The Head-Direction Signal Plays a Functional Role as a Neural Compass during Navigation

Abstract

The rat limbic system contains head direction (HD) cells that fire according to heading in the horizontal plane, and these cells are thought to provide animals with an internal compass. Previous work has found that HD cell tuning correlates with behavior on navigational tasks, but a direct, causal link between HD cells and navigation has not been demonstrated. Here, we show that pathway-specific optogenetic inhibition of the nucleus prepositus caused HD cells to become directionally unstable under dark conditions without affecting the animals' locomotion. Then, using the same technique, we found that this decoupling of the HD signal in the absence of visual cues caused the animals to make directional homing errors and that the magnitude and direction of these errors were in a range that corresponded to the degree of instability observed in the HD signal. These results provide evidence that the HD signal plays a causal role as a neural compass in navigation.

Keywords: brainstem; head direction cell; navigation; optogenetics; path integration; thalamus.

Figure 1. Optogenetic targeting of inputs to the HD network

A, Cells in the NPH with projections to the HD circuit were infected with Cre-dependent halorhodopsin via retrogradely transported Cre-recombinase that was injected into the dorsal tegmental nucleus (DTN). HD cells were recorded downstream in the anterodorsal nucleus of the thalamus (ADN). B, Expression of halorhodopsin was confined to the NPH. These NPH→DTN cells were inhibited by 593.5 nm yellow-orange light delivered through a fiber implant; red line indicates the track of the fiber implant. C, After viral infection, animals were screened and recorded in a cylindrical environment under standard (with visual landmark) and dark conditions. See also Figure S1.

Figure 2. Projection-specific inhibition of NPH neurons causes HD cell instability under dark conditions

A, Outline of the experimental timeline. HD cell activity was monitored with and without laser illumination under either standard or dark conditions. For the dark condition sessions, there was no landmark cue present. Table S1 describes how many observations were made under each of these conditions. B, Two HD cells recorded across three sequential 8 min sessions under either standard (top) or dark conditions (bottom). The cell recorded under dark conditions displayed a dramatic shift in directional tuning (~200°) over the course of the photoinhibition session. Grey: animal’s heading; black: headings in which the cell fired at a rate > 70% of its peak firing rate; red: the calculated preferred firing direction (PFD) of the cell. C, Optogenetic disruption of the NPH→DTN inputs causes significant HD cell instability only under dark conditions in eNpHR3.0 rats (n=5) relative to eYFP control rats (n=4) (interaction, three-way ANOVA, Virus x Laser x Condition, F_1,353 = 40.04, P < 0.001). See also Figure S2 and Figure S7.

Figure 3. HD cell network organization is maintained during directional instability

A, The direction of ‘drift’ in HD cells’ PFD tended to correlate significantly either positively (left) or negatively (right) with the animals’ head turns (mean r² = 0.49 ± 0.05). Grey: the animal’s cumulative head turns; red: the cumulative change in the HD cell’s PFD. B, HD cell pairs maintain a constant orientation relative to one another, despite their overall preferred firing direction (PFD) instability. The difference in directional tuning between co-recorded HD cell pairs (n=13) did not change between Dark Laser Off and Dark Laser On sessions (F_1,12 = 0.85, P = 0.38). Grey: animal’s heading; red: the calculated PFD of one HD cell; blue: the calculated PFD of another simultaneously recorded HD cell. C, The instability observed in the HD signal during Dark Laser On sessions is best modeled by a combination of gain change in the inputs to the HD network relative to animals’ actual head turns (in both the CW and CCW direction) and a constant drift term that results in the signal accumulating error at a constant rate. The gain (left) and drift (right) parameters recovered for eNpHR3.0 animals’ individual sessions (Std or Dark Laser On) are indicated by open circles. Dark lines: means of each group. Dashed lines indicate the values corresponding to a perfectly stable HD signal. See also Figures S3, S4, and S5.

