| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Physiology, University of Munich, 80336 Munich, Germany
Submitted 11 June 2003; accepted in final form 23 July 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
), preferentially from afferent vestibular nerve fibers that remained intact after RA nerve section (i.e., posterior vertical canal and lagena). In addition to this expansion of afferent nerve signals, excitatory commissural inputs increase in their efficacy as well, whereas inhibitory commissural inputs become weaker (Goto et al. 2001). These changes in chronic RA frogs (i.e., about 2 mo after RA nerve section) may be considered to be beneficial because they tend 1) to restore the resting discharge in disfacilitated 2°RA neurons, 2) to reduce the bilateral asymmetry in the spontaneous discharge rates of central vestibular neurons, and 3) to enlarge the depth of modulation for remaining inputs.
However, these additional vestibular inputs can be expected to alter the spatial response tuning of disfacilitated 2°RA neurons. If the synaptic strength of these new inputs were strong enough and the projection patterns of the axons of disfacilitated 2°RA neurons remained unaltered, vestibular reflexes with inappropriate spatial characteristics might develop. The horizontal maculo-ocular reflex is of particular interest in this respect for the following reasons. First, this reflex is not organized in a push–pull fashion (as is the horizontal canal–ocular reflex) because it lacks an uncrossed inhibition (Rohregger and Dieringer 2002). As a consequence, the resting activity of abducens motoneurons on the intact side is still modulated acutely after unilateral labyrinthectomy (UL) during horizontal angular acceleration through ipsilateral inhibitory projections, but no longer in response to horizontal linear acceleration. Second, if the responsiveness to horizontal linear head acceleration recovers postoperatively, the signals can be expected to originate in the remaining utricle and to be mediated by commissural fibers to disfacilitated 2°RA neurons. Because the polarization vectors of utricular hair cells cover 360° and because uncrossed inhibition is absent, the direction of the spatial response vector for recovered abducens nerve best responses could be reoriented in any possible direction in the horizontal plane.
For control frogs the spatial orientations of the response vectors of extraocular motor nerves for angular or translational vestibulo-ocular reflexes are precisely known (Rohregger and Dieringer 2002). This detailed information provides a solid platform for a comparison with data obtained from chronic RA frogs. Abducens nerve responses evoked by horizontal linear acceleration are activated in controls by hair cells that are located on the contralateral utricle in a sector medial to the striola (Wadan and Dieringer 1994). Recent in vivo studies characterized the width of this sector and its vector orientation in canal coordinates (Pantle and Dieringer 1998; Rohregger and Dieringer 2002). A possible contribution from the lagena, a vertically oriented macula organ that is the functional equivalent of the mammalian sacculus (see Straka et al. 2002), was studied but no contribution was detected (Rohregger and Dieringer 2002). Because the lagenar nerve branch remains intact after RA nerve section, lagenar afferent nerve signals might expand onto disfacilitated 2°RA neurons and could possibly contribute to maculo-ocular reflexes in chronic RA frogs. We therefore determined in chronic RA frogs the direction of the vector for abducens best responses and the presence of a possible contribution from the lagena to compare these results with control data.
