Abstract
Gravity may provide a ubiquitous allocentric reference to the brain's spatial orientation circuits. Here we describe neurons in the macaque anterior thalamus tuned to pitch and roll orientation relative to gravity, independently of visual landmarks. We show that individual cells exhibit two-dimensional tuning curves, with peak firing rates at a preferred vertical orientation. These results identify a thalamic pathway for gravity cues to influence perception, action and spatial cognition.
This is a preview of subscription content, access via your institution
Access options
Subscribe to this journal
Receive 12 print issues and online access
$209.00 per year
only $17.42 per issue
Buy this article
- Purchase on Springer Link
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
References
Angelaki, D.E., Shaikh, A.G., Green, A.M. & Dickman, J.D. Nature 430, 560–564 (2004).
Laurens, J., Meng, H. & Angelaki, D.E. Nat. Neurosci. 16, 1701–1708 (2013a).
Laurens, J., Meng, H. & Angelaki, D.E. Neuron 80, 1508–1518 (2013b).
Zhou, W., Tang, B.F., Newlands, S.D. & King, W.M. J. Neurophysiol. 96, 2915–2930 (2006).
Merfeld, D.M., Zupan, L. & Peterka, R.J. Nature 398, 615–618 (1999).
Laurens, J. & Angelaki, D.E. Exp. Brain Res. 210, 407–422 (2011).
Yakusheva, T.A. et al. Neuron 54, 973–985 (2007).
Taube, J.S. Annu. Rev. Neurosci. 30, 181–207 (2007).
Yoder, R.M., Clark, B.J. & Taube, J.S. Trends Neurosci. 34, 561–571 (2011).
Shinder, M.E. & Taube, J.S. J. Neurophysiol. 111, 2479–2492 (2014).
Finkelstein, A. et al. Nature 517, 159–164 (2015).
Wilson, J.J., Page, H. & Jeffery, K. Preprint at bioRxiv http://dx.doi.org/10.1101/043711 (2016).
Meng, H., May, P.J., Dickman, J.D. & Angelaki, D.E. J. Neurosci. 27, 13590–13602 (2007).
Meng, H., Green, A.M., Dickman, J.D. & Angelaki, D.E. J. Neurophysiol. 93, 3418–3433 (2005).
National Primate Research Center, University of Washington. BrainInfo http://www.braininfo.org/ (–1991–present).
Zar, J.H. Biostatistical Analysis 4th edn. 592–615 (Prentice Hall, 1998).
Acknowledgements
This work was supported by NIH grant R01 EY012814 (D.E.A.), NIH grant R01DC004260 (D.E.A.) and NIH grant R01 DC014686 (J.D.D.).
Author information
Authors and Affiliations
Contributions
J.L. analyzed the data and wrote the manuscript. B.K. performed the experiments. J.D.D. supervised the experiments and wrote the manuscript. D.E.A. designed and supervised the experiments and wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Integrated supplementary information
Supplementary Figure 1 Recording sites and experiment setup.
Reconstructed cell location based on stereotaxic coordinates and MRI scans. Green (Animal V: n=19; Animal L: n=7), cyan (Animal V: n=12; Animal L: n=2), gray (Animal V: n=4; Animal L: n=4) symbols: G-tuned, dG-tuned and G+dG-tuned cells, respectively. White symbols: non-responsive cells (Animal V: n=28; Animal L: n=19). Blue stars: Head direction cells (n=57) recorded in a head-free setup (data from Kim, B., Dickman, J.D., Tauble, J.S. & Angelaki, D.E. Society for Neuroscience abstr. 444.11/CC2 (2015).). Ca: Tail of the Caudate Gyrus. Fo: Fornix. LV: Lateral ventricle. The dorsal, ventral and medial nuclei of the anterior thalamus are color-coded red, orange and magenta, respectively. (c) Schematic of the motion setup used in the experiments here, which includes a yaw-axis rotator (clockwise and counterclockwise yaw rotations are represented by red/blue arrows) inside a visual enclosure that has the capability of tilting (green arrows) – see also Supplementary Video 1. Note the head-stationary inner most gimbal that moves with the animal, which might have somewhat restricted the view of distant landmarks on the inside of the spherical enclosure. (d) Coordinate system used to represent small angle head tilts. Upper panel: illustration of head orientation relative to gravity in upright (UP), right ear down (RED), nose-up (NU), left ear down (LED) and nose-down (ND) orientations. Lower panel: representation of the head tilt in polar coordinates, with the radius representing the tilt angle and the angular coordinate the direction of the tilt (ND, LED, etc). The changes of head tilt associated with yaw, pitch and roll rotations (bottom right inset) are indicated. The projection of G onto the head-fixed X and Y axes, GX and GY, can be represented in Cartesian coordinates on the same plot (green). Note that this polar representation is a small tilt angle approximation of a 2D spherical topology that represents any arbitrary head orientation.
