(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Boundary vector cells in the goldfish central telencephalon encode spatial information [1]

['Lear Cohen', 'Department Of Biomedical Engineering', 'Ben-Gurion University Of The Negev', 'Beer-Sheva', 'Zlotowski Center For Neuroscience', 'Ehud Vinepinsky', 'Institut De Biologie De L École Normale Supérieure', 'Paris', 'Opher Donchin', 'Ronen Segev']

Date: 2023-05 

Navigation is one of the most fundamental cognitive skills for the survival of fish, the largest vertebrate class, and almost all other animal classes. Space encoding in single neurons is a critical component of the neural basis of navigation. To study this fundamental cognitive component in fish, we recorded the activity of neurons in the central area of the goldfish telencephalon while the fish were freely navigating in a quasi-2D water tank embedded in a 3D environment. We found spatially modulated neurons with firing patterns that gradually decreased with the distance of the fish from a boundary in each cell’s preferred direction, resembling the boundary vector cells found in the mammalian subiculum. Many of these cells exhibited beta rhythm oscillations. This type of spatial representation in fish brains is unique among space-encoding cells in vertebrates and provides insights into spatial cognition in this lineage.

Abbreviations: Dc, large-celled subdivision of Dm; Dd, dorsal division of area dorsalis; Dld, dorsal subdivision of lateral division of area dorsalis; Dlv, ventral subdivision of lateral division of area dorsalis; Dlv-d, dorsal part of Dlv; Dlv-v, ventral part of Dlv; Dm, medial subdivision of area dorsalis; Dmc, caudal part of medial subdivision of area dorsalis; Dmr, rostral part of medial subdivision of area dorsalis; Vd, dorsal nucleus of area ventralis; Vdi, intermediate subnucleus of Vd; Vs, supracommissural nucleus of area ventralis; Vv, ventral nucleus of area ventralis

Funding: We gratefully acknowledge financial support from THE ISRAEL SCIENCE FOUNDATION—FIRST Program (Grant no. 281/15 to RS and OD), THE ISRAEL SCIENCE FOUNDATION—FIRST Program (Grant no. 555/19 to RS), THE ISRAEL SCIENCE FOUNDATION (Grant no. 211/15 to RS), The Human Frontiers Science Foundation Grant (RGP0016/2019 to RS), and the Helmsley Charitable Trust through the Agricultural, Biological and Cognitive Robotics Initiative of Ben-Gurion University of the Negev (OD and RS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2023 Cohen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

(A) Schematic overview of the experimental setup: A fish swims freely in a water tank with a wireless recording system (see Materials and methods ) mounted on its head. The fish’s movements are recorded by a Raspberry Pi camera positioned in front of the tank. Dashed circle presents the directions used in all data analyses in this paper. We used camera-east as zero with an anti-clockwise progression. ( B) Example of the recording site in the goldfish central telencephalon and the corresponding brain region (right panel, anatomical diagram based on [ 40 ]). Black x’s show a location in which boundary vector cells were recorded. ( C) Example of a raw recording from a tetrode (black traces) and a reference electrode (gray) in the fish’s central telencephalon. Neural activity can be seen in the tetrode alone. Blue and red asterisks correspond to the blue and red clusters in panels D and E. The blue cluster corresponds to the cell presented in S2B Fig . The red cluster corresponds to the cell in Fig 2A–2D . (D) Waveforms of the two neurons after spike sorting. ( E ) Projections on the main principal components of the data from the tetrode of all spike candidates that crossed the threshold. Other clusters were not distinguishable from other multiunit activity and neural noise. The underlying data supporting panels D and E can be found in a file named Fig 1_data.mat (see Data Availability). Dld, dorsal subdivision of lateral division of area dorsalis; Dlv, ventral subdivision of lateral division of area dorsalis; Dlv-d, dorsal part of Dlv; Dlv-v, ventral part of Dlv; Dmr, rostral part of medial subdivision of area dorsalis; Vd, dorsal nucleus of area ventralis; Vdi, intermediate subnucleus of Vd; Vv, ventral nucleus of area ventralis.