Whereas all labyrinthine endorgans were routinely removed during surgery for UL, they remained in situ after RA nerve section. As a consequence, a reinnervation of vestibular endorgans by outgrowing, regenerating afferent nerve fibers was excluded after UL but had to be expected after RA nerve lesion (Goto et al. 2002). We used this opportunity to investigate in addition a possible reversibility of the functional consequences of a postlesional synaptic reorganization for maculo-ocular reflexes. We therefore investigated and compared the spatial tuning of contralesional abducens nerve best responses in chronic RA frogs with and in chronic RA frogs without a functional reinnervation of the utricular macula. Preliminary results were published in abstract form (Rohregger and Dieringer 2003).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Surgery was performed in deeply anesthetized grass frogs (0.1% MS-222; 3-aminobenzoic acid ethyl ester dissolved in tapwater; Sigma; Rana temporaria; n = 38) and consisted of 3 steps. Two months before an experiment the otic capsule was opened on the right side and the ramus anterior (RA) of N.VIII was sectioned under visual control distal to the entry of the saccular nerve branch (see Fig. 1A). The distal transected end of the nerve was bent peripherally to impede regeneration. This nerve section eliminated afferent nerve inputs from the utricle, the anterior vertical, and the horizontal semicircular canal, whereas inputs from the posterior vertical semicircular canal, from the lagena, and from auditory organs were spared (Fig. 1A). In some RA frogs the labyrinthine cavity was opened again about 1 mo later and outgrowing RA fibers were sectioned once more close to the previous nerve section to reduce the possibility of a reinnervation of the labyrinthine endorgans. In later experiments the RA nerve was sectioned proximal with respect to the saccular nerve branch. The distance for regenerating RA fibers was thereby larger and reinnervation of the utricle was significantly delayed. Because the saccular nerve mediates vibratory and acoustic signals in frogs (Lewis and Narins 1999) this more proximal RA nerve section had the same functional consequences for the vestibular system as an RA nerve section that spared inputs from the saccule. About 2 mo after the 1st RA nerve lesion the forebrain and parts of the diencephalon were disconnected from the brain by electrocoagulation, the abducens nerve on the intact (left) or on either side was dissected free for multiunit recordings, and in some cases the labyrinthine capsule on the operated side was opened again to prepare the animal for labyrinthectomy after the 1st recording session. After recovery from anesthesia frogs were put back to their home cages. The next day the decerebrated frog was immobilized (tubocurarine 0.03 mg intralymphatic injection) and electrodes were attached for stable long-term multiunit recordings (for details see Pantle and Dieringer 1998). The National Institutes of Health Principles of Laboratory Animal Care were followed, and permission for the experiments was granted by Regierung von Oberbayern (211-2531-98/99).
|
Stimulation
The frog was placed in a small box, with the maxilla oriented parallel to the ground plate of the box. Body and eyes were covered with moist gauze to prevent desiccation and vision. The box was mounted on a platform that was attached to a linear sled. Head orientation with respect to sled motion could be altered by turning the platform in the horizontal plane (Fig. 1B). The head of the frog was pitched up by 15° to bring the horizontal semicircular canals close to the earth-horizontal plane (standard head position; see Blanks and Precht 1979). The maximal activation direction (MAD) of the abducens nerve was determined with the null-point technique (Estes et al. 1975). To that end the orientation of the static head position on the sled was systematically altered before each horizontal linear acceleration test in steps of 5, 10, or 15° over a range of 180° (0° corresponded to an acceleration along the body length axis and 90° to an acceleration along the interaural axis). For a more detailed description of the stimulation protocol see Rohregger and Dieringer (2002).
Combinations of horizontal and vertical linear accelerations were delivered in a ramplike manner (Fig. 1C). To compensate the inclined static position of the frog in space and to keep the head in standard position during each test, a platform on the sled could be tilted by 
90° (see Fig. 1C). A horizontal linear velocity step (peak acceleration 11% of g) in the ON-direction of the recorded abducens nerve was used to measure the onset latency of the evoked responses. The response sensitivity was analyzed with sinusoidal horizontal linear oscillations at a frequency of 0.5 Hz and different peak velocities.
Data processing
The recorded multiunit nerve discharge was amplified, filtered (band-pass 300–1,500 Hz, Krohn Hite 3550), rectified, and displayed on an oscilloscope together with the upper and lower trigger levels of a spike amplitude window discriminator (Mentor, N-750). Trigger levels were set above background noise and below the peak of small, spontaneous action potentials. Normalized action potentials delivered by the window discriminator were stored together with the sled position on the computer (CED 1401 and Spike 2, Cambridge Electronics Design) for off-line analysis. The instantaneous firing rate was calculated from reciprocal interspike intervals and smoothed with a uniform average filter (0.3 s corresponding to a corner frequency of 1.66 Hz). The onset latency of responses evoked by velocity steps was calculated as the time interval between the onset of sled motion and a point at which the smoothed average instantaneous firing rate at rest and in response to sled motion diverged. Responses to several (4–30, not necessarily consecutive) cycles of oscillations were averaged after the elimination of blink-related bursts. If the response exhibited only one peak per stimulus cycle, a sine wave was fitted to the modulated part [a · sin (
t - 
) + b] after having removed electronically the unmodulated part. The parameter was fixed to the stimulus frequency. The parameters 
and were fit values representing the response magnitude and the phase value, respectively. The parameter b represented the offset value. If the response exhibited 2 peaks per stimulus cycle (see, e.g., Figs. 4G, 5, A and D) a sine wave as described above was fitted, however, to each of the 2 response half cycles separately. The range of head positions over which 2 response peaks per horizontal linear acceleration cycle occurred defined the width of the opening of a functional sector on the contralateral utricular macula, as originally introduced by Wadan and Dieringer (1994). The phase of the response was related to the phase of maximal sled acceleration and positive values indicate that the response was lagging sled acceleration. The depth of the modulation of the responses varied between different preparations, in part because of the variable number of axons recorded from. To facilitate a comparison of data from different individuals, we normalized the responses by taking the largest response of a given nerve in a series of experiments as 100%. Averages from a population of animals were expressed by mean values ± SDs. These SDs were used to outline the range of control responses (gray tone in Fig. 3) to facilitate a comparison with data from chronic RA frogs.