Supplementary Figure 2 Summary of responses in the horizontal plane with the animal upright
Of 95 cells tested with upright (Fig. 1a) and tilted (Fig. 1b,c) yaw rotations, 5 and 48 cells, respectively, exhibited significant tuning (at p = 0.01, permutation test on the resultant vector length). The 5 cells that responded significantly to rotations around an upright yaw axis, both in light and in darkness (permutation test on the resultant vector, p<0.01) and are classified as azimuth-tuned cells (filled symbols). (a) Resultant (R) vector length during rotation around an upright versus 30° tilted axis. Azimuth responses were weaker (mean R = 0.17 ±0.09 CI for these 5 cells) than responses during rotations around a tilted axis (mean R = 0.37 ± 0.06 CI, n=48, Wilcoxon rank sum test, p = 0.01) (b) Comparison between the resultant vector length during upright rotation in light and darkness. The R vector lengths are not significantly different (Wilcoxon signed rank test, p = 0.8, mean R = 0.15 ± 0.09). (c) Comparison between the Preferred Direction (PD) of cell responses during rotation around an upright versus a 30° tilted axis (computed relative to the visual reference frame). The difference in PD of azimuth-tuned cells was distributed uniformly during CCW (red) and CW rotations (blue) (circular R test p = 0.05 across both directions), indicating that tuning during upright and tilted yaw rotations were independent. (d) Comparison between PD during upright rotation in light and darkness. The difference in PD clusters around 0° (circular mean 0° ± 25° CI).
Using a less stringent criterion (at p = 0.05 on the Rayleigh vector, in either the CW or CCW direction), 34 cells passed the horizontal modulation criterion, 54 cells passed the gravity modulation criterion, and 31 had both. We also recomputed the histograms in c with the larger number (n=34) of cells that would be obtained with less strict criterion for Raleigh vector upright yaw tuning and found that they were not different from uniform (Chi square test, p>0.5 for both directions). Because responses during upright yaw rotation in the horizontal plane were comparatively weak, they have not been further considered here.
Supplementary Figure 3 Comparison between the gravity and gravity-derivative responses during yaw and pitch/roll rotations
(a,b) Response of the example cell in Fig. 1 during pitch and roll rotations. This experiment was repeated in n=48 cells. (c-f) Population analysis. A linear regression (see Methods) was used to compute the cells’ response gain and preferred direction (PD) (e.g. NU, LED, ND…) during tilted yaw rotation (±30° tilt, protocol 2-3). An identical regression analysis was performed on pitch/roll rotation (protocol 4) responses. Thus, we obtained two independent estimates of the cell’s gravity gain and PD, one based on yaw rotation responses and one based on pitch/roll rotation responses. We then compared the gain ratio and difference in PD on a cell-by-cell basis (G-tuned cells: green; dG-tuned cells: cyan; G+dG-tuned cells: gray). (c) The G gain ratios were not significantly different from 1 (bootstrap-based statistics: mean ratio = 0.9, 0.7 to 1.1 CI, n = 34), and (d) PDs were predominantly aligned (mean ΔPD = 10°, -6 to 29° CI). (e) The dG gain ratios were significantly larger than one (mean ratio = 1.6, CI = 1.3 to 2, n=22), likely due to nonlinearities (dG stimuli had a higher amplitude during pitch/roll than during yaw rotation; see on-line Methods). (f) ΔPDs of the dG component were not different from zero (mean ΔPD = 9°, -26 to 8° CI). The similarity in tuning during tilted yaw and pitch/roll movements and independence on the plane/axis of rotation provides strong support that the observed direction tuning arises from G and dG (rather than angular velocity)-coding.
