Periodic Reporting for period 4 - NATURAL_BAT_NAV (Neural basis of natural navigation: Representation of goals, 3-D spaces and 1-km distances in the bat hippocampal formation – the role of experience)
Période du rapport: 2021-05-01 au 2022-04-30
The overall goal of this project was to elucidate the neural bases of navigation under evermore naturalistic conditions. We had 5 specific aims (objectives):
Aim 1: Answering, for the first time, the question: Is there a vectorial representation of goals by single cells in the mammalian hippocampal formation? Namely, are there neurons representing the direction and distance to the goal, which could support goal-directed navigation?
Aim 2: Elucidating 3D spatial representations by 3D grid cells in the mammalian entorhinal cortex: Do grid fields have spherical 3-D shapes in 3-D space, and are they stacked in a 3D optimal-packing configuration – i.e. a lattice? Or are they cylindrical in shape, insensitive to the Z-dimension?
Aim 3: Elucidating 3D spatial representations by 3D border cells: Do 3D border cells represent all the walls, floor and ceiling? Can we find border/boundary-sensitive cells that fire at-a-distance from the wall, as predicted by boundary-vector cell (BVC) models?
Aim 4: Answering, for the first time, the question: Are there kilometer-sized place fields and grid-fields in the mammalian hippocampus and entorhinal cortex? Are the place-fields large in the middle of the flyway, and small around the reward zones and other important locations? Or perhaps the representation of very large spaces is fundamentally different as compared to results of small-scale laboratory experiments?
Aim 5: Examining the effect of natural experience on spatial representations: What is the difference between spatial representations (place cells and grid cells) in our 1-km test facility in wild-born bats that flew routinely such distances, versus laboratory-born bats that have never experienced a 1-km environment?
Answering these Aims is expected to revolutionize our understanding of the neural basis of natural navigation.
Aim 1: Answering, for the first time, the question: Is there a vectorial representation of goals by single cells in the mammalian hippocampal formation? Namely, are there neurons representing the direction and distance to the goal, which could support goal-directed navigation?
Aim 2: Elucidating 3D spatial representations by 3D grid cells in the mammalian entorhinal cortex: Do grid fields have spherical 3-D shapes in 3-D space, and are they stacked in a 3D optimal-packing configuration – i.e. a lattice? Or are they cylindrical in shape, insensitive to the Z-dimension?
Aim 3: Elucidating 3D spatial representations by 3D border cells: Do 3D border cells represent all the walls, floor and ceiling? Can we find border/boundary-sensitive cells that fire at-a-distance from the wall, as predicted by boundary-vector cell (BVC) models?
Aim 4: Answering, for the first time, the question: Are there kilometer-sized place fields and grid-fields in the mammalian hippocampus and entorhinal cortex? Are the place-fields large in the middle of the flyway, and small around the reward zones and other important locations? Or perhaps the representation of very large spaces is fundamentally different as compared to results of small-scale laboratory experiments?
Aim 5: Examining the effect of natural experience on spatial representations: What is the difference between spatial representations (place cells and grid cells) in our 1-km test facility in wild-born bats that flew routinely such distances, versus laboratory-born bats that have never experienced a 1-km environment?
Answering these Aims is expected to revolutionize our understanding of the neural basis of natural navigation.
We have answered in full all the 5 Aims of this ERC project.
On the technical side, we have developed new generations of 64-channel neural loggers that allow us to record spikes from dozens of neurons simultaneously in freely-flying bats (important for all Aims 1-5). We also set up a large 200-meter flight tunnel and a localization system that allows tracking the bat’s position with a 10-cm accuracy over large spatial scales (important for Aims 4-5).
