A local anchor for the brain's compass

NEWS AND VIEWS npg © 2014 Nature America, Inc. All rights reserved. recorded S1 responses to the whisker stimuli and found that the effect of adaptation on the oddball stimulus response was much smaller than the effect on the ongoing pulses. Thus, adaptation enhanced the difference between responses to the oddball target and the ongoing pulses. This suggests that adaptation might increase the salience of rapid changes in input against a background of ongoing stimulation: for instance, those associated with stick-slip events during whisking against textured objects12–15. Overall, Musall et al.3 offer powerful causal evidence that adaptation profoundly affects tactile perception. The strength of their work lies largely in their careful, laborious matching of tactile and optogenetic stimuli in terms of psychometric curves and neural responses. In this respect, Musall et al.3 set a rigorous benchmark for future work and point the way for the next generation of studies that aim to causally link sensory cortex activity and perception. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Ollerenshaw, D.R., Zheng, H.J., Millard, D.C., Wang, Q. & Stanley, G.B. Neuron 81, 1152–1164 (2014). 2. Wang, Q., Webber, R.M. & Stanley, G.B. Nat. Neurosci. 13, 1534–1541 (2010). 3. Musall, S. et al. Nat. Neurosci. 17, 1567–1573 (2014). 4. Khatri, V., Hartings, J.A. & Simons, D.J. J. Neurophysiol. 92, 3244–3254 (2004). 5. Romo, R., Hernandez, A., Zainos, A. & Salinas, E. Nature 392, 387–390 (1998). 6. Sachidhanandam, S., Sreenivasan, V., Kyriakatos, A., Kremer, Y. & Petersen, C.C. Nat. Neurosci. 16, 1671–1677 (2013). 7. O’Connor, D.H. et al. Nat. Neurosci. 16, 958–965 (2013). 8. Luna, R., Hernandez, A., Brody, C.D. & Romo, R. Nat. Neurosci. 8, 1210–1219 (2005). 9. Gerdjikov, T.V., Bergner, C.G., Stuttgen, M.C., Waiblinger, C. & Schwarz, C. Neuron 65, 530–540 (2010). 10. von Heimendahl, M., Itskov, P.M., Arabzadeh, E. & Diamond, M.E. PLoS Biol. 5, e305 (2007). 11. Huber, D. et al. Nature 451, 61–64 (2008). 12. Jadhav, S.P., Wolfe, J. & Feldman, D.E. Nat. Neurosci. 12, 792–800 (2009). 13. Wolfe, J. et al. PLoS Biol. 6, e215 (2008). 14. Ritt, J.T., Andermann, M.L. & Moore, C.I. Neuron 57, 599–613 (2008). 15. Arabzadeh, E., Zorzin, E. & Diamond, M.E. PLoS Biol. 3, e17 (2005). A local anchor for the brain’s compass Martin J Chadwick & Hugo J Spiers Retrosplenial cortex neurons provide a signal akin to a compass readout. Evidence in humans now demonstrates that these neurons anchor their representations locally, locking to the geometry of a room rather than to the city beyond. As anyone with a poor sense of direction can attest, the ability to orient yourself in an environment is fundamental for finding your way in the world. Although the modern age has provided us with GPS built into our mobile devices, rendering our own sense of direction less important, it is worth considering the historical importance that this ability has had for humanity. Through much of history, maintaining an accurate sense of direction over a long journey could be the difference between life and death. Indeed, it is no coincidence that the fear induced by the feeling of being completely lost has been the central theme of countless myths and fairy tales over the ages, from the Labyrinth of Knossos to Hansel and Gretel. Now Marchette et al.1 add to our understanding of this sense of direction by showing that neural populations encoding direction can become anchored to the immediate environment. We know from studies of rodents that neurons in limbic structures known as head direction cells act analogously to an internal compass. Each head direction cell becomes active whenever the rat faces in a particular Martin J. Chadwick and Hugo J. Spiers are in the Institute of Behavioural Neuroscience, Department of Experimental Psychology, University College London, London, UK. e-mail: [email protected] 1436 direction relative to the environment2, with each cell preferring a slightly different orientation. This means that, for every facing direction in a given environment, a unique population of head direction cells is active, providing a compass-like signal that can be read by connected brain structures. To provide a meaningful sense of direction within any specific environment, this compass needs to be ‘anchored’ to that environment, just as a real compass needs to latch on to the Earth’s magnetic field. In a room that is shielded from the magnetic field, a real compass becomes useless, and the same principle applies to the brain’s own internal compass: without some way of fixing to the external world, it cannot provide a usable sense of direction. This anchoring could occur relative to either the global environment (for example, streets aligned north in a city) or relative to the local environment occupied (for example, the walls in a room). But which brain region, or regions, accomplishes this? How does the brain decide which direction will be ‘north’ in any new environment? These are the key questions posed by Marchette et al.1. The authors created a realistic virtual environment consisting of a set of small museums set in a large park (Fig. 1). The subjects learned the spatial relationships among eight unique objects located along the walls of each museum. After learning, they were tested on their ability to imagine facing each object and indicate whether another object would be on their left or right. Before investigating the neural basis of direction anchoring, the authors first tested for behavioral evidence of a global or local anchor in their sense of direction. They found that judging direction for a specific facing direction was faster when subjects had just been asked to imagine that same facing direction on the previous trial. Crucially, this was only the case when the direction matched relative to the museum’s room geometry (for example, facing away from the door, or ‘local north’), not relative to the park (for example, facing toward the mountains, or ‘global north’), and it generalized across the museums. Thus, the local geometry of the rooms dominated over the global geometry of the park in the subjects’ representations of the object locations. Having demonstrated that the local anchoring of direction representations were not random in their task, but were instead seemingly determined by the shared geometries of the four museums, the authors next investigated the brain mechanisms supporting these representations. To this end, they scanned a second group of subjects with functional magnetic resonance imaging (fMRI) while the subjects performed the same task. The analysis focused on a