Head direction cells
|Jeffrey Taube (2009), Scholarpedia, 4(12):1787.
|revision #91349 [link to/cite this article]
Head direction cells (HD cells) are neurons found in several brain areas that discharge in relation to the animal’s directional heading with respect to the environment in the horizontal (yaw) plane. For example, a particular neuron might discharge whenever the animal points its head northeast, independent of its location. In this way, head direction cells are similar to a compass in that their discharge is always tuned to a particular direction and can fire at any location provided the animal’s head is facing the appropriate direction. However, unlike a compass, head direction cells are not dependent on the Earth’s geomagnetic field, but rather on landmarks and self-motion cues, such as vestibular and proprioceptive cues. Head direction cell firing occurs in relation to the animal’s head direction, and not its body position or orientation of its head relative to its body. Head direction cells discharge whether the animal is moving or still, and is primarily independent of the animal’s on-going behavior. Each cell has only one direction in which it fires maximally. The firing properties of head direction cells in the same environment remain stable across many days and even months. To view a video of a typical head direction cell online go to: . Head direction cells are believed to represent the neural substrate of the organism’s perceived directional heading in the environment and enable accurate navigation.
Because head direction cells use the surrounding environment as its reference (coordinate) frame, and not the organism’s body, they are classified as allocentric. They therefore complement other allocentric cells found in the rat limbic system – place cells in the hippocampus which encode the animal’s location, and grid cells in the entorhinal cortex which discharge at multiple locations within an environment where these locations form a regular repeating pattern, or grid, across the entire environment.
Head direction cells were discovered in the dorsal presubiculum of rats in 1983 by James Ranck, Jr. They have subsequently been observed in other species including mice, chinchillas, guinea pigs, and monkeys. The recordings in monkeys showed that head direction cell activity was not dependent on eye movements or where the monkey was directing its gaze, but rather the orientation of the its head in space.
Most studies on head direction cells have been conducted in rats. The general properties of head direction cells were first described by Taube, Muller, and Ranck in 1990. Each head direction cell can be characterized by a number of parameters based on its tuning curve, which is often plotted on an x-y graph, with firing rate represented on the ordinate axis and the animal’s head direction represented on the abscissa ( Figure 1), although head direction cell firing is sometimes depicted using polar plots. Using an x-y graph, each head direction cell’s tuning curve is generally triangular or Gaussian in shape and contains a single peak. The direction at which the neuron fires maximally is referred to as the cell’s preferred firing direction, and all preferred directions are equally represented within a population of head direction cells. Each head direction cell’s firing rate is at or near zero when the animal’s head is not pointing in the cell’s preferred firing direction, and then increases near linearly as the animal moves its head into the proper orientation. The range of head directions in which the firing rate is above the cell’s baseline firing rate is referred to as the cell’s directional firing range and averages around 90° for most cells, although it can vary from 60° up to about 150° across cells. The maximal firing rate of the cell at its preferred firing direction is referred to as the peak firing rate. Peak firing rates vary across different head direction cells and range from ~5 to 200 spikes/sec. The determinants of a cell's peak firing rate, and the role served by having cells with different characteristic peak firing rates are not known. If an animal is restrained and passively rotated back-and-forth through the cell’s preferred firing direction, a number of different responses are observed: 1) some head direction cells continue to fire in a direction-specific manner, 2) some cells continue to fire, but have a reduced firing rate at the cell's preferred direction, and 3) some cells fail to respond, suggesting that motor/proprioceptive inputs may be required for normal cell activity.
Head direction cells in some brain areas have a secondary correlate with angular head velocity. For these cells, firing rates are slightly higher when the animal is turning its head rapidly through the cell’s preferred firing direction. Nonetheless, even for these cells the firing rate remains high in the cell’s preferred direction when the animal turns its head at slow rates or is motionless. Similarly, most head direction cells discharge at a slightly higher rate when the animal is moving faster (high linear speeds) than when it is motionless. For both linear and angular head velocity, firing rate increases are less than 10% of the cell’s peak firing rate, and the correlation to angular velocity only accounts for about 1% of the firing rate variance. Head direction cells in the anterodorsal thalamus, but not the dorsal presubiculum (see below), actually anticipate the animal's future heading by about 25 msec. Thus, a plot of head direction tuning curves that distinguish between clockwise and counter-clockwise turns leads to functions that have slightly offset peaks (See Figure 2). This difference in peaks widens as the animal increases the speed of its head turns.
