-timing correction, co-registration, normalization, and 6 mm FWHM smoothing. We used a generalized linear model (GLM) to analyze main effects of egocentric (navigate from car/phone box) and allocentric (navigate from intermediate intersection) conditions, for both imagination (GLM1) and retrieval (GLM2) phases. Motion parameters, global signal, and top 10 anatomical components were included as nuisance variables. Group-level statistics were estimated from individual z maps.
Results:
Behavior: Orienting choice RTs were longer (indicating higher cognitive cost) on tasks without imagination phase (Fig. 2a) compared to tasks with imagination (Fig. 2b), which suggests participants benefited from imagined navigation. Interestingly, the farther from the navigation start (intersection 3 > 2 > 1), the longer were the RTs to identify the location and choose the direction (Fig. 2a).
fMRI Results: During imagined navigation, brain activation was higher in parahippocampal gyrus (PHC), bilateral lingual gyrus, and retrosplenial cortex (RSC) in allocentric conditions, and higher in right superior parietal cortex (SPC) in egocentric conditions (Fig. 2c, GLM1). During actual navigation, activation was stronger in anterior and posterior cingulate gyrus in allocentric conditions, and in caudate and occipital cortex in egocentric conditions (Fig. 2d, GLM2).
fMRI Results: During imagined navigation, brain activation was higher in parahippocampal gyrus (PHC), bilateral lingual gyrus, and retrosplenial cortex (RSC) in allocentric conditions, and higher in right superior parietal cortex (SPC) in egocentric conditions (Fig. 2c, GLM1). During actual navigation, activation was stronger in anterior and posterior cingulate gyrus in allocentric conditions, and in caudate and occipital cortex in egocentric conditions (Fig. 2d, GLM2).
Conclusions:
Our findings suggest that planning a route to the goal in the imagination phase could benefit to actual navigation. Allocentric navigation mainly relied on PHC, RSC, lingual, and cingulate cortex, while egocentric mainly relied on SPC and caudate. Imagined and actual navigation using both ego and allocentric strategies may engage different brain mechanism.
Higher Cognitive Functions:
Space, Time and Number Coding 2
Higher Cognitive Functions Other 1
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Keywords:
Cognition
Other - Spatial Navigation
1|2Indicates the priority used for review
Provide references using author date format
2. Wiener, J.M., et al. (2020), 'A novel virtual-reality-based route-learning test suite: Assessing the effects of cognitive aging on navigation', Behav Res Methods, vol. 52, no. 2, pp. 630-640