To probe navigation in fish, we recorded multiple single-cell activities from the central area of the goldfish telencephalon. During the recordings, the goldfish could freely explore a water tank along its vertical and horizontal axes (see Fig 1A ). The recorded site in the goldish brain is located in the central area of their telencephalon, although its boundaries and function remain somewhat undetermined [ 22 – 24 ]. Studies have, however, suggested that this area integrates information from multiple neighboring pallial regions [ 23 ]. We hypothesized that it would integrate representations of space from the entire telencephalon.

One fundamental difference between fish and other vertebrates is the environment in which they navigate in. Navigation in aquatic environments is unique for several reasons. Fish navigate in full 3D, unlike terrestrial animals that are limited by the vertical dimension of their world. Furthermore, unlike flying animals, fish are subjected to a steep pressure gradient while navigating in the vertical dimension of their environment. This might imply that information about position in space is encoded in the fish brain differently than the locally activated cells shown to exist in 3D environments in bats [ 20 ] and rats [ 21 ].

A recent study described the existence of single cells that encode spatial and kinematic features of the environment in the goldfish lateral pallium [ 14 ]. The homology of this brain region between fish and the mammalian hippocampal formation is supported by anatomical and lesion studies [ 15 ]. The findings in goldfish include cells that encode the environment’s edges, the fish’s head direction, speed, and velocity [ 14 ]. However, no evidence of neuronal activity in a specific place field was found, although place cells have been identified in many studies in mammals [ 16 ] and recently were also reported to exist in birds [ 17 , 18 ]. In addition, no evidence of rhythmic neural oscillations associated with navigation has been found in the goldfish comparable to those in the theta frequency range of the mammalian hippocampal formation [ 19 ]. Therefore, a comparative approach is needed to better understand the fundamental mechanisms of spatial cognition across vertebrates. This can shed light on whether the mammalian model is valid for all vertebrates or whether different classes evolved different computational schemes. For this purpose, it is crucial to determine how position is represented in the brain of fish, the largest vertebrate class.

The place cell found in the hippocampus of rats [ 10 ] was the first spatial cell type to be identified. These cells are activated once the animal is in the cell’s preferred location (place field) in the environment. After the discovery of place cells, the existence of boundary vector cells was hypothesized to explain the dependency of place cells on environment borders [ 11 ]. Later, boundary-vector cells that encode proximity to allocentric boundaries [ 12 ] were found, together with other spatial cell types in the brain areas connected to the hippocampus (e.g., the entorhinal cortex and the subiculum). These spatial cells include the grid cells [ 13 ], and the head direction cells [ 13 ], which encode the animal’s allocentric head direction.

For most animals, the ability to locate themselves in the environment is crucial for survival [ 1 – 4 ]. This ability requires encoding information about self-position and locomotion [ 5 , 6 ]. Studies in navigating mammals have identified several types of neurons that encode self-position and locomotion in the hippocampal formation [ 5 – 9 ].

Results

To better understand how space is represented in the teleost brain, we recorded the activity of multiple single cells in the goldfish telencephalon. Before the recordings, we trained the fish to swim continuously in a rectangular water tank (Fig 1A) measuring 0.7 m on the horizontal axis and 0.7 m on the vertical axis (0.2 m in the third, foreshortened axis) by feeding the fish in various places in the tank. In most cases after 1 to 2 weeks of training sessions (every other day), the fish became familiar with the water tank and adapted to exploring its entire environment freely. At the end of the last training session, we implanted an extracellular tetrodes and wireless recording system (see Materials and methods) in its central telencephalon (Fig 1B; additional histology examples are presented in S1 Fig). After a day of recovery from the implant surgery, we let the goldfish explore the quasi-2D water tank embedded in a 3D environment while recording neuronal activity and tracking its position. Neuronal activity was recorded using 3 to 4 tetrodes, and single cells were identified using spike sorting (Fig 1C–1E), ranging from 1 to 12 single cells per recording session and from 1 to 8 single cells per tetrode. Subsequently, we analyzed the associations between cell activity and the fish’s trajectory (see Materials and methods).

