The brain’s GPS for an imaginary world

“I’ll meet you outside in 10 minutes,” you text your friend as you look forward to catching up over lunch. As you message her, you envision yourself turning left towards her office building from your front door, to watch for her to arrive. This sense of direction is not just critical for navigating through the world, but also for imagining or planning interactions with the environment, like when planning to meet a friend. Research in rodents has pinned down several types of neurons that support this neural “GPS,” including cells that sense where in space the animal is, or the direction they’re facing. While recent studies suggest that humans may have similar brain circuitry that supports a sense of direction, the details of such a system are not fully understood. A new functional MRI (fMRI) study published in eLife now reports that humans have neurons that detect imagined spatial directions, and that these cells closely resemble the spatially tuned “grid cells” of the rodent brain.

Over the past decade, countless studies have contributed to our knowledge of spatial coding in the rodent brain. These investigations have revealed hippocampal place cells tuned to particular locations, entorhinal cortex grid cells that are spatially tuned with a radially symmetric pattern of the environment, and more specialized cells such as head direction cells, goal-directed cells and border cells. These various cell types are thought to support an animal’s navigation, integrating information about where it is in space with where it’s headed. More recently, intracranial and fMRI studies have shown that the human brain similarly possesses neurons tuned to location during virtual navigation. However, these studies haven’t clarified the full functional range of these cells; for instance, beyond subserving our ability to move through the world, might they also support our more fundamental sense of direction, similar to the ‘head direction’ neurons in rodents? Could this directional tuning extend to our internal mental map of the world, to guide planning and anticipating the future?

Exploring the virtual “Donderstown”

To test these questions, Jacob Bellmund and colleagues put 24 young adults to the test, performing fMRI of the participants’ brains while they imagined their locations in a virtual world. Participants first explored the virtual European city “Donderstown” until they were familiar with its layout and its 18 buildings. They were trained to learn the names and locations of these buildings, which were spaced evenly across 12 directions from a reference point (Figure 1). During fMRI scanning, participants were asked to imagine being at a start building and to indicate the direction of a target building.

Direction-tuned brain responses

The researchers performed multi-voxel pattern analysis of the fMRI data – which assesses patterns in the single-voxel signal – in several brain regions implicated in spatial processing, including the parahippocampal gyrus, entorhinal cortex and retrosplenial cortex. They found an area of the parahippocampal gyrus in which brain activity patterns were more similar when participants imagined similar directions (≤ 30°) than when they imagined dissimilar directions (≥ 60°). Thus, this parahippocampal region seemed to be tuned to the absolute direction of imagined locations. Importantly, they also performed several control analyses to ensure that the similarity in brain responses to the directions was not an artifact of imagining the same start or target building, of the distance between buildings, or of the visual features of the imagined scene.

Grid-like representations of direction

Next they tested the follow-up hypothesis that the human brain shows grid-like signaling of imagined direction, similar to grid cell responses to spatial location. Specifically, they looked for fMRI signal showing six-fold rotational symmetry; i.e., similar activity patterns when participants imagined directions separated by 60° (directions of 60°, 120°, 180°, etc.). They found such a pattern in the posterior, medial portion of the entorhinal cortex, which corresponds to the rodent entorhinal cortex area containing grid cells. Notably, there was no activity that systematically related to cardinal directions or to other directional spacing (e.g., four-fold symmetry). Although the study focused on the entorhinal cortex since animal studies have identified the most abundant number of grid cells here, they also observed similar grid-like patterns in parietal and occipital cortex. Bellmund explains that “A model of spatial memory and mental imagery suggests that allocentric spatial representations in the medial temporal lobe might drive egocentric parietal representations. This is a speculative interpretation of why we observed the 60°-symmetric activation patterns in other brain areas as well in our exploratory whole-brain analysis. Previous studies investigating grid-like coding in the human brain have also reported hexadirectional activity beyond the entorhinal cortex, for example in parietal and medial prefrontal cortices.”