On the scientific side, we achieved our goals on all the five Aims:
Aim 1: We finished this Aim, and published a paper in Science summarizing our exciting results (Sarel et al., Vectorial representation of spatial goals in the hippocampus of bats, Science 2017). In this paper, we found neurons that represent the direction and distance to navigational goals – i.e. a vectorial representation of spatial goals. Following this study, we asked whether not only navigational goals are represented in the hippocampus, but also other things “out there” are represented – and indeed we found that the location of other conspecifics (other bats) is represented by neurons in the bat hippocampus: A paper summarizing these exciting results was published in Science (Omer et al., Social place-cells in the bat hippocampus, Science 2018).
Aims 2-3: We finished these two Aims, and recently published a paper in Nature summarizing our exciting results on these two aims (Ginosar et al., Locally ordered representation of 3D space in the entorhinal cortex, Nature 2021). In this paper, we described our discovery of 3D grid cells (Aim 2) and 3D border cells (Aim 3) in the entorhinal cortex of flying bats. Grid fields were not cylindrical bur spherical, and exhibited fixed distances to nearest-fields – i.e. a fixed local distance scale – but did not exhibit a global lattice. These results revolutionize our understanding of grid cells and suggest that, at least in 3D, grid cells form not a global metric but a local metric for 3D space. In addition, we have used some of the previously acquired 3D flight- and neural-data (and also 2D data) to ask a question about the relevance for hippocampal and entorhinal processing of the theta oscillation. We found that there was no theta oscillation in bats, but we did find phase-locking (synchronization) and phase-precession (phase-coding) to the nonoscillatory field activity in the hippocampus – suggesting a novel nonoscillatory phase coding in the hippocampus. These exciting results were recently published in Cell, and they suggest that although oscillations do not generalize across mammalian species, coding principles do generalize (Eliav et al., Nonoscillatory phase coding and synchronization in the bat hippocampal formation, Cell 2018).
Aims 4-5: We finished these two Aims, and recently published a paper in Science summarizing our exciting results on these two aims (Eliav et al., Multiscale representation of very large environments in the hippocampus of flying bats, Science 2021). To this end, we first established all the technical foundations necessary for this ambitious project – including a new 64-channel neural logger (wireless electrophysiology system) and a positional-tracking system that is 100 times more accurate than GPS. We have built a long flight-tunnel at the Weizmann Institute (200 meters long) and recorded hundreds of hippocampal neurons in bats flying in this tunnel – both in wild-born bats (Aim 4) and laboratory-born bats (Aim 5). We found that the bat hippocampus exhibited a multifield multiscale spatial code in this large environment, which is very different from the single-field single-scale neural codes found in small environments, in both rodents and bats. Theoretical decoding analysis showed that this multiscale code is much more efficient for representing very large environments. Surprisingly, we found that in lab-born bats, the same multifield multiscale neural code exists – suggesting that the multifield multiscale code does not require previous experience with large environment, and indicating that the multifield multiscale code is a very robust characteristic of the hippocampus, regardless of experience.
On the technical side, we have developed new generations of 64-channel neural loggers that allow us to record spikes from dozens of neurons simultaneously in freely-flying bats (important for all Aims 1-5). We also set up a large 200-meter flight tunnel and a localization system that allows tracking the bat’s position with a 10-cm accuracy over large spatial scales (important for Aims 4-5).
On the scientific side, we achieved our goals on all the five Aims:
Aim 1: We finished this Aim, and published a paper in Science summarizing our exciting results (Sarel et al., Vectorial representation of spatial goals in the hippocampus of bats, Science 2017). In this paper, we found neurons that represent the direction and distance to navigational goals – i.e. a vectorial representation of spatial goals. Following this study, we asked whether not only navigational goals are represented in the hippocampus, but also other things “out there” are represented – and indeed we found that the location of other conspecifics (other bats) is represented by neurons in the bat hippocampus: A paper summarizing these exciting results was published in Science (Omer et al., Social place-cells in the bat hippocampus, Science 2018).