The results showed that a considerable portion of the recorded cells had neural activity that was modulated by the fish’s position in the environment (39 out of 123; i.e., about 32% of all the well-isolated units recorded in the 15 fish used in this study. See examples in Fig 2). Of the 39 spatially modulated cells, 4 cells had a diffuse firing pattern and were removed from further analyses. The remaining 35 spatially modulated cells were recorded from the brains of 7 of the 15 fish used in this study, ranging from 1 to 11 spatially modulated cells per fish during up to 5 recording sessions, with a maximum of 5 spatially modulated cells in a single experiment.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Boundary vector cells in the goldfish brain. (A-D) An example of a boundary vector cell tuned to distance from the left wall of the water tank (experimental setup is shown in Fig 1A). (A) Left: Fish trajectory (black curve) is presented together with the location of the fish when each spike of a single cell occurred (red dots). The neuron was mainly active when the fish was near the left wall of the tank. Right: trajectory (top panel) and spike locations (bottom panel). (B) Firing rate heatmap of the cell in A, color coded from dark blue (zero firing rate) to dark red (maximal firing rate, indicated on the upper right side of the panel). The preferred boundary direction of this cell is indicated (Θ max ; see Materials and methods). The heatmap is occupancy corrected (see Materials and methods) to obtain a reliable estimate of the firing rate as a function of position. (C) In-session stability of the cell in A. The similar rate of the cell in the first (top panel) and second (bottom panel) halves of the experiment suggest stable activity within the recording session (correlation coefficient and p-value are indicated; see Materials and methods). (D) Spatial coherence (red arrow, top panel) and spatial information (red arrow, bottom panel) values of the cell in A are higher than those of 5,000 shuffled spike trains obtained from the same dataset (blue histograms; see Materials and methods). (E-H) Example of a boundary vector cell tuned to the bottom of the water tank. (I-P) Two other examples of boundary vector cells with a preferred boundary direction, which is neither horizontal nor vertical. The underlying data supporting all panels in this figure can be found in a file named Fig2_data.mat (see Data Availability). https://doi.org/10.1371/journal.pbio.3001747.g002

This population of space encoding cells reported here can best be described as boundary vector cells. This encoding scheme is characterized by a gradually decreasing firing rate with the animal’s distance from a boundary located at a specified direction. The presented cells, recorded from the central telencephalon of goldfish, had a gradual activity in a specified direction (Θ max ; see Materials and methods) rather than activity that is localized in space (as in place cells) or tightly tuned to physical barriers (as in border cells). This makes them appear to resemble the mammalian boundary vector cells [12].

Fig 2A–2D and S1 Video show an example of this type of cell whose firing rate gradually decreased with the distance of the fish from the left wall of the tank (Θ max = 179o). This was manifested in its spiking activity (red dots, Fig 2A) throughout the fish’s trajectory (black curve) as well as in the occupancy-corrected rate map in the water tank (Fig 2B).

To test statistically whether this neuron encoded components of position, we tested for in-session stability, spatial coherence, and for spatial information, as well as for crossing a spatial information threshold of 0.1 bits/spike (see Materials and methods). The similarity between the firing rate maps for the first (0 to 37 min) and the second (38 to 75 min) halves of the session indicated in-session stability (Fig 2C, correlation coefficient = 0.93, p < 1 × 10−3; see Materials and methods). The cell’s spatial coherence (Fig 2D, top panel, red arrow) and spatial information (Fig 2D, bottom panel, red arrow) were higher than the corresponding values for 5,000 shuffled spike trains generated using an interspike interval (ISI) shuffling procedure (blue histograms; see Materials and methods).

Another example of a boundary vector cell tuned to distance from the bottom of the water tank (Θ max = 285o) is presented in Fig 2E–2H. Additional examples of boundary vector cells with a horizontal (i.e., Θ max = 15o or 180o±15o) or a vertical (i.e., Θ max = 90o±15o or 270o±15o) preferred boundary direction are presented in S2B–S2N Fig, as well as in S2 and S3 Videos.