|
|
|
MAD directions in the horizontal plane were defined by the stimulus direction orthogonal to the minimal activation direction in the horizontal plane. Ramplike stimulation with the frog positioned in horizontal MAD direction produced 2 different types of responses. One response type, mainly seen in intact frogs or chronic RA frogs with regeneration, consisted of one response maximum per stimulus cycle that decreased in amplitude as a cosine function with the angle of sled inclination (see Fig. 6B). This response was fitted by a sine wave a · sin ( - ), where a is the amplitude, is the angle of inclination, and is a possible inclination or depression of the MAD. The 2nd response type, mainly seen in chronic RA frogs without regeneration, consisted of one or 2 response maxima per stimulus cycle depending on whether the operated side of the frog was facing upward or downward (see Fig. 6D). Responses with 2 maxima per stimulus cycle were fitted by 2 separate sine waves: the one had a peak at 0° and the other at 90° inclination with respect to earth horizontal [a1 · sin () and a2 · sin ( - 90°)]. Responses with one maximum per stimulus cycle were fitted by the sum of the 2 sine waves [a1 · sin () + a2 · sin ( - 90°); see Fig. 6E]. From the ratio of the 2 amplitudes a1 and a2 the elevation or depression of the MAD inclination was calculated [ = a · tan (a2/a1)].
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
). Significant differences in the mean resting rates between RA frogs and controls or between the abducens nerves recorded on the operated or on the intact side of RA frogs were not detected. As a rule, the resting rates remained stable over time; otherwise results were discarded. Spontaneous saccade-related bursts of activity were absent and blink-related bursts of activity were removed electronically before response cycles were averaged (see Agosti et al. 1986; Pantle and Dieringer 1998; Pantle et al. 1995; Rohregger and Dieringer 2002; Wadan and Dieringer 1994).
In view of a possible functional reinnervation of the utricular epithelium by regenerating afferent nerve fibers we established criteria that allowed a subdivision of chronic RA frogs. However, the rate of postoperative normalization of head and body posture was very similar in both groups (Goto et al. 2002). Therefore no prediction could be made from the normalized head posture as to whether a utricular reinnervation was present. Instead, we used the latency and the sensitivity of abducens nerve responses to horizontal linear acceleration as criteria. In analogy to results in an earlier study (abducens nerve responses of frogs 2 mo after UL to angular velocity steps; Agosti et al. 1986), we expected to find significantly lower sensitivities and longer latencies in chronic RA frogs without macular reinnervation than in controls. In chronic RA frogs with a macular reinnervation, however, these response parameters might have normalized again (see Goto et al. 2002). Accordingly, we assumed the absence of a functional macular reinnervation if the response sensitivities and latencies were significantly different from control values. The assumed presence of a functional macular reinnervation was tested in some chronic RA frogs by comparing the abducens nerve responses recorded before and immediately after UL on the side of the former RA nerve section (see following text).
Recovery of responses attributed to central reorganization or to peripheral reinnervation?
The multiunit abducens nerve responses evoked by sinusoidal horizontal linear oscillation in controls consisted predominantly of an excitatory half cycle (Fig. 2A). This response asymmetry is readily explained by low discharge rates at rest (only a fraction of abducens motoneurons are spontaneously active; Dieringer and Precht 1986). The depth of modulation of the discharge rate depended on the angle between the static head position on the sled and the direction of linear acceleration in the horizontal plane. At the beginning of an experiment, we always analyzed first the orientation of this maximal activation direction (MAD), before we continued to characterize the sensitivity of abducens nerve responses for accelerations in this direction.