Supplementary Figure 4 Rotation with congruent gravity and visual cues
(a) Response of the example cell in Fig. 1 when the rotation axis was tilted while the visual environment was kept upright (thus, there was no gravity/vision conflict as in Fig. 1b,c). The peak response occurred in nose-down (ND) orientation, as in Fig. 1, indicating that it remained anchored to gravity. This experiment was repeated in n=48 cells. (b-e) R vector strength and (f-i) preferred direction difference summaries (similar in format as Fig. 2). Panels b and f show that the modulation depth (R vector length) was similar (paired Wilcoxon sign rank test, p = 0.1) with (|R|=0.39 ± 0.2 SD) and without (|R|=0.37 ± 0.2 SD) gravity/vision conflict and PDs were generally aligned (|ΔPD|<45° in 80/96, i.e. 83% data points; |ΔPD|>90° in 7/96, i.e. 7% data points). Panels c and g compare CW and CCW rotations for stimuli without conflict. Cells had similar modulation strength (|R| = 0.38 ± 0.2 SD in CW, 0.36 ± 0.2 SD in CCW, p = 0.49) and |ΔPD| spread between 0° (response to G) and 180° (response to dG/dt). Panels d and h show similarity between +30° and -30° tilt rotations without conflict (|R| = 0.39 ± 0.2 SD versus 0.36 ± 0.2 SD, p = 0.05) and most had similar PD (|ΔPD|<45° in 75/96, i.e. 78% data points; |ΔPD|>90° in 9/96, i.e. 9% data points). Panels e and i show similarity between tuning in light and dark (|R| = 0.37 ± 0.2 SD versus 0.38 ± 0.2 SD, p = 0.23) and most had similar PD (|ΔPD|<45° in 80/96, i.e. 83% data points; |ΔPD|>90° in 4/96, i.e. 4% data points).
Supplementary Figure 5 Example cell exhibiting gravity derivative tuning
Responses are shown during (a) yaw rotation around an earth-vertical axis, with the animal upright (in light), (b-d) yaw rotations around tilted axes in light (b,c) and darkness (d), (e-f) pitch and roll rotations (in light). Upper row: illustration of the stimuli. Middle row: neuronal responses as a function of head orientation relative to the visual environment (a-d, top axis) or relative to gravity (b-f, lower axis). The drawings below the plots illustrate head position relative to gravity. Red/Blue raster plots show individual spikes during multiple cycles of CCW/CW rotation (a-d), ±30° forward/backward pitch (e) or ±30° rightward/leftward roll (f). Red/Blue curves show corresponding average firing rate (scale shown in ordinate). The lower plots show the x- and y- component of gravity in an egocentric frame of reference (i.e. Gx, Gy, green curves, which are identical during CCW and CW rotation) and the derivative (velocity) of these components (i.e. dGx/dt, dGy/dt) during CCW (red) and CW (blue) rotations. Note that there is a ≍180° phase shift between the responses to CCW and CW rotation (b-d), indicating that the cell responds to the derivative of head tilt. The cell’s firing rate increases during leftward roll (f, blue curve) and decreases during rightward roll (f, red curve), indicating that the cell responds to movements towards the left side. This is consistent with the responses in b-d: during CCW rotation (b-d), the cell responds maximally in NU orientation when the head rotates towards LED. During CW rotation, the cell responds in ND orientation, which is also when the head rotates towards LED. A leftward head tilt corresponds to Gy > 0 (lower panel, green curves) and a movement towards left tilt corresponds to dGy/dt > 0 (b-f, lower panels, blue and red curves). That is, the cell’s firing rate increases whenever dGy/dt > 0. Note that neuronal modulation is higher in b-d (when dGy/dt peaks at 0.25 G/s) than in f (when dGy/dt peaks at 0.5 G/s). Neuron recorded in Animal V, left hemisphere. This experiment was repeated in n=48 cells.
Supplementary Figure 6 Decomposition of responses into G and dG tuning curves.