Not all boundary vector cells were tuned to boundaries in the vertical or horizontal directions. Rather, some cells had a firing pattern that gradually decreased with the distance of the fish from the corners of the water tank. An example of this type of cell is presented in Fig 2I–2L. For this cell, firing gradually decreased with distance from the top right corner (Θ max = 25o) of the water tank. Other examples of boundary vector cells with a preferred boundary in the direction of the corners of the water tank are presented in Figs 2M–2P and S2O–S2W.

To assess the preferred boundary direction of the boundary vector cells, we calculated the 2D correlation coefficient of the cells’ activity and the fish’s position (Fig 3A, red dot, which corresponds to the cell presented in Fig 2A–2D). The correlation coefficient was then compared to the 2D correlation coefficients of 5,000 shuffled spike trains (grey dots, the 97.5 percentile is depicted; see Materials and methods). This analysis showed that the example cell (Fig 2A) had a left boundary tuning (Θ max = 179o, the preferred and null directions and corresponding correlation coefficients are indicated; see Materials and methods). This was further confirmed by the spiking activity of the cell (red dots, Fig 3B) over the distance of the fish from the preferred boundary. This cell’s tuning curve of firing rate versus distance from the preferred boundary showed a clear gradually decreasing pattern (Fig 3C, blue curve), whereas no clear tuning pattern was visible for distance from a boundary in the orthogonal direction (Θ max +90o, yellow curve). Similar results are presented in panels D-F, G-I, and J-L and correspond to the cells presented in Fig 2E–2H, 2I–2L and 2M–2P, respectively.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Spatial characteristics of boundary vector cells. (A-C) Spatial tuning properties of the cell presented in Fig 2A-2D. (A) 2D correlation coefficient of firing rate and the fish’s position of the cell (red dot) and 5,000 shuffled spike trains (gray dots, the 97.5 percentile is depicted), suggesting this cell had a left boundary tuning. Preferred and null tuning directions (Θ max and Θ min , respectively; see Materials and methods) and correlation coefficients (cc) are indicated. (B) Spiking activity (red dots) superimposed on the distance of the fish (black curve) from the preferred boundary. (C) Tuning curves (mean ± standard deviation) of the cell’s firing rate to the distance of the fish from the boundary in the preferred direction (Θ max , blue curve) and its orthogonal direction (Θ max +90o, orange curve). A gradually decreasing firing pattern is shown for the Θ max direction. (D-L) The spatial tuning properties of the boundary vector cells presented in Fig 2E–2H, 2I–2L and 2M–2P, respectively. The underlying data supporting all panels in this figure can be found in a file named Fig 3_data.mat (see Data Availability). https://doi.org/10.1371/journal.pbio.3001747.g003

Furthermore, we tested whether the activity of the boundary vector cells is modulated by the allocentric swimming direction of the fish. Fig 4 presents the result of this analysis for the cells in Fig 2, respectively. We show each cell’s heatmap of firing rate to position, bisected into the periods during which the fish swam towards its preferred boundary direction (Θ max ±90o; Fig 4A, 4D, 4G and 4J, left panels, see Materials and methods) or away from its preferred boundary direction (Θ max ±180o±90o; Fig 4A, 4D, 4G and 4J, right panels). In addition, we show the tuning curves of firing rate to the distance from the preferred boundary of each cell while swimming towards it (Fig 4B, 4E, 4H and 4K, blue curve; see Materials and methods) and away from it (yellow curve). To assess whether this analysis of allocentric swimming direction is biased by the fish’s swimming speed, we also present the speed distribution along distance from the preferred boundary in each of the allocentric swimming directions (Fig 4C, 4F, 4I and 4L).