|
CONTRALESIONAL ABDUCENS NERVE RESPONSE SENSITIVITY. In controls sinusoidal linear acceleration in the horizontal plane evoked abducens nerve responses that increased rapidly in magnitude in relation with the stimulus intensity up to about 8% of g (Fig. 2B; Table 1). At higher stimulus intensities the response magnitude tended to saturate. The peak discharge rates differed between individuals, probably in correspondence with the variable number of recorded abducens nerve fibers. However, the normalized response characteristics (response saturation was taken as 100%) recorded in a population of control frogs (Fig. 3A) were rather similar between different individuals. To facilitate a comparison between these control data and data obtained from chronic RA frogs the average values ± SDs of the population responses of controls were outlined in gray tone (see Fig. 3).
|
Linear horizontal acceleration toward the operated side of chronic RA frogs evoked a response in the abducens nerve on the intact side (Fig. 2C). This response was a clear demonstration of a functional gain recovery, given that no responses can be evoked by similar or even stronger accelerations acutely after RA nerve section on the intact side (Rohregger and Dieringer 2002). At higher stimulus intensities the responses in chronic RA frogs reached average peak discharge rates that were very similar to those of controls (compare responses for a stimulus intensity of 19.2% of g in Fig. 2, A and C and B and D). Peak discharge rates in RA frogs with or without utricular reinnervation or between abducens nerve responses recorded on the intact or on the operated side of chronic RA frogs were not significantly different.
The sensitivity of responses below saturation level was measured as the rate of increase in response magnitude with increasingly higher stimulus intensities. In the control frog shown in Fig. 2, A and B this response sensitivity was 69 imp/s per 1% of increase in g. In the chronic RA frog 10 shown in Fig. 2, C and D, however, it was only 15 imp/s per 1% of increase in g. The more sensitive responses of control frogs tended to saturate already at stimulus intensities above 5% of g (Fig. 2B), whereas the responses of the chronic RA frog 10 reached saturation level at a stimulus intensity of more than 20% of g (Fig. 2D). A similarly low response sensitivity was seen in other chronic RA frogs (e.g., frog 6 in Fig. 3B). The responses of this frog exhibited a clear threshold (at about 4% of g), a slow rate of increase and response saturation at stimulus intensities above 20% of g (Fig. 3B). However, other chronic RA frogs exhibited responses to the same stimuli that were practically identical to those of control frogs (i.e., most data points overlapped with the range of control data; see Fig. 3C). These results were taken as evidence against (frog 6 in Fig. 3B) or in favor (frog 7 in Fig. 3C) of a functional utricular reinnervation in chronic RA frogs.
For a population of 20 chronic RA frogs we characterized the abducens nerve response sensitivities during horizontal linear acceleration. Ten of these individuals were classified as chronic RA frogs without evidence for a functional utricular reinnervation. Linear acceleration evoked in these animals responses that were significantly less sensitive and that saturated at much higher stimulus intensities than in controls (Fig. 3D, Table 1). For the remaining 10 chronic RA frogs we assumed the presence of a utricular reinnervation. The response parameters of these frogs for linear acceleration were practically identical to those of controls (Fig. 3E, Table 1).
IPSILESIONAL ABDUCENS NERVE RESPONSE SENSITIVITY. The immediate consequences of UL or RA nerve section for the response characteristics and sensitivity of the ipsilateral abducens nerve to horizontal linear acceleration was studied by comparing the responses before and a few minutes after the appropriate nerve lesion. Horizontal linear acceleration in the direction of the MAD of the recorded abducens nerve (Fig. 4, A, C, and E) or perpendicularly to this direction (Fig. 4, B, D, and F) evoked responses that were practically identical (Fig. 4G). The sensitivity of this abducens nerve for linear horizontal acceleration of different intensities remained again unaltered after ipsilateral UL (Fig. 4H). The implications of these findings are discussed in the following text.
In 4 chronic RA frogs we compared the bilateral abducens nerve response sensitivity for horizontal linear acceleration by simultaneously recording the responses on either side of the brain stem. Two of these chronic RA frogs belonged to the group of frogs with and the other 2 to the group of frogs without utricular reinnervation. The abducens nerve responses measured on the operated side of these 4 and of 3 other chronic RA frogs were very similar with respect to each other. The mean sensitivity of these responses for linear sinusoidal acceleration was not significantly different from control data (Fig. 3F; Table 1).
