Whiskers-R-Us
Home Background Research Methods Lab Publications PowerPoint Presentation
RESEARCH STRATEGY AND METHODS FOR THE STUDY OF WHISKING
Problems and solutions
Introduction: While studies of the rodent whisking system has told us a great deal about its anatomy and physiology we know relatively little about its functional organization. We suggest that this may reflect (a) the emphasis of previous studies upon sensory, rather than sensorimotor function, and (b) the related fact that these analyses have lacked an essential behavioral dimension. For example, our knowledge of sensory processing in this system has come from examining neuronal responses to precisely controlled but passively applied displacements of individual whiskers, rather than to the input patterns generated by active contacts. We lack studies which combine recordings from individual neurons and neuronal populations with “on-line” monitoring of whisking behavior in awake, behaving animals. The absence of such correlated neural-behavioral studies severely constrains our ability to understand the operation of a closed-loop sensorimotor system such as whisking. For this reason, we have devoted considerable attention to the development of preparations that would facilitate such studies.
Progress in the analysis of other sensorimotor (oculomotor, somatomotor) systems has involved not only advances in chronic unit recording from behaving animals but also refinements in behavioral methodology. These include immobilization/restraint procedures, “real-time” high-resolution movement transduction and behavioral training paradigms. These improve control of stimulus and response specification and measurement. They permit the experimenter to manipulate and transduce response parameters (e.g., onset, frequency, amplitude,termination) normally under the control of the animal. They are especially important in studies of “voluntary” behaviors (such as whisking) that involve continuous interactions between the acquisition of information and the control of adaptive responses. The application of such procedures to the study of whisking behavior presents a formidable methodological challenge.
Problems:
1. Response monitoring: Whisking behavior may involve the movements of up to 30 individual whiskers occuring at frequencies between 1 and 25 Hz, across a velocity range from < 500° to > 1000°/sec, and over an amplitude range from microns to millimeters. Videographic methods for monitoring whisking have relatively low low spatial and temporal resolution and the time-consuming nature of videographic data analysis restricts behavioral sample sizes, and constrains the study of long-term changes in whisking patterns during discrimination learning.
2. Stimulus control. Each whisker is set in a movable pad that is fixed to a movable head that is attached to a movable body. Even the simplest response sequences emitted by a freely-moving rat may involve the simultaneous activation of a variety of receptor types and loci, and the generation of response sequences involving diverse effector systems (head, limbs). Moreover, sensory input during whisking may reflect exafference produced by whisker contacts, reafference produced by displacement of the whisker, or even “efference copy” related to efferent commands to the vibrissal motor system. All of these processes will engage different components of whisking circuitry. Furthermore, the typical discrimination tasks --maze running or gap-detection --are incompatible with tight stimulus control, while the intrusion of head movements during these behaviors introduces potential confounds.
3. Behavioral control. Whisking is not a reflex (i.e., it is not predictably elicitable by a specified stimulus) and it tends to habituate (i.e. become less probable) fairly rapidly. The use of “free” exploration as a method for neuronal activation, while it produces a “quasi-natural” behavior, precludes experimenter control of whisking behavior. Even when the animal is whisking, the onset and termination, as well as the parameters of its whisker movements (frequency, amplitude, velocity) are under the control of the rat, not the experimenter. The absence of behavioral control severely constrains studies of neural activity in awake, behaving animals.
We have developed instrumentation for the monitoring of whisking movements and contacts that allow us to monitor whisker movements in real time and on a time scale comparable to that of unit activity (i.e., in msec). The use of operant conditioning paradigms provides control of whisking behavior; its onset, termination and movement parameters. By combining operant conditioning and discrimination paradigms with real-time monitoring, we can study the modulation in whisking activity which accompany acquisition of whisker-mediated discriminations. Our collaborationz with the Ahissar, Keller and Kleinfeld laboratories have enabled us to combine these behavioral methodologies with electromyographic and unit recording in awake, whisking animals. We describe here the relevant methodologies.
A. Monitoring of whisker and pad movements in the head-fixed rat.
Figure 2: Monitoring whisking movements: A laser emitter-detector system provides high spatio-temporal resolution (1 ms: 11 mm) of whisking movements (Fig. 2 Left: A). Interruption of the emitted beam (laser curtain) by the shadow of a whisker produces a voltage shift in a subset of shaded sensors (CCDs). Whisker movement results in successive displacements in the position of that voltage shift which are linearly related to whisker position. A comparator circuit identifies the successive positions of voltages above a preset threshold and outputs the data to a microprocessor for computation and display of the whisker movement trajectory. To monitor an individual whisker trajectory with all whiskers present, a light (< 5 mg) self-adhesive foam marker is attached to the side of a vibrissa, which increases its relative “visibility” but does not affect its kinematics (Bermejo, et al. ’98: Table 1. B. Schematic diagram illustrating the basic principle of the monitor system and the procedure for transforming whisker displacements from CCD units to angular whisker positions. In most studies all whiskers on both sides are intact.
NOTE: By positioning two sets of monitoring devices at right angles to each other, it is possible either to monitor whisking in both the AP and DV planes (Fig. 3 Left) or both whisker movements and pad displacements (Fig. 3 Right)
(Fig. 3, above, left). Whiskogram: Movements of the right and left C-1 vibrissa during whisking in air. Top: 8.5 sec. The shaded portion highlights a 850 ms sample. This is displayed at higher temporal resolution at the bottom of the figure for computer-assisted kinematic analysis (amplitudes, peak velocity, rise times, etc) of individual vibrissa movements, using specially written software. Arrows = start, peak and end points of a selected whisk. (from Gao et al.’01). Fig. 3 (above, right: Simultaneous monitoring of whisker and pad movements. Translation movements of the pad (right: mm scale) and rotation movements of a single whisker at a different location (left: degree scale). The whisker and pad record plots the summed movements of the two as monitored from the whisker movement. The bottom record represents an estimate (by subtraction) of the contribution of the pad movement to the whisker record (From Bermejo, et al., in preparation).