(a-d) Example cells (shown for 30° tilt). The responses during CCW (red) and CW (blue) rotations (upper curves) are assumed to be the sum of a G response curve (green, middle plots) and a dG response curve (cyan, lower plots). The G response curves are identical during CCW and CW rotation (since position is the same). The dG response curves are shifted 180° but otherwise identical (since velocity leads position by 90° during CCW and CW rotations). These mathematical constraints are translated into a system of overdetermined equations (see Methods). The G and dG curves are reconstructed by solving this system of equations. The black curves, which are the sum of G and dG tuning curves, as well as the average firing rate (FR0, dashed black line), show how well this decomposition approximates the experimentally-measured responses during both CW and CCW rotations. (a) Example G-tuned cell, where CW and CCW responses have similar phase (resulting in strong G tuning and weak dG tuning). (b) Example dG-tuned cell, where CW and CCW responses are 180° out of phase (resulting in strong dG tuning and weak G tuning). (c, d) Example G+dG-tuned cells. (e) Scatter plot of the partial correlation coefficients of the Gravity (r2G|dG) and Gravity-derivative (r2dG|G) components, used as a criterion for classifying the cells into G-tuned, dG-tuned and G+dG. (f) As a proof of principle, we applied this analysis to simulated Poisson-firing cells with combinations of G and dG cosine-tuned components drawn from the values measured in the population of recorded neurons. The scatter plot illustrates the modulation depth |R| of the G and dG components for each cell. The color code indicates cell classification based on the reconstructed tuning: applying the same analyses as in real neurons, cells with higher G modulation are classified as G-tuned (green), whereas cells with higher dG modulation are classified as dG-tuned (cyan). Intermediate cells are classified as G+dG-tuned (gray). The G and dG-tuned cells had comparable tuning strength (Fig. 2c): the average |R| of the G component in G cells was |R(G)| = 0.37 ± 0.2 (SD) whereas the the average |R| of the G component in dG-tuned cells was |R(dG)| = 0.33 ± 0.1.
Supplementary Figure 7 Simulated response of otolith afferents.
We simulated the response of five otolith afferents whose preferred directions correspond to RED tilt with an angle of 0, 15, 30, 45 and 60° respectively. Otolith afferents are cosine tuned: their firing rate is represented by FR=1+cos(α), where α is the angle between gravity and the preferred direction and the baseline firing rate is assumed to be 1. (a) Two-dimensional tuning curves exhibit a peak (white diamond), but tuning curves are far broader than those typically observed in anterior thalamus cells. (b) Simulated responses to yaw rotation around tilted axes, similar to our experimental protocols. In all cells, the amplitude of modulation always increases when the tilt angle increases. This contrasts with the responses reported here (e.g. Fig. 3b and Supplementary Fig. 8). We conclude that the gravity responses observed in anterior thalamus cells don't reflect unprocessed otolithic signals.
Supplementary Figure 8 Additional examples of two-dimensional tuning.
The neuron (a) fires preferentially at a tilt angle of 20° nose-down and has a very broad tuning curve. In neuron (b), the highest firing rate was recorded at the highest tilt angle used in the study. It is plausible, and likely, that this neuron’s firing rate would increase when tilt exceeds 45°. Thus we cannot determine the location of the neuron’s peak response, but we can conclude that it equals or exceeds 45°. Neuron (c) is one of 3 cells that showed bimodal tuning. Neuron (d) is dG-tuned and fires preferentially at a tilt angle of 11°. Neuron (e) is dG-tuned and fires preferentially at a tilt angle equal or in excess of 45°. In d and e, ND (nose-down), LED (left ear down), NU (nose-up) and RED (right ear down) refer to the direction of tilt velocity. These measurements were obtained in n=31 cells
Supplementary information
Supplementary Text and Figures
Supplementary Figures 1–8 (PDF 1490 kb)
Illustration of the motion stimuli.
All motion stimuli used in this study are shown using an animated computer-generated model of the experimental setup. The setup is about 2.5m high; the size of the animal is exaggerated for better visibility. The black diagram in the upper right corner indicates head orientation in the visual enclosure, and the green diagram in the lower right corner indicates head orientation relative to gravity. (MOV 16093 kb)
Rights and permissions
About this article
Cite this article
Laurens, J., Kim, B., Dickman, J. et al. Gravity orientation tuning in macaque anterior thalamus. Nat Neurosci 19, 1566–1568 (2016). https://doi.org/10.1038/nn.4423
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/nn.4423