Figure 4. Locomotor activity is not affected by inducing HD cell instability

A, Animals’ locomotion is not affected by the optogenetic manipulation (three-way interactions of Virus x Condition x Laser for each measure: F_1,163 = 0.016, P = 0.90; F_1,163 = 0.22, P = 0.64; F_1,163 = 0.09, P = 0.77). B, Experimental (eNpHR3.0) animals’ sampling of the environment did not vary between Dark Laser Off and Dark Laser On recording sessions (left; representative place/time heat maps of sessions from two different experimental animals). The proportion of the session that eNpHR3.0 animals (n = 5) spent in the center portion (radius = 75% of total radius) of the apparatus relative to the periphery did not significantly change between Dark Laser Off and Dark Laser On sessions (right) (paired t₍₄₎ = −2.02, P = 0.11).

Figure 5. Testing path integration through a food-carrying task

A, The food-carrying apparatus was composed of a circular platform (182 cm diameter) with a home refuge placed just below the edge of the platform at one of four possible locations on its periphery. Animals were trained to forage (grey) for a food pellet (green) and carry the food pellet back to the refuge along a homing path (red) for consumption. B, Examples of eNpHR3.0 animals’ foraging (grey) and homing paths (red) under each experimental condition. The blue line indicates the most direct path between the food pellet and the refuge for each trial.

Figure 6. Inducing instability in the HD signal results in inaccurate homing

A, Photoinhibition of NPH caused inaccurate (left, F_1,536 = 19.76, P < 0.0001) and less efficient (more circuitous) homing paths (right, F_1,535 = 15.51, P < 0.0001) under dark conditions in eNpHR3.0 rats (n=8) relative to eYFP controls (n=8). B, Left, animals still displayed a kinematically typical increase in speed and reached a similar peak velocity on their homeward trip, despite their inaccuracy during homing (F_1,536 = 0.89, P = 0.35). Middle, animals’ impairments in homing were not due to longer foraging paths (F_1,536 = 0.39, P = 0.53). Right, animals’ peak error, a measure of their innate ability to estimate the upcoming length of their homing path and modulate their speed accordingly, also was not disrupted (F_1,526 = 1.057, P = 0.30). C, Animals’ behavior is unaffected by manipulating the HD signal after path integration. Left, Examples of homing paths under dark conditions when laser illumination is confined to the homing path alone. Right, Under dark conditions, eNpHR3.0 animals’ (n=2) homing is just as accurate as laser off conditions when laser illumination is confined to the homing path alone (homing error: unpaired t₍₉₄₎ = 0.83, P = 0.41; homing circuity: unpaired t₍₉₃₎ = 1.54, P = 0.13).

Figure 7. The degree and directionality of navigational error matches the instability observed in the HD network

A, The degree to which individual animals (n=4) were impaired in the food-carrying task correlated with the degree of instability observed in their HD cells (under the same recording conditions) and the length of time spent searching for the food pellet on each trial (r = 0.306, P < 0.0001); Blue: fit line; shaded grey area: s.e.m. B, The direction of eNpHR3.0 animals’ (n=8) homing paths in the food-carrying task relative to the direction of their outward path (top) matches the direction of drift in the HD signal relative to the direction eNpHR3.0 animals’ (n=5) head turns in the recording experiment (bottom) (χ² = 2.18, P = 0.34). C, Alternative sources of self-movement information for the HD network are not directly disrupted by the optogenetic methods. The medial vestibular nucleus (MVN) is a likely source for self-movement inputs to the HD circuit and has projections to not only the NPH, but also the dorsal paragigantocellular reticular nucleus (PGRNd) and supragenual nucleus (SGN), which also likely contribute to HD cell stability. The NPH→DTN pathway (orange) is the primary pathway targeted directly in the current approach; it is unknown if the NPH→SGN pathway (red) is also affected, since retrogradely infected NPH cells may also send collaterals to SGN. See also Figure S6.