1. Control of whisking rate and amplitude. Operant conditioning is used to bring whisking under the control of the experimenter rather than the rat. During training, protraction amplitudes are continuously monitored. Whisks meeting a specified amplitude criterion are reinforced on some trials in the presence of a discriminative stimulus and extinguished on others in its absence- “Go/No Go” (Fig. 4). The rate of response is controlled by varying the “density” (i.e., reinforcers/time interval) of the reinforcement schedule.
Fig. 4 (above). Operant control of whisking by stimulus/reinforcer contingencies. Cumulative records of operant whisking under differential stimulus control.[Left; First conditioning session: Right; last conditioning session](Top) (VI 30s/Ext). whisking is reinforced in the presence of a tone CS (S+, shaded: Go), and extinguished in its absence (S- unshaded: No Go). (Middle) Mult VI 20s/120s). Note the difference in whisking rates under the 20 s (shaded) and 120 s (unshaded) components. (Bottom ). Cue/reinforcer relations are reversed. In all panels whisking response rate is proportional to the slope of the cumulative record, which resets after every 70 responses.
Fig. 5. Conditioned whisking patterns recorded from the Right and Left C-1 whisker during the discriminated operant paradigm shown in Fig 4. (Left panel) An S+ trial (VI 30-s): (Right panel) An S- trial (EXT) corresponding to shaded and unshaded components of the MULT VI 30-s EXT schedule. Note: the presence of large number of whisking responses in the VI30-s component (Left Panel) and the almost total absence of whisking in the EXT component.
1. Operant control of whisker position (Fig. 6): This paradimg requires the head-fixed animal to position and maintain a single marked whisker (C-1) within an electronically defined “window” of variable size and location. On each trial, whisker trajectories are continuously monitored and movement of a marked whisker into and out of the window is signaled, respectively, by the onset and offset of a tone/light pair (CS). Reinforcement is contingent upon the presence of the whisker within the window for a predetermined period (1- 3 s) terminated either by the delivery of a reinforcer or after 20 s. without reinforcement.
Fig. 6. Monitoring of
whisking movements during performance of a conditioned “whisker-positioning”
task. Panel A
shows the position of the identified whisker with respect to the target
“window”. Panel B shows 10 continuous
seconds of whisking plotted separately in the AP and DV planes. At about 9 s, a
single selected, whisking movement is identified by open circles marking its
onset, peak protraction and retraction.
In Panel C the trajectory of that movement is shown in a two-dimensional plot.
Note that movement is minimal in the DV plane
(From
Bermejo,et al.“01)
(A) If the location of the “target” window is randomly shifted on each trial the rat generate periods of active whisking in air under stimulus control, and organized into defined “epochs” (Fig.4, below) across many testing sessions. This makes the paradigm ideal for generating very large numbers of whisks during a single unit recording session. (B) If the whisker is maintained in the same position over successive reinforced trials, the rat tends to reduce its rate of whisking and adopt the reinforced position as its “resting” position for that whisker. (C) By initially reinforcing the whisker’s resting position, and then shifting the reinforced target location gradually over trials, animals can be trained to place their whisker into the cued location.
C. Discriminative whisking behavior. Whisking tasks require that the animal indicate to us that it has received specific information about some feature of the object. It may do so by making an “indicator” response (such as lever pressing or gap-jumping) and varying some property of that response (e.g. rate or jump direction) as we vary some stimulus parameter of the object. Fig 7. illustrates the experimental arrangements and a sample of data gathered on a single trial during the acquisition of a tactile discrimination
Fig. 7. Discriminative whisking in the head-fixed rat. Left: Schematic of the testing situation. Stimuli are rotated into position by a stepping motor. An operant lever press serves as the “indicator” response and movements of an individual whisker are monitored optoelectronically. Right: Top panel: A record of “whisks” recorded from an identified whisker during a single trial. Ticks at the bottom of each record indicate lever-pressing responses. The interruption of the record represents a period during the middle of the trial when data were not being recorded. Middle panel: A low–resolution view of a 7s epoch of individual whisker movements. A cursor-driven analysis program is used to select segments of the record (shaded portion) for analysis. Vertical lines indicate the occurrence of lever presses. Bottom panel: The data segment selected in the middle panel is displayed at higher resolution. Arrows indicate the start, peak and end of a selected whisk [From Harvey,et al ’01).
D. Monitoring of whisker contacts:
Although the optoelectronic system allows us to monitor whisking in air, it provides no information on the nature of the contacts made by the whiskers during exploratory or discriminative behaviors. We use an inexpensive piezoelectric element as a sensor for the detection of vibrissa contacts during whisking. By reinforcing contacts on appropriate schedules of reinforcement we may generate epochs of whisking behavior which model the rat’s “active touch” behavior. By combining the optoelectronic system and the contact detector it is possible to monitor both the trajectory of the whisking movement and the onset and offset of whisker contacts with an object surface.
Fig. 8. Detection of whisker contacts. (Left) Location of the contact sensor with respect to the position of the head-fixed rats. (Right) Several examples of vibrissa contact detection and the whisking trajectories with which these contacts were associated. In each panel, the top trace displays movement data for a one second period prior to and following a defined initial contact. The bottom trace presents a continuous record of variation in the voltage output of the piezo-sensor during the same period. The contact "threshold" of the device is indicated by a dashed line at the bottom of each panel. To cover the entire 1 s period prior to and following the first contact signal, only 1/3 of the data points are plotted. (From Bermejo and Zeigler, ’01).