Degradation of head direction cell activity during inverted locomotion

Abstract

Head direction (HD) cells in the rat limbic system carry information about the direction the head is pointing in the horizontal plane. Most previous studies of HD functioning have used animals locomoting in an upright position or ascending/descending a vertical wall. In the present study, we recorded HD cell activity from the anterodorsal thalamic nucleus while the animal was locomoting in an upside-down orientation. Rats performed a shuttle-box task requiring them to climb a vertical wall and locomote across the ceiling of the apparatus while inverted to reach an adjoining wall before ascending into the reward compartment. The apparatus was oriented toward the preferred direction of the recorded cell, or the 180 degrees opposite direction. When the animal was traversing the vertical walls of the apparatus, the HD cells remained directionally tuned as if the walls were an extension of the floor. When the animal was locomoting inverted on the ceiling, however, cells showed a dramatic change in activity. Nearly one-half (47%) of the recorded cells exhibited no directional specificity during inverted locomotion, despite showing robust directional tuning on the walls before and after inversion. The remaining cells showed significantly degraded measures of directional tuning and random shifts of the preferred direction relative to the floor condition while the animal was inverted. It has previously been suggested that the HD system uses head angular velocity signals from the vestibular system to maintain a consistent representation of allocentric direction. These findings suggest that being in an inverted position causes a distortion of the vestibular signal controlling the HD system.

Figure 1.

Diagram of recording track and position of a rat over the course of a single trial. The rat begins on one side of the floor surface (A). To initiate a trial, the animal approaches and ascends the outer wall (B). When the animal reaches the junction between the wall and ceiling, it pitches backward 90° to move across the ceiling (C) in an upside-down orientation. When the rat reaches the far wall (D), it descends into the reward compartment of the floor (E) for a food reward. The current floor compartment then becomes the start compartment, and the animal reverses direction for the next trial.

Figure 3.

Representative firing rate × HD tuning curves across each of the four surfaces. HD is shown on the abscissa, the firing rate is plotted as a solid line relative to the left ordinate, and the number of samples at each directional bin is plotted as a dotted line relative to the right ordinate using a logarithmic scale. A, Tuning curves from three different HD cells categorized as CD. Note the present but degraded directional specificity while the animal was traversing the ceiling in an inverted orientation. B, Tuning curves from three HD cells categorized as CND. Note the absence of directional specificity while the animal was traversing the ceiling. Plots are constructed using 6° directional bins. deg., Degree.

Figure 2.

Firing rate × HD × time plot for two consecutive trials. Time is represented on the abscissa with the distance between tic marks indicating 5 s. The HD of the animal is indicated as a dotted gray line using the left ordinate, whereas the firing rate of the recorded cell is plotted as the black line on the right ordinate. The label under each panel indicates the current surface being traversed and the general direction of locomotion relative to the preferred direction of the cell. This cell had a preferred direction of ∼0° (east). Note the influence of HD on the firing rate on the floor and when the animal ascends and descends the walls (compare panels labeled preferred and nonpreferred). In contrast, when the animal was locomoting on the ceiling, the firing rate of the cell was generally not affected by HD. The firing rate trace was smoothed by averaging each value across five points forward and backward in time. The HD trace was not smoothed. deg., Degree.

Figure 4.

Frequency distributions of mean vector lengths obtained from Rayleigh analysis of directional tuning. Mean vector length is a measure of directional specificity, with increasing values indicating cellular activity dependent on the HD of the animal. The filled bars indicate cells showing significant (p < 0.05) directional specificity on that surface, whereas the open bars indicate cells failing to show directional specificity on that surface.

Figure 5.

Firing rate × HD tuning curve of a representative HD cell plotted as a function of whether the animal was ascending a wall, locomoting inverted across the ceiling, or descending a wall. Note that the directional specificity is maintained when the animal descended, although the cell was nondirectional on the ceiling. deg., Degree.

Figure 6.

Scatter diagrams showing the amount of angular shift of preferred directions between different surfaces. While traversing the walls, cells tended to show preferred directions similar to the floor (top row). In contrast, preferred directions tended to shift randomly while the animal was locomoting in an inverted position (bottom left panel). Cells showed stable tuning when the east and west compartments of the floor surface are compared (bottom right panel). The angular position of each filled circle represents the shift of preferred direction for individual cells between the two surfaces. For each comparison, only angular shifts from the cells showing maintenance of directional tuning on both surfaces are shown. The dotted arrow denotes an angular shift of zero, and the solid arrow denotes the observed mean vector angle. The length of the solid arrow denotes the mean vector length; a length of 1.0 (no variability in shift scores) is represented by a vector spanning the radius of the circle.

Figure 7.

Mean values (± SEM) of directional characteristics across each of the four surfaces. For each statistic, the directional characteristic was degraded on the ceiling relative to the other three surfaces. Means from each surface were calculated only from cells categorized as directional on that surface. One and two cells had background firing rates of zero on the east and west walls, respectively. Because a zero background rate for these cells would have resulted in an infinite signal/noise ratio, these cells were assigned background values equal to the lowest values within their respective condition for the purpose of calculating the signal/noise ratio. Comparisons were relative to the floor condition using the Wilcoxon signed rank test (^*p < 0.05; ^**p < 0.01).