Head direction cell activity in mice: robust directional signal depends on intact otolith organs

Abstract

The head direction (HD) cell signal is a representation of an animal's perceived directional heading with respect to its environment. This signal appears to originate in the vestibular system, which includes the semicircular canals and otolith organs. Preliminary studies indicate the semicircular canals provide a necessary component of the HD signal, but involvement of otolithic information in the HD signal has not been tested. The present study was designed to determine the otolithic contribution to the HD signal, as well as to compare HD cell activity of mice with that of rats. HD cell activity in the anterodorsal thalamus was assessed in wild-type C57BL/6J and otoconia-deficient tilted mice during locomotion within a cylinder containing a prominent visual landmark. HD cell firing properties in C57BL/6J mice were generally similar to those in rats. However, in C57BL/6J mice, landmark rotation failed to demonstrate dominant control of the HD signal in 36% of the sessions. In darkness, directional firing became unstable during 42% of the sessions, but landmark control was not associated with HD signal stability in darkness. HD cells were identified in tilted mice, but directional firing properties were not as robust as those of C57BL/6J mice. Most HD cells in tilted mice were controlled by landmark rotation but showed substantial signal degradation across trials. These results support current models that suggest otolithic information is involved in the perception of directional heading. Furthermore, compared with rats, the HD signal in mice appears to be less reliably anchored to prominent environmental cues.

Figure 1.

Schematic top view of the recording arena (with white cue card indicated by white band) and rotation protocol used to assess HD cells. Each HD cell was recorded during a standard session (1), followed by a rotation session in which the visual cue card was rotated 90° CW or CCW (2), a second standard session (3), a dark session with no visual cue card (4), and a final standard session (5).

Figure 2.

Coronal section of mouse brain at bregma −0.50 mm, stained with thionin. Dashed line indicates boundaries of ADN. Electrode penetration of ADN was verified by the presence of electrode tracks (indicated by arrows) extending through ADN. Scale bar, 300 μm.

Figure 3.

Directional tuning curves of HD cells in C57BL/6J and tilted mice. A, Robust directional tuning of three classic HD cells recorded simultaneously on different electrodes from the same C57BL/6J mouse. Peak firing rate (PFR), preferred direction (PFD), directional firing range (DFR), and background firing rate (BGR) are unique to each HD cell. B, Tuning curve of a representative HD cell recorded from a tilted mouse. Note the variable background firing rate (outside of the directional firing range) for the tilted HD cell, which contrasts with the relatively uniform background firing rates of C57BL/6J HD cells. Rayleigh's r value is noted for each cell.

Figure 4.

A, Directional tuning curve recorded from a bursty ADN cell recorded from a tilted mouse. B, Firing rate × HD × time plot showing the activity peaks of the bursty cell depicted in A. Qualitatively, this ADN cell showed firing characteristics similar to those of HD cells in C57BL/6J mice but lacked a consistent preferred firing direction. As the mouse navigated around the arena in a CW direction, the cell fired in bursts, with most bursts followed by a period of inactivity. Solid line indicates actual HD throughout the recording session, and points indicate HD and time at which the firing rate reached 75% of the maximum firing rate of the cell. C, Firing rate plots depicting cellular activity throughout the 8 min recording session. Top, The bursty cell shown in A and B (burst index, 0.66). Middle, A typical C57BL/6J HD cell (burst index, 0.74). Bottom, A typical C57BL/6J nondirectional ADN cell (burst index, 0.18). Note the presence of distinct bursts, followed by relative inactivity, for the bursty cell and the HD cell, which contrasts with the uniform firing rate of the nondirectional cell. Temporal resolution, 1 bin = 1 s.

Figure 5.

Activity of simultaneously recorded bursty cells in a tilted mouse during head turns. Timescale for firing rate plots in inset corresponds to head direction plot (bottom). At the beginning of the recording session, the animal remained relatively motionless for ∼100 s, after which the animal began to move about the arena. During CCW head turns, activity burst onset for cell 1 (indicated by vertical green line) preceded activity burst onset for cell 2 (indicated by vertical red line). After a transition to CW head turns at t = 155 s, activity burst onset for cell 2 preceded activity burst onset for cell 1. After another transition at t = 193 s, burst onset for cell 1 again preceded burst onset for cell 2. After this period of behavioral activity, the animal remained relatively motionless for the remainder of the recording session, with the exception of a continuous CCW head turn between t = 333 and t = 352 s. Temporal resolution, 1 bin = 1 s; inset, 1 bin = 250 ms.

Figure 6.

HD signal of tilted HD cells degraded across recording sessions. A, Histogram of Rayleigh's r values across standard recording sessions. Most cells from tilted mice had lower r values than C57BL/6J and are clustered to the left in each plot. In C57BL/6J mice, the activity of most HD cells maintained significant directional tuning across recording sessions. In contrast, most tilted cells that showed a significant preferred firing direction in session 1 showed reduced directional tuning across sessions. By the third recording session, nearly one-half of the HD cells became nondirectional. By the fifth recording session, most HD cells in tilted mice became nondirectional, with Rayleigh's r values that fell below the directional significance criterion (dashed line). B, Percentage of HD cells that remained significantly directional across standard recording sessions. In session 1, all HD cells were significantly directional. Across subsequent recording sessions, most C57BL/6J HD cells remained significantly directional, whereas most tilted HD cells became nondirectional. C, Total distance (centimeters) traveled during standard recording sessions for C57BL/6J and tilted mice.

Figure 7.

