Behaviors of the rat vibrissal system during an object localization task: acuity, strategies, kinematics, and closed loop effects.
2012-11-08T05:08:20Z (GMT) by
<p>The large mystacial hairs (whiskers) of rodents support a variety of functions, ranging from social interactions to spatial navigation. Whisking behavior has previously been studied in the context of radial object localization, object detection as well as texture and shape discrimination. The role and importance of active whisker movements, their kinematics and motor-sensory coupling in such tasks is, however, still unclear. We addressed these issues in context of a novel horizontal object localization task in unrestrained, behaving rats.</p> <p><strong>PERFORMANCE</strong>. We trained rats in a staircase procedure to discriminate the relative horizontal location of two vertical bars, one on each side of the head. Their localization proficiency was only temporarily affected by removing selected whiskers. Differences between the average best-performances (mm, deg) were insignificant between groups of animals trained either with all (2.7, 5.1), a row (2.8, 5.3) or an arc (3, 5.7) of whiskers from onset of training. Surprisingly, leaving only a single-whisker on either side of the face neither impaired performance and animals in this group reached progressively lower thresholds with further training (1.5, 2.8). The lowest thresholds recorded across all conditions fell in the range of previous estimates of passive deflection detection thresholds of awake rats (0.5-1.63 deg; Hutson and Masterton, 1986). However, localization proficiency was permanently impaired if only a single whisker was left intact on each side from the onset of training (average best performance 4.9 mm, 9.5 deg). Thus, fine discriminations with a single whisker were highly dependent on prior exposure to the task with multiple whiskers.</p> <p><strong>MOTOR-SENSORY STRATEGIES</strong>. Head and whisker movements were recorded with a high-speed (500 fps) video camera and tracked offline using semi-automatic methods. Active whisker movements were ubiquitous and diverse during object localization. Movement amplitudes were lower and periods shorter in the localization task (medians 15 deg, 62 ms) compared to behavior in a simpler search task (medians 26 deg, 78 ms; p < 0.001, Wilcoxon). Across trials, whisker angle at moment of contact was well correlated with object position relative to the head (R = 0.86 +/- 0.1) and the horizontal offset between objects (R = 0.74 +/- 0.08; 4 rats). We also observed that whisker curvature was correlated, albeit weaker, to object position relative to head (R = 0.35 +/- 0.1). Thus, parameters that define both the position and shape of whiskers during object touch contribute to an objective read-out of relative object location. We observed a large variety of behaviors during object localization, suggesting that these different sensory cues are used to different extents by different animals.</p> <p><strong>WHISKING KINEMATICS</strong>. The apparent, volitional modulation of movement amplitudes and periods during object localization was contrasted by stereotyped kinematics across repetitions of the different movement components during rapid whisking. The average protraction trajectory was characterized by an initial, sharp rise in velocity that peaked at about 0.2 duty-cycle. This was followed by a linear reduction in velocity throughout the remainder of the protraction phase. Individual protractions, however, were less smooth and typically had more than one dominant velocity peak. In comparison, most retraction trajectories were characterized by a bell-shaped profile. During protraction, peak-velocity and movement duration increased with protraction amplitude although peak-velocity was independent of the movement duration. During retraction, peak-velocity also increased with movement duration. The distribution of protraction peak-velocities suggests that movement velocity does not saturate during protraction but rather reaches a target value which is maintained throughout protraction.</p> <p><strong>CLOSED-LOOP BEHAVIOR</strong>. Whisk movements were altered both within and across whisks upon contact with objects. During object localization, movement amplitudes and their periods changed rapidly across subsequent cycles following initial contact with objects. On average, movement amplitudes were reduced and the timing of touch occurred progressively earlier in the protraction phase across subsequent movement cycles. These changes coincided with head movements along the horizontal axis in the direction of the objects. A fast (~10ms) within-whisk amplification of the ongoing movement was also observed when whiskers contacted objects during protraction.</p> <p><strong>CONCLUSIONS</strong>. Our results show that the rat vibrissal system has an excellent capability of discriminating horizontal object location. Discrimination thresholds can reach levels comparable to passive deflection detection thresholds, even with just a single whisker intact on either side of the face. However, optimal performance with only a single whisker requires prior exposure to the task with multiple whiskers. Parameters that describe both whisker location and shape contribute to an objective encoding of horizontal object location either relative to the head or in world-centered coordinates. Although fine kinematic features of whisker movements are highly stereotyped, timing and amplitudes of whisker movements are rapidly and continuously modified in response to object contact. These observations provide evidence of closed-loop effects in the rat vibrissal system, whereby motor behavior is continuously modified by sensory input.</p>