For three of the examples (Fig 4D–4L), the bisected heatmaps were similar (Fig 4D, 4G, and 4J), but the firing rate was attenuated while swimming away from the preferred boundary. This is shown in the bisected tuning curves (Fig 4E, 4H and 4K, modulation ratio indices are indicated; see Materials and methods). The additional cell shows no tuning to allocentric swimming direction (Fig 4A–4C).

To estimate the strength of the rate modulation, we calculated a modulation ratio index for all 35 boundary vector cells (see Materials and methods). This value indicates the unsigned ratio (i.e., stronger divided by weaker) between the tuning curves of firing rate to distance from a boundary in the preferred direction while swimming towards it versus away from it (e.g., the blue and yellow tuning curves in Fig 4B, respectively). Thus, a modulation ratio index around 1 indicates no clear difference between the tuning curves. The modulation ratio indices of all boundary vector cells are shown in Fig 4M, distributed in the range of [1, 3.6] with a mean ± standard deviation of 1.67 ± 0.67.

To address the possibility that our results might be mediated by a difference in swimming speed towards the boundary versus away from it, we used a 2 × 2 ANOVA (see Materials and methods). Under this approach, in only 2 of the 35 cells, we found a clear effect of direction on speed (ANOVA F-value in the range 11.8 to 20.89). For 23 of the cells, we found similar swimming speeds in the two directions (ANOVA F-value in the range 0 to 0.91), and the remaining 10 cells had a speed–direction relationship that was neither clearly different nor very similar (ANOVA F-value in the range 1.04 to 11.6).

Although this analysis supports the suggestion that there is a directionality tuning in the activity of the boundary vector cells in fish, it is still hard to confidently differentiate the effect of allocentric swimming direction from other behavioral aspects in space. Therefore, further investigations are needed to establish this point.

Last, in both fish and rodents, position was more important than direction: As was shown in rats [12], we calculated the spatial information carried by the goldfish boundary vector cells to be 5 times greater on average than the corresponding directional information (0.5 ± 0.37 bits/spike versus 0.1 ± 0.07 bits/spike, respectively; mean ± standard deviation; see Materials and methods).

Boundary tuning while changing environmental geometry To test the hypothesis that the recorded cells were tuned to a preferred boundary rather than to a specific place field in the environment, we conducted an additional experiment. In this control experiment, we measured the neural activity of the same cells before and after we changed the tank geometry. This was done by adding an additional horizontal half-wall (control in 8 of the 35 boundary vector cells). Examples of these cells are presented in Fig 5. Each cell was recorded for a full recording session (60 to 75 min; Fig 5A), after which the fish was blindfolded, and a Perspex shelf was inserted into the water tank (Fig 5B). Another full recording session was then initiated. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Changing environmental geometry. (A) The first session was recorded in the main experimental water tank. (B) Before the second recording session, a shelf was inserted into the water tank to modulate the geometry of the environment. (C) Example rate map of a boundary vector cell tuned to distance from the bottom of the water tank, color coded from dark blue (zero firing rate) to dark red (maximal firing rate, indicated on the top right side of the panel). (D) Rate map over the water tank of the same cell after adding the shelf. Firing was modulated by both the bottom of the water tank and above the shelf. (E-H) Additional examples of boundary vector cells before and after the geometric change in the environment. The underlying data supporting all panels in this figure can be found in a file named Fig 5_data.mat (see Data Availability). https://doi.org/10.1371/journal.pbio.3001747.g005 In one example, in the first recording session, the cell’s neural activity gradually decreased with distance from the bottom of the water tank (Fig 5C). After adding the shelf (gray mark in Fig 5D), the cell’s rate map (Fig 5D) shows it responded to both the bottom of the water tank and the Perspex shelf, as expected from a boundary vector cell. Additional examples showing firing patterns of boundary vector cells before and after the geometric change in the environment are presented in Figs 5E–5H and S3A–S3G.

