The navigation system is usually believed to stand on three pillars, represented by three types of neuronal correlates of space: head-direction, place and grid cells (1). However, in the last ten years, other cell classes have been described, for example boundary vector cells in the subiculum (2) (note that a slight variant of this cell class, the border cells, was also described in the entorhinal cortex (3)). The report of Olson and colleagues published in Nature Neuroscience (4) as its companion paper from Jacob (5), focus on a new type of cell that robustly encode for the axis of travel. Those cells are equivalent to HD cells with two different preferred firing directions separated by 180°. As is the case for HD cells, their firing should be independent of the location.
In the report by Jacob M Olson et al, they performed single neuron recordings during a navigational task in a triple T-maze as well as in a wall-less arena. The two fields are surrounded by the same distal cues. The rats were trained to explore every branch of the maze always with the same direction of travel. Then, the authors recorded ensembles of neurons in the subiculum and ran the following analysis: first they selected the neurons characterized by a bimodal head-direction tuning curve (they fitted the curves with a mixture of von Mises distribution, that is the equivalent of a mixture of Gaussian distributions on the angular axis). The authors also made sure that the firings of these neurons were independent of location. The angular difference between the two peaks of the head-direction tuning curves was on average 180° and the authors thus concluded that these were ‘axis-tuned neurons’.
Neurons that responded to those criteria where only found in track-running task on the T-maze but not during the free foraging task in the open arena. The authors suggest that this kind of activity would be generated by intense training on the track and/or restricted available trajectories.
They then asked the question of whether the neurons were anchored to the track or to the distal cues. To answer this question they rotated the maze by 90° relative to the cues and correlated the activity of axis-tuned neurons in this experiment with the activity in the non-rotated maze. The directionality of the neurons on the maze shifted by 90°, thus indicating that neurons are more likely to be anchored to distal cues.
It is noteworthy that the firing of these axis-tuned neurons share two key properties with the head-direction neurons: as we just said, they were aligned to distal cues and their firing was preserved in darkness, suggesting that they could be updated by idiothetic inputs (e.g. vestibular). These observations suggest that axis-tuned neurons and head-directions neurons may be tightly related to each other, an interesting starting point for future studies. Recording those neurons in other types of arenas could also address how their firings depend on the geometry of the environment, for example in square or rectangle fields or on a circular track. It would be interesting to control also for the modulation of those neurons by other behavioural parameters such as velocity or egocentric (i.e. body-aligned) coordinates by fitting Generalized Linear Models on the firing (6) and the authors recently did (7). These parameters can be further controlled by recoding in other types of mazes : indeed, in the T-maze, the animal’s speed differs from one axis to the other. Finally, training animals to run along a given trajectory in an open area would allow to study how a limited number of available trajectories affect the firing of axis-tuned neurons.
Overall, this article by Jacob M Olson provides a foundation for the study the neuronal correlates of navigation relative to environmental boundaries. Further investigations are needed to understand their relationship with other cell classes and their potential role in route planning.
1. R. M. Grieves, K. J.Jeffrey. The representation of space in the brain. Behavioural Processes, 135:113-131, 2017
2. J. O’Keefe & N. Burgness, Geometric determinants of the place fields of hippocampal neurons. Nature 381: 425 – 428, 1996
3. T. Solstad, et al. Representation of Geometric Borders in the Entorhinal Cortex, Science 322(5909): 1865-1868, 2009
4. J. M Olson, K. Tongprasearth & D. A. Nitz. Subiculum map the current axis of travel, Nature Neuroscience 20: 170–172, 2017
5. P. O. Jacob et al. An independent, landmark-dominated head-direction signal in dysgranular retrosplenial cortex. Nature Neuroscience 20: 173-175, 2017
6. A. Peyrache, L. Roux, N. Schieferstein, G. Buzsaki. Transformation of the head-direction signal into a spatial code, bioRxiv: 075986
7. A. S. Alexander, D. A Nitz. Spatially periodic activation patterns of retrosplenial cortex encode route sub-spaces and distance travelled, bioRxiv: 100537