particular region of brain in the medial parietal cortex, referred VOLUME 17 | NUMBER 11 | NOVEMBER 2014 NATURE NEUROSCIENCE NEWS AND VIEWS Imagining being at point A Facing north Facing away from door Imagining being at point B Facing north Facing left from door Imagining being at point C Facing east Facing away from door Marina Corral Spence/Nature Publishing Group Building 1 A Building 2 MRI voxel pattern B C npg © 2014 Nature America, Inc. All rights reserved. Figure 1 Local room-bound direction coding in the RSC. Left, a map of an environment containing two buildings. Inside the buildings three specific viewpoints (A–C) are shown. Right, MRI activity patterns evoked in medial parietal cortex (RSC) in response to imagining each of the viewpoints and judging the relative to direction to other locations in each room. MRI voxel patterns are correlated (A versus C) when the views match the same local directions (for example, facing away from the door), but not when they differ or when they match purely in global direction (A versus B). to by the authors as the retrosplenial complex (RSC). The authors applied pattern analysis to the fMRI activity in this region and found a pattern of results that showed a marked agreement with the behavioral data. There was a robust direction code within each museum, but, notably, this code generalized across the different museums such that the neural pattern for local north in museum A matched the pattern for local north in museum B. By contrast, the authors found no evidence for the presence of a global code for direction that stayed constant across the entire park environment (Fig. 1). These results provide clear evidence that the RSC is capable of providing a neural code for direction within the local environment that is anchored to the local geometry of that environment. So far, we have focused on the factors determining the anchoring of direction representations. However, the authors were also able to investigate location coding in both the behavioral and fMRI data. Unexpectedly, they found a pattern of results that precisely mirrored that for the direction coding. Not only did they find evidence of precise location coding within each museum, as was predicted, they also found that location generalized across the different museums on the basis of the shared geometry. Remarkably, the authors found it possible to reconstruct the layout of museum A on the basis of subject’s RSC activity for museum B. These results suggest that, at least for this set of environments, there is a spatial code for direction and location that is not absolute, but is defined in relation to the local environment. Furthermore, this spatial code appears to generalize across environments with the same geometric spatial arrangement, suggesting that this geometry is important in each case for anchoring both direction and location coding. The evidence clearly suggests that both direction and location codes are present in the human RSC and that they are anchored to the local environment through salient cues—in this case, the geometry of the room boundaries. But what neural mechanisms might underpin these results? Populations of head direction cells have been reported in rodent RSC3. The anchoring of such cells to local geometry could explain the sensitivity to direction, but not location. The authors propose that the presence of both direction and location codes could be explained if their data were the result of a theoretical neuronal ensemble that encodes the egocentric bearings to specific landmarks. These theoretical neurons they dub reference vector cells. If these cells respond to the walls of each environment, then this could explain the generalization across the different museums, given the shared geometric arrangement. Although this is an interesting proposal that is certainly worthy of consideration, it may be premature to posit the existence of an entirely new class of specialized spatial neuron. As the authors acknowledge, the same data could equally well be explained by a type of cell that has been found in the parahippocampal cortex of rodents: the boundary vector cell4. These cells respond whenever a rodent is at a specific distance and environment-centric direction from a boundary (for example, 20 cm from the east wall in a 1-m square enclosure), and can also be sensitive to heading direction. Regardless of what the precise neural mechanism turns out to be, this new study provides an important advance in our knowledge of how a sense of direction is generated in each new environment that we encounter. Clearly, knowing which direction you are facing and where you are located in a building is useful to avoid becoming lost. However, NATURE NEUROSCIENCE VOLUME 17 | NUMBER 11 | NOVEMBER 2014 it is of little use when navigating the streets of the city outside. A similar problem has been identified for the hippocampus, which contains neurons known as place cells that encode current location in the environment5. Place cells have recently been shown to be highly dependent on the local geometry of the environment, expressing similar spatial firing patterns in each of a set of geometrically identical compartments connected by a corridor6. A more global code for direction and location must exist to allow navigation across the different fragments of space, each with its own local code. But how it operates and whether it resides in the RSC remains unclear. Similarly, it is unknown which brain region signals the direction to future goal locations. Given recent observations that the human entorhinal cortex encodes the distance along a vector to the goal during navigation7, the entorhinal cortex, rather than the RSC, would seem a potential candidate for this function. Regardless, the results of Marchette et al.1 provide an important step forward in our understanding of how the human brain has allowed our species to navigate to safety through the ages, both in modern-day environments and in the dark and forbidding forests of antiquity. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Marchette et al. Nat. Neurosci. 17, 1598–1606 (2014). 2. Taube, J.S., Muller, R.U. & Ranck, J.B. J. Neurosci. 10, 420–435 (1990). 3. Lever, C., Burton, S., Jeewajee, A., O’Keefe, J. & Burgess, N. J. Neurosci. 29, 9771–9777 (2009). 4. Chen, L.L., Lin, L.H., Green, E.J., Barnes, C.A. & McNaughton, B.L. Exp. Brain Res. 101, 8–23 (1994). 5. O’Keefe, J. & Dostrovsky, J. Brain Res. 34, 171–175 (1971). 6. Spiers, H.J., Hayman, R.M., Jovalekic, A., Marozzi, E. & Jeffery, K.J. Cereb. Cortex published online, doi:10.1093/cercor/bht198 (13 August 2013). 7. Howard, L.R. et al. Curr. Biol. 24, 1331–1340 (2014). 1437