Beta oscillations in boundary vector cells About half of the boundary vector cells exhibited rhythmic neural activity. An example of one such cell is presented in Fig 7A–7C. This cell is a boundary vector cell with a gradually decreasing firing pattern with distance from the bottom of the water tank (Fig 7A). Examining the histogram of interspike interval of this cell (Fig 7B) revealed a periodic spacing pattern between the peaks of the histogram. This was further manifested in the frequency domain: After the histogram was normalized by the total number of spikes, we calculated the power spectral density function of the normalized histogram (Fig 7C), where a local maximum appeared at around 16Hz; in other words, this neuron oscillated rhythmically in the low-beta frequency range. A counter-example of a cell that was not tuned to space is presented in Fig 7D–7F. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. Beta oscillations in boundary vector cells. (A-C) Example of a boundary vector cell with a periodic interspike interval (ISI) pattern. (A) Firing rate heatmap of a boundary vector cell tuned to the bottom of the water tank, color coded from dark blue (zero firing rate) to dark red (maximal firing rate, indicated on the upper right side of the panel). (B) ISI histogram of the cell in A. Even spacing between the peaks suggests periodic oscillations of the neural activity. (C) Power spectral density of the histogram in B (normalized). Colored background marks the typical frequency range of beta waves (12.5–30 Hz). A local peak is shown at approximately 16 Hz, suggesting that this cell exhibited beta-rhythm oscillations. (D-F) A counter-example of a cell with no clear spatial tuning and no specific pattern in the ISI histogram. (G) Maximal power spectral density in the beta waves range for the entire population. The color bar spans the population’s max-p values (see Materials and methods) on a logarithmic scale. As shown, roughly half of the boundary vector cells (blue dots) exhibited beta rhythm oscillations at 15.25 ± 1.62 Hz (mean ± standard deviation). (H) Different thresholds were tested to estimate the prevalence of beta oscillations in the recorded population (see Materials and methods). The median value of the population’s max-p was used to divide the data into groups similar in size. Regardless of the threshold tested, the cells that were more spatially tuned (i.e., below median-max-p = 0.11, blue curve) were more abundant above the threshold than others (i.e., above median-max-p value, orange curve). The underlying data supporting all panels in this figure can be found in a file named Fig 7_data.mat (see Data Availability). https://doi.org/10.1371/journal.pbio.3001747.g007 Examining the position and magnitude of the peak power spectral density in the beta range of the entire population (Fig 7G, color-coded by the max-p values; see Materials and methods) revealed a clear connection between the beta oscillations and the boundary vector cells in the central telencephalon of the goldfish. To estimate the properties and prevalence of the beta oscillations in the population, we used a threshold of 0.0015 dB/Hz for the peak power of the spectral density. The findings showed that 16 of the 35 boundary vector cells (approximately 46%) and 1 out of the other 84 cells crossed this threshold with a peak spectral density in 15.25 ± 1.62 Hz (mean ± standard deviation). Different thresholds were tested to verify that these results were independent of the chosen threshold (Fig 7H; see Materials and methods).

Sigmoidal firing patterns in space To further characterize the spatial firing patterns observed in the boundary vector cells, we tested two tuning models: (1) Sigmoid tuning along the cell’s preferred direction, resembling distance encoding; and (2) Gaussian tuning in the form of a 2D Gaussian centered in proximity to the environmental boundaries, resembling local encoding. We fitted the two models for the firing rate map of each cell and tested the goodness of fit for the two models. This process was first validated using a simulated dataset (examples are presented in S4 Fig; see Materials and methods). Using the simulated data, we set the thresholds (S4J and S4K Fig; see Materials and methods) for classifying a cell as distance encoding or Gaussian encoding (as in a place cell). As expected, the results showed that most boundary vector cells fit better into the sigmoid model than the Gaussian model (see S1 Table). This emerged in the correlations between the data and the models (S4K Fig). Since a 2D Gaussian with a very broad distribution on only one axis is similar to a sigmoid distribution, these two models are not mutually exclusive. Nevertheless, this analysis helped us to rule out the hypothesis that the cells we recorded were a subset of other typical locally activated space-encoding cells found in vertebrates, characterized by a narrow Gaussian distribution that happen to be centered near boundaries.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001747

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/