ATI of HD cells in C57BL/6J and tilted mice. A, Firing rate as a function of head direction during CCW and CW head turns for a C57BL/6J HD cell. Preferred firing directions during CCW (solid line) and CW (dashed line) turns were shifted CW and CCW, respectively. For this session, anticipatory time interval was 36.9 ms. B, Histogram of ATI values for all HD cells recorded from C57BL/6J mice (n = 24) and tilted mice (n = 8) that met criteria for inclusion in ATI analyses. For C57BL/6J mice (open bars), plot indicates continuous range of ATI with no distinct clustering at any ATI; all except two cells show a positive ATI, indicating cell activity best predicted where the rat's head direction would be in the future. Lack of otoconia did not appear to affect the ATI, because most HD cells in tilted mice (filled bars) showed ATI values within the range of C57BL/6J ATI values.

Figure 8.

Shift of preferred firing direction across trials. A, For C57BL/6J mice, two clusters of preferred direction shifts occurred between sessions 1 and 2 (S1-S2) in response to 90° rotation of the visual cue. One cluster consisting of 66.7% of the HD cells (recorded in 9 sessions) appears near the 90° rotation (mean, 82.4°), indicating that these HD cells were heavily influenced by the visual cue card. The second cluster, consisting of 33.3% of the HD cells (recorded in 5 sessions), shows only a slight shift (mean, 16.8°). For tilted mice, rotation of the visual cue caused a rotation of the preferred firing direction for most HD cells (mean, 85.5°). B, C, The preferred firing direction of HD cells in C57BL/6J mice remained fairly stable across recording sessions in which the visual cue was located in the same position relative to the room. For HD cells that remained significantly directional across sessions, cells from tilted mice showed greater instability than those of C57BL/6J mice, indicated by greater mean vector deviation from 0° in the session 1–5 plot (S1-S5) relative to the session 1–3 plot (S1-S3). For presentation, the direction of preferred firing direction shift was standardized to the direction of cue rotation, in which 0° represents the preferred firing direction of each HD cell during session 1. Black points represent individually recorded HD cells, and corresponding colors within each group represent simultaneous records from multiple HD cells. Rayleigh's vector was calculated from the mean angular shift when multiple HD cells were recorded simultaneously. Note that fewer cells are displayed across sessions for tilted mice than for C57BL/6J mice because many cells in tilted mice lost their directional tuning after the first recording session and fell below the r = 0.4 criterion level.

Figure 9.

Loss of robust HD signal across trials in tilted mice. The preferred firing direction of each cell is indicated by vertical solid and dashed lines; Rayleigh's r values are shown for all tuning curves. Left, C57BL/6J mouse: signals from two HD cells recorded simultaneously on different electrodes remained stable across lighted recording sessions. Cell 1 (solid) and cell 2 (dashed) showed robust directionality throughout standard session 1 (S1). In session 2 (S2), cell 1 shifted 96° and cell 2 shifted 84° after a 90° CCW rotation of the visual cue (vertical lines indicate 90° shift relative to each cell's preferred direction in S1). In session 3 (S3), the preferred firing direction of both cells returned to the standard alignment, corresponding to replacement of the visual cue to the standard position. During session 4 (S4) in darkness, both HD cells became unstable, resulting in broad tuning curves with no well defined peak. In session 5 (S5), the cells regained their directional tuning, indicated by return of the preferred directions to the standard alignment. Note that in all sessions recorded during lighted conditions, the background firing rate remained low. Right, Otoconia-deficient tilted mouse: degradation of HD signal across recording sessions for two HD cells recorded simultaneously on different electrodes. In S1, the tuning curves of cell 1 (solid) and cell 2 (dashed) indicate a significant directional preference. In S2, the preferred direction of these HD cells shifted 96 and 66°, respectively, after a 90° rotation of the visual cue, indicating visual information influenced the HD signal. The tuning curves of both cells became slightly broader during S3 relative to S1, indicating increased activity across a greater directional range. In the darkness/no cue session (S4), the preferred directions of both cells maintained the angular relationship to one another despite the absence of visual information. Unlike C57BL/6J HD cells, tuning curves of tilted HD cells showed decreasing directionality across recording sessions. All cell waveforms remained well isolated throughout the recording sessions. STD, Standard session under light conditions.

Figure 10.

Firing rate × HD × time plot showing C57BL/6J HD cell stability characteristics during light and darkness. A linear slope, indicated by a black line, was fit to data from cells in which the preferred direction shifted <360° within the recording session. Cell 1 (A, B) showed an increased firing rate within a limited range of head directions throughout session 1 (standard, 8 min) and session 4 (darkness/no cue, 16 min), indicating HD cell stability during light and darkness. Cell 2 (C, D) showed an increased firing rate within a limited range of head directions throughout session 1 but failed to maintain directional tuning during darkness (session 4), indicating that this HD cell was unstable in the absence of visual cues. Inset in D includes depiction of actual head direction (solid line). Cells 1 and 2 were recorded from different animals. E, Preferred firing direction (PFD) drift for all HD cells recorded during the dark recording session. HD cell stability was indicated by significant directionality (Rayleigh's r > 0.40) and showed a PFD drift of <0.5°/s. HD cells that failed to maintain directionality showed a PFD drift of >0.5°/s. Black points represent cells recorded individually; colored points represent simultaneously recorded cells and correspond to Figure 8 and supplemental Figure S6 (available at www.jneurosci.org as supplemental material). Dashed line indicates significance criterion for directionality.