When we remember a city that we have visited, we retrieve places related to finding our goal but also non-target locations within this environment. Yet, understanding how the human brain implements the neural computations underlying holistic retrieval remains unsolved, particularly for shared aspects of environments. Here, human participants learned and retrieved details from three partially overlapping environments while undergoing high-resolution functional magnetic resonance imaging (fMRI). Our findings show reinstatement of stores even when they are not related to a specific trial probe, providing evidence for holistic environmental retrieval. For stores shared between cities, we find evidence for pattern separation (representational orthogonalization) in hippocampal subfield CA2/3/DG and repulsion in CA1 (differentiation beyond orthogonalization). Additionally, our findings demonstrate that medial prefrontal cortex (mPFC) stores representations of the common spatial structure, termed schema, across environments. Together, our findings suggest how unique and common elements of multiple spatial environments are accessed computationally and neurally.

Research into the behavioral and neural correlates of spatial cognition and navigation has benefited greatly from recent advances in virtual reality (VR) technology. Devices such as head-mounted displays (HMDs) and omnidirectional treadmills provide research participants with access to a more complete range of body-based cues, which facilitate the naturalistic study of learning and memory in three-dimensional (3D) spaces. One limitation to using these technologies for research applications is that they almost ubiquitously require integration with video game development platforms, also known as game engines. While powerful, game engines do not provide an intrinsic framework for experimental design and require at least a working proficiency with the software and any associated programming languages or integrated development environments (IDEs). Here, we present a new asset package, called Landmarks, for designing and building 3D navigation experiments in the Unity game engine. Landmarks combines the ease of building drag-and-drop experiments using no code, with the flexibility of allowing users to modify existing aspects, create new content, and even contribute their work to the open-source repository via GitHub, if they so choose. Landmarks is actively maintained and is supplemented by a wiki with resources for users including links, tutorials, videos, and more. We compare several alternatives to Landmarks for building navigation experiments and 3D experiments more generally, provide an overview of the package and its structure in the context of the Unity game engine, and discuss benefits relating to the ongoing and future development of Landmarks.

Older adults typically perform worse on spatial navigation tasks, although whether this is due to degradation of memory or an impairment in using specific strategies has yet to be determined. An issue with some past studies is that older adults are tested on desktop-based virtual reality: a technology many report lacking familiarity with. Even when controlling for familiarity, these paradigms reduce the information-rich, three-dimensional experience of navigating to a simple two-dimensional task that utilizes a mouse and keyboard (or joystick) as means for ambulation. Here, we utilize a wireless head-mounted display and free ambulation to create a fully immersive virtual Morris water maze in which we compare the navigation of older and younger adults. Older and younger adults learned the locations of hidden targets from same and different start points. Across different conditions tested, older adults remembered target locations less precisely compared to younger adults. Importantly, however, they performed comparably from the same viewpoint as a switched viewpoint, suggesting that they could generalize their memory for the location of a hidden target given a new point of view. When we implicitly moved one of the distal cues to determine whether older adults used an allocentric (multiple landmarks) or beaconing (single landmark) strategy to remember the hidden target, both older and younger adults showed comparable degrees of reliance on allocentric and beacon cues. These findings support the hypothesis that while older adults have less precise spatial memories, they maintain the ability to utilize various strategies when navigating.

An important question regards the nature of our spatial memories for the paths that we have walked and, in particular, whether such distortions might violate the topological properties of the shape of the paths (i.e., creating an intersection when two paths did not intersect or vice versa). To investigate whether and how this might occur, we tested humans in situations in which they walked simple paths and idiothetic and visual cues either matched or mismatched, with the mismatching cues creating the greatest potential for topological distortions. Participants walked four-segment paths with 90° turns in immersive virtual reality and pointed to their start location when they arrived at the end of the path. In paths with a crossing, when the intersection was not presented, participants pointed to a novel start location suggesting a topological distortion involving non-crossed paths. In paths without a crossing, when a false intersection was presented, participants pointed to a novel start location suggesting a topological distortion involving crossed paths. In paths without crossings and without false intersections, participants showed reduced pointing errors that typically did not involve topological distortions. Distortions more generally, as indicated by pointing errors to the start location, were significantly reduced for walked paths involving primarily idiothetic cues with limited visual cues; conversely, distortions were significantly increased when idiothetic cues were diminished and navigation relied primarily on visual cues. Our findings suggest that our spatial memories for walked paths sometimes involve topological distortions, particularly when resolving the competition between idiothetic and visual cues.

No abstract

Spatial navigation deficits in older adults are well documented. These findings are based on experimental paradigms that require using a joystick or keyboard to navigate a virtual desktop environment. In the present study, cognitively normal young and older adults navigated in each of two virtual reality (VR) conditions: a desktop VR condition which required using a mouse and keyboard to navigate and an immersive and ambulatory VR condition which permitted unrestricted locomotion. Consistent with past studies, older adults navigated to target locations less precisely than did younger individuals in the desktop condition. These age differences were significantly attenuated when tested in the immersive environment. Additional analyses indicated that older adults showed a preference for route-based search strategies compared to young adults, regardless of condition. These findings suggest that certain aspects of navigation performance in older adults are improved in paradigms that offer a fuller range of enriched and naturalistic cues.

Previous research has demonstrated that humans combine multiple sources of spatial information, such as allothetic and idiothetic cues, while navigating through an environment. However, it is unclear whether this involves comparing multiple representations from multiple sources during encoding (parallel hypothesis) or primarily accumulating idiothetic information until the end of the navigation to integrate with allothetic information (serial hypothesis). We tested these two hypotheses in an active navigation task using mobile scalp EEG recordings. Participants walked through an immersive virtual hallway with or without conflicts between allothetic and idiothetic cues and pointed toward the starting position of the hallway. By analyzing the scalp oscillatory activities during the navigation phase, we found that path segments including memory anchors -- such as path intersections -- were more strongly associated with the pointing error, regardless of when they appeared during encoding. This indicates that the integration of spatial information of a walked path likely begins in the early stages of navigation rather in late stages alone, supporting the parallel hypothesis. Furthermore, theta oscillations in frontal-midline regions during active navigation were related to memory of the path rather than only movement through the path, supporting a mnemonic role of theta oscillations.

No abstract