Some studies have examined head direction cell properties when the animal is locomoting in planes other than the azimuth (Earth horizontal). When locomoting in the vertical plane, head direction cells display ‘normal’ direction-specific firing, where firing occurs in only one direction within the vertical plane. For instance, a particular cell might fire only when the animal points its head upward to the left independent of where it is located on the vertical surface and independent of how the head is oriented relative to its trunk. This finding is consistent with the notion that head direction cells define the horizontal reference frame as the plane of locomotion of the animal. Thus, when the animal is locomoting in the vertical plane, the neuron treats this plane as its horizontal plane and cell firing is similar to that seen when on the ground. These findings in the vertical plane were replicated in a shuttle box task that required rats to climb a vertical wall and then locomote inverted across the ceiling of the apparatus to reach an adjoining vertical wall where they descended into a reward compartment. Although directional discharge was normal along the walls – either ascending or descending – it was severely degraded when the rats traversed the ceiling inverted, but returned immediately when the animal moved back onto the wall. Directional firing was compromised on the ceiling even in rats that had been well-trained to perform this task for months. Why the directional signal should be disrupted only when an animal is inverted is not known, and suggests that inputs from the otolith organs may be important for normal head direction cell discharge.
Landmark and Cue Control of the Preferred Firing Direction
A major distinction is drawn between the mechanisms and types of cues involved in generating the directional signal, and those that are responsible for controlling the signal’s preferred firing direction. Two different strategies are used by animals for navigation - landmark and path integration. Landmark navigation occurs whenever an animal derives its current position and orientation in the environment relative to surrounding landmarks. The sensory information the animal uses is usually referred to as allothetic cues, and can be obtained from any of the sensory modalities (e.g., visual, auditory, olfactory). In contrast, in path integration, the animal registers its starting position and orientation, but thereafter, estimates its current location and direction by integration of internally available self-motion information, such as proprioceptive and vestibular inflow, and motor outflow (''efference copy''). The sensory/motor systems involved in path integration are referred to as idiothetic cues.
The preferred firing direction of a head direction cell is controlled by a number of different types of allothetic cues. The most common cue studied has been visual landmarks. When head direction cells are recorded in a cylinder-shaped arena that contains a prominent visual cue attached to the inside wall, rotation of the salient visual landmark leads to a corresponding shift in the head direction cell’s preferred firing direction, indicating that head direction cells can be controlled by landmarks ( Figure 3). Odor cues have also been shown to be an effective landmark for head direction cells. Background visual cues are more effective at cue control than foreground visual cues. A prominent novel visual cue can gain control over a cell’s preferred firing direction within minutes of exposure. Removing the familiar visual cues or turning off the lights, however, does not lead to a change in cell activity, although the preferred firing direction may drift after some time or shift to a new orientation. Thus, the preferred firing direction of head direction cells can be maintained purely via idiothetic information as the animal moves about its environment in the dark. When the lights are restored the visual landmark can reset the cell’s preferred firing direction within 80 msec. Importantly, when multiple head direction cells are recorded simultaneously, landmark removal leads to an equal angular shift in the preferred direction of all cells by a random amount ( Figure 4). This finding indicates that afferent input driving one head direction cell similarly influences other head direction cells, and that head direction cells within a particular brain area behave as a network and their preferred directions always remain a fixed angle apart (in register) from one other. Finally, there is also evidence that optic flow (a self-movement cue) can be used by head direction cells to update its signal.
Several studies have explored how head direction cells respond when information from landmark sources conflicts with idiothetic (self-motion) information about the animal’s movements. Most studies have found that when well-established and prominent visual cues are placed in conflict with idiothetic cues, the spatial information derived from visual landmarks usually overrides that of idiothetic information, and head direction cells respond by aligning their preferred directions based on the spatial information from landmarks.
Generation of the head direction signalHead direction cells were first identified in the dorsal presubiculum (sometimes referred to as the postsubiculum), an area within the subicular complex, which lies adjacent to the hippocampus. Since their discovery, head direction cells have been found in several other brain areas, many of which are within the classic Papez circuit. They are particularly abundant in the anterodorsal thalamus. Other areas where they have been identified include the lateral mammillary nuclei, retrosplenial cortex, lateral dorsal thalamus, striatum, and entorhinal cortex ( Figure 4). parietal cortex, but is dependent on intact lateral mammillary nuclei. Lesions of the postsubiculum, however, lead to a disruption in the ability of a visual cue to control the preferred firing direction of an head direction cell, suggesting that the postsubiculum is critical for processing landmark information.