Extrapolating from animals to humans

These findings add critically to mounting evidence for a homologous human neural system of spatial signaling to that so well characterized in rodents. Using intracranial recordings in humans during virtual navigation, place cells were discovered in the human hippocampus, and shortly after, grid cells were observed in the entorhinal cortex. FMRI studies supported these findings, reporting grid-like activity in the entorhinal cortex during virtual navigation, and even during imagined navigation. Given our wealth of knowledge about the neural circuits supporting spatial processing in rodents, such human studies are crucial for validating their extrapolation to humans.

Bellmund and colleagues take this evidence one step further to show parahippocampal responses and grid-like entorhinal cortex activation to an imagined direction–in the absence of real or virtual sensory stimulus, or any real or imagined movement. This neural system may therefore serve broader functions than simply helping us move through our world or sense where we are in space; spatially-tuned cells and grid cells could integratively support our sense of direction to feasibly aid in making decisions or anticipating the future–for instance, when planning where to meet your friend for a lunch date. Although it’s of course not possible to directly assess an animal’s imagination, Bellmund notes that “future anticipation is key to survival for many species and evidence from a growing body of research suggests that spatial processing mechanisms might also support this ability in rodents. For example, hippocampal place cells do not only represent the animal’s current location, but can also carry information about previous and upcoming trajectories, referred to as replay and preplay, respectively. This might support the idea that spatially tuned cells provide the building blocks for the simulation of upcoming events in service of decision-making and planning. The majority of findings in line with this idea are based on recordings of place cells, but replay of grid cell activity suggests that the brain’s spatial navigation system as a whole might support future anticipation and planning.”

Multi-talented spatially-tuned neurons

Given the growing number of cognitive functions supported by spatially tuned cells, including navigation, sensing direction, and now imagining or planning, it’s possible the roles of these neurons are even more diverse than we’re currently aware. Bellmund explains,

“Recently, a grid-like code was observed while participants navigated a conceptual space, hinting at a role for the brain’s spatial navigation system in representing knowledge about our world. Understanding how exactly the brain’s spatial navigation system supports such different functions is a great challenge for future research.”

References

Bellmund JLS, Deuker L, Schröder TN, Doeller CF. (2016). Grid-cell representations in mental simulation. eLife. 5:e17089. doi: 10.7554/eLife.17089

Byrne P, Becker S, Burgess N. (2007). Remembering the past and imagining the future: a neural model of spatial memory and imagery. Psychol Rev. 114(2):340–75. doi: 10.1037/0033-295X.114.2.340

Constantinescu AO, O’Reilly JX, Behrens TEJ. (2016). Organizing conceptual knowledge in humans with a gridlike code. Science. 352(6292):1464-8. doi: 10.1126/science.aaf0941

Doeller CF, Barry C, Burgess N. (2010). Evidence for grid cells in a human memory network. Nature. 463(7281):657-61. doi: 10.1038/nature08704

Dragoi G, Tonegawa S. (2011). Preplay of future place cell sequences by hippocampal cellular assemblies. Nature. 469(7330):397-401. doi: 10.1038/nature09633

Ekstrom AD et al. (2003). Cellular networks underlying human spatial navigation. Nature. 425(6954):184-8. doi: 10.1038/nature01964

Hafting T et al. (2005). Microstructure of a spatial map in the entorhinal cortex. Nature. 436(7052):801-6. doi: 10.1038/nature03721

Hormer AJ et al. (2016). Grid-like Processing of Imagined Navigation. Curr Biol. 26(6):842-7. doi: 10.1016/j.cub.2016.01.042

Jacobs J et al. (2013). Direct recordings of grid-like neuronal activity in human spatial navigation. Nat Neurosci. 16(9):1188-90. doi: 10.1038/nn.3466

O’Keefe J, Dostrovsky J. (1971). The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Research. 34(1):171-5. doi: 10.1016/0006-8993(71)90358-1

Ólafsdóttir HF, Carpenter F, Barry C. (2016). T Coordinated grid and place cell replay during rest.. Nat Neurosci. 19(6):792-4. doi: 10.1038/nn.4291

Any views expressed are those of the author, and do not necessarily reflect those of PLOS.

Emilie Reas received her PhD in Neuroscience from UC San Diego, where she used fMRI to study memory. As a postdoc at UCSD, she currently studies how the brain changes with aging and disease. In addition to her tweets for @PLOSNeuro she is @etreas.