ABDUCENS NERVE RESPONSE LATENCIES TO STEPS OF LINEAR ACCELERATION. Horizontal linear velocity steps (peak acceleration 11% of g) evoked abducens nerve responses in controls with a short average onset latency of about 27 ms (Table 1). The abducens nerve response latencies recorded on the operated side of RA frogs and on the intact side of RA frogs with utricular reinnervation were similarly short as in controls (Table 1). The response latencies of the abducens nerve on the intact side of RA frogs without utricular reinnervation, however, were significantly delayed (Table 1). Therefore differences in response sensitivity and in response latency independently supported our classification of chronic RA frogs as belonging to subgroups with or without a functional utricular reinnervation.
Our results concerning the spatial tuning of abducens nerve responses in chronic RA frogs were organized in the following sections:
Spatial tuning of contralesional abducens nerve responses in chronic RA frogs without utricular reinnervation
Similar to our protocol for the study of the spatial tuning characteristics of abducens nerve responses in controls (Rohregger and Dieringer 2002) we first determined the direction of horizontal linear acceleration that evoked maximal responses of the recorded abducens nerve (MAD) and used this information to characterize the abducens nerve responses in addition for ramplike and vertical linear accelerations in a population of chronic RA frogs (n = 10).
RESPONSES TO HORIZONTAL LINEAR ACCELERATION. Horizontal linear oscillation evoked abducens nerve responses that depended in magnitude and phase angle on the orientation of the static head position in the yaw plane (Fig. 5, A–C). Large responses were evoked in controls by oscillations at a frequency of 0.33 Hz along or close to the interaural axis (90° in Fig. 5A). Clockwise reorientation of the frog's static head position in yaw toward an oscillation along the body length axis (0°) resulted in a decrease in the amplitude of this response (closed squares in Fig. 5, A and B) and in the emergence of a small 2nd-response component (open circles in Fig. 5, A and B) that was phase-shifted by about 180° with respect to the initial component (Fig. 5C). This new response component increased and the initial response component decreased in magnitude the more the static head position was reoriented toward an interaural oscillation (-90° in Fig. 5, A and B). Instead of a "null point" (no response at a particular head position) a range of head positions with 2 response peaks per stimulus cycle was present (shaded area in Fig. 5B). This range of head positions defined the width of the opening of a sector on the utricle from which the activation of the contralateral lateral rectus muscle originated (see Wadan and Dieringer 1994). The average, calculated "null point" (two response peaks of equal amplitude per stimulus cycle) for a population of control frogs was -20° (±8°; Rohregger and Dieringer 2002). The orientation of the MAD of the responses is orthogonal to this null point (i.e., 70°; see Fig. 7A) and represents the orientation of the axis of symmetry of the utricular response sector. In controls, the lateral rectus sector on the utricle was therefore defined by a width of about 60° and an axis of symmetry of about 70° lateral with respect to the body length axis (Fig. 7A; Rohregger and Dieringer 2002).
|
In chronic RA frogs the magnitude and phase values of recovered abducens nerve responses on the intact side depended on the static head position with respect to the direction of sled motion as in controls (Fig. 5, D–F). Because of the lower response sensitivity the peak discharge rates evoked by a standard horizontal linear acceleration were smaller than in controls. To evoke responses that were similar in magnitude to those of control frogs, we used a stimulus intensity for chronic RA frogs that was higher than for controls. Because the length of our sled was limited to about ±50 cm we increased the frequency of oscillation from 0.33 to 0.5 Hz (see Fig. 5, A and D). With these stimulus protocols, the depths of modulation of the evoked responses in chronic RA frogs and in controls were very similar (compare Fig. 5, B and E). The null points of responses (about 45° in Fig. 5, D and E), however, and the MADs recorded in chronic RA frogs differed strongly between individual RA frogs (see Fig. 7C; Table 2). The range of static head positions (between 15 and 95°; mean 55 ± 25°) over which the abducens nerve responded with 2 maxima per stimulus cycle was comparable to those in controls. The orientation of the MAD vectors of abducens nerve responses was in most chronic RA frogs turned toward a more caudal direction than in control frogs, but varied considerably (Fig. 7C; Table 2).