Aims 2-3: We finished these two Aims, and recently published a paper in Nature summarizing our exciting results on these two aims (Ginosar et al., Locally ordered representation of 3D space in the entorhinal cortex, Nature 2021). In this paper, we described our discovery of 3D grid cells (Aim 2) and 3D border cells (Aim 3) in the entorhinal cortex of flying bats. Grid fields were not cylindrical bur spherical, and exhibited fixed distances to nearest-fields – i.e. a fixed local distance scale – but did not exhibit a global lattice. These results revolutionize our understanding of grid cells and suggest that, at least in 3D, grid cells form not a global metric but a local metric for 3D space. In addition, we have used some of the previously acquired 3D flight- and neural-data (and also 2D data) to ask a question about the relevance for hippocampal and entorhinal processing of the theta oscillation. We found that there was no theta oscillation in bats, but we did find phase-locking (synchronization) and phase-precession (phase-coding) to the nonoscillatory field activity in the hippocampus – suggesting a novel nonoscillatory phase coding in the hippocampus. These exciting results were recently published in Cell, and they suggest that although oscillations do not generalize across mammalian species, coding principles do generalize (Eliav et al., Nonoscillatory phase coding and synchronization in the bat hippocampal formation, Cell 2018).
Aims 4-5: We finished these two Aims, and recently published a paper in Science summarizing our exciting results on these two aims (Eliav et al., Multiscale representation of very large environments in the hippocampus of flying bats, Science 2021). To this end, we first established all the technical foundations necessary for this ambitious project – including a new 64-channel neural logger (wireless electrophysiology system) and a positional-tracking system that is 100 times more accurate than GPS. We have built a long flight-tunnel at the Weizmann Institute (200 meters long) and recorded hundreds of hippocampal neurons in bats flying in this tunnel – both in wild-born bats (Aim 4) and laboratory-born bats (Aim 5). We found that the bat hippocampus exhibited a multifield multiscale spatial code in this large environment, which is very different from the single-field single-scale neural codes found in small environments, in both rodents and bats. Theoretical decoding analysis showed that this multiscale code is much more efficient for representing very large environments. Surprisingly, we found that in lab-born bats, the same multifield multiscale neural code exists – suggesting that the multifield multiscale code does not require previous experience with large environment, and indicating that the multifield multiscale code is a very robust characteristic of the hippocampus, regardless of experience.
We have progressed beyond expected on the technical side: We have developed the smallest neural logger (wireless electrophysiology) system in the world, a tiny 64-channel system that is now being used by a number of other labs in Europe and worldwide; and we have established a location-tracking methodology that is 100 times more accurate than GPS. All these technical developments push the envelope of behavioral neuroscience experiments worldwide.
We have also progressed beyond expected on the scientific side: As described above, Aim 1 has led us to ask questions about the representation of the positions of other individuals by hippocampal neurons, which indeed we discovered in the bat hippocampus – a result published in Science (Omer et al., Science 2018). We have several other projects in this social-neuroscience direction that have now started as a direct spin-off from this ERC-CoG project, and in particular from its Aim 1. In addition, Aims 4-5 led us to another spin-off project where 2 bats are flying simultaneously in the long flight tunnel (hundreds of meters), and we are finding very interesting neural responses during the interactions between the two bats.
So overall, the scientific output of this ERC-CoG project has achieved all its 5 scientific aims and technical aims, and moreover, went far beyond them.
We have also progressed beyond expected on the scientific side: As described above, Aim 1 has led us to ask questions about the representation of the positions of other individuals by hippocampal neurons, which indeed we discovered in the bat hippocampus – a result published in Science (Omer et al., Science 2018). We have several other projects in this social-neuroscience direction that have now started as a direct spin-off from this ERC-CoG project, and in particular from its Aim 1. In addition, Aims 4-5 led us to another spin-off project where 2 bats are flying simultaneously in the long flight tunnel (hundreds of meters), and we are finding very interesting neural responses during the interactions between the two bats.
So overall, the scientific output of this ERC-CoG project has achieved all its 5 scientific aims and technical aims, and moreover, went far beyond them.