Current theories postulate that the head direction signal originates in the reciprocal connections between two subcortical nuclei - the dorsal tegmental nucleus and lateral mammillary nuclei. The dorsal tegmental nuclei receive prominent inputs from the supragenual nucleus and the nucleus prepositus hypoglossi, both of which receive major innervation from the vestibular nuclei. The dorsal tegmental nuclei contain a large percentage of cells that are sensitive to the animal’s angular head velocity. From the lateral mammillary nuclei, which contain both angular head velocity sensitive and head direction cells, the head direction signal is projected to the anterodorsal thalamus → postsubiculum → entorhinal cortex → hippocampus. In this way directional heading information can be integrated with information about the animal’s location in the hippocampus and entorhinal cortex. Together, information about place and directional heading provide a sense of one’s spatial orientation in the environment.
The vestibular system is critically involved in the generation of the head direction signal as interference of the vestibular signal disrupts the directional firing patterns in head direction cells. Motor signals, whether proprioceptive feedback from muscles or efference copy signals from cortical motor systems, also play an important role in head direction signals. For example, head direction cells maintain their preferred firing direction when an animal locomotes into a novel environment where it is unfamiliar with the surrounding landmarks. However, when the animal is transported passively via a small cart into the novel environment, the preferred firing directions are not maintained between the two environments. Because passive transportation does not disrupt vestibular cues, but only deprives the animal of motor efference and proprioceptive cues, these findings indicate that these cues play an important role in updating the head direction signal when the animal undergoes normal locomotion.
Of course, head direction cells maintain their directional firing properties even when the animal is still, and there is little, if any, reduction in cell firing when the animal points its head continuously in the cell’s preferred firing direction. Thus, although motor/proprioceptive feedback systems can influence head direction cell firing, normal directional activity can continue in the absence of movement. The problem of what sustains directional firing in the absence of movement is usually solved computationally by applying the principles of continuous ring attractor networks. These network models, which involve recurrent excitation onto head direction cells of similar preferred directions and inhibition of cells with different preferred firing directions, can sustain activity without outside excitation. The ‘hill of activity’ is moved around the ring to different directional headings following inputs from idiothetic or allothetic sources.
- Sharp PE, Blair HT, Cho J (2001) The anatomical and computational basis of the rat head-direction cell signal. Trends in Neuroscience 24: 289-294.
- Taube JS (2007) The head direction signal: Origins and sensory-motor integration. Annual Reviews of Neuroscience 30: 181-207.
- Taube JS, Muller RU, Ranck JB Jr (1990a) Head-direction cells recorded from the postsubiculum in freely moving rats. I. Description and quantitative analysis. Journal of Neuroscience 10: 420-435.
- Taube JS, Muller RU, Ranck JB Jr (1990b) Head-direction cells recorded from the postsubiculum in freely moving rats. II. Effects of environmental manipulations. Journal of Neuroscience 10: 436-447.
- Wiener SI, Taube JS (2005) Head Direction Cells and the Neural Mechanisms of Spatial Orientation. MIT Press: Cambridge, MA, USA.
- Valentino Braitenberg (2007) Brain. Scholarpedia, 2(11):2918.
- Keith Rayner and Monica Castelhano (2007) Eye movements. Scholarpedia, 2(10):3649.
- Edvard Moser and May-Britt Moser (2007) Grid cells. Scholarpedia, 2(7):3394.
- Peter Jonas and Gyorgy Buzsaki (2007) Neural inhibition. Scholarpedia, 2(9):3286.
- Rodolfo Llinas (2008) Neuron. Scholarpedia, 3(8):1490.
- Wolfram Schultz (2007) Reward. Scholarpedia, 2(3):1652.
- Philip Holmes and Eric T. Shea-Brown (2006) Stability. Scholarpedia, 1(10):1838.
- S. Murray Sherman (2006) Thalamus. Scholarpedia, 1(9):1583.
- Kathleen Cullen and Soroush Sadeghi (2008) Vestibular system. Scholarpedia, 3(1):3013.