|
RESPONSES TO VERTICAL LINEAR ACCELERATION. The response evoked by horizontal linear acceleration along the MAD of the abducens nerve declined in magnitude as a function of the inclination of the sled with respect to earth-horizontal (Fig. 6, A and B; see METHODS). Vertical linear oscillation evoked no abducens nerve responses in control frogs (Fig. 6, A and B; see Rohregger and Dieringer 2002). In all chronic RA frogs without utricular reinnervation (n = 10), however, vertical linear oscillation evoked a response in the contralesional abducens nerve, either during the upward (see Fig. 6, D and E) or during the downward motion of the sled. This response component evoked by vertical linear oscillation originated from the lagena (the vertical macula organ of frogs), either on the operated or on the intact side. Ramplike linear acceleration evoked in chronic RA frogs a combination of utricular and lagenar responses. Depending on the phase relationship between the lagenar and the utricular response components, the maxima of these 2 components were either phase shifted by 180° (with the result that 2 response maxima per stimulus cycle emerged; see Fig. 6D) or the 2 components were superimposed (with the result that only one maximum per stimulus cycle was present). In the majority of chronic RA frogs (7 out of 10) such a vertical macular response component was present. In the remaining 3 chronic RA frogs the lagenar response component was small, 2 response peaks per stimulus cycle were difficult to detect, but the MAD of the responses were shifted. The superposition of lagenar and utricular response components resulted in a displacement of the calculated "null-point" of the responses (about 60° in Fig. 6E). For the analysis of our data, the utricular response component (dashed curves in Fig. 6, B and E) was fitted separately from the lagenar response component (continuous fit curve in Fig. 6E). Because of the presence of this lagenar response component, the orientation of the abducens MAD vector was no longer coplanar with the ipsilateral horizontal semicircular canal as it is in controls (Fig. 7B). Rather, the abducens MAD vector of chronic RA frogs without utricular reinnervation exhibited an elevation (in 5 out of 10) or a depression (in 5 out of 10) with respect to the plane of the horizontal semicircular canals (Fig. 7D). The elevation or depression component represented the lagenar response component that was evoked by upward or downward vertical linear acceleration, respectively.
Spatial tuning of contralesional abducens nerve responses in chronic RA frogs with utricular reinnervation
As for chronic RA frogs without utricular reinnervation we first determined the direction of the contralesional abducens MAD in the horizontal plane and then analyzed a possible contribution by the lagena with ramplike and vertical accelerations in a population of 10 chronic RA frogs with utricular reinnervation.
Most of the individual MAD directions of this population of RA frogs were caudally oriented in the horizontal plane when compared with the range of MAD directions of controls (Fig. 7E). With some exceptions (RA 09 and 26) were these individual MAD directions less spread out than the MAD directions of chronic RA frogs without utricular reinnervation (compare Fig. 7, C and E). Ramplike or vertical linear acceleration evoked in none of these animals 2 response peaks per stimulus cycle. The calculated contribution of lagenar signals to the macular responses was small (Fig. 7F) compared with the lagenar contributions determined for chronic RA frogs without utricular reinnervation (Fig. 7D). In 2 of the RA frogs with utricular reinnervation (RA 07 and 11) the MAD vectors were very similar to those of control animals.
Origin of macular responses recorded from the contralesional abducens nerve of chronic RA frogs
The contralesional abducens nerve responses evoked by horizontal linear acceleration in chronic RA frogs originated either in the contralesional utricle (because of reorganization; see Fig. 9, C and D), in the ipsilesional utricle (because of reinnervation; see above), or on either side (because of a combination of reorganization and reinnervation). The response component evoked by vertical linear acceleration could have originated from the lagena on either side (see Fig. 9D), given that the lagenar nerves on either side remained intact after the RA nerve section. A few chronic RA frogs (n = 3) were prepared before the abducens nerve recordings for a section of the N.VIII on the side of the earlier RA nerve section to eliminate all afferent nerve inputs from the operated side. Response latencies and sensitivities as well as MADs of macular abducens nerve responses were recorded before and after N.VIII section for a comparison.
|
A relatively long response latency (44 ms) and a low sensitivity (4.0% of response increase per 1% of g) for the frog RA 35 suggested the absence of an RA nerve regeneration (see Table 1). In fact, after N.VIII section were the onset latency (43 ms) and the response sensitivity (3.7% of response increase per 1% of g) practically unchanged, the MAD still had the same orientation in the horizontal plane (57° before and 58° after nerve section; Fig. 8, A and D), and ramplike linear acceleration evoked responses with one or with 2 peaks per stimulus cycle before as after the 2nd nerve section (Fig. 8, B and E). Even though the magnitude of the evoked response was slightly reduced after the nerve section, very similar utricular and lagenar response components were still present (Fig. 8, C and F), demonstrating that these macular responses originated entirely on the intact side, as predicted by the measured onset latencies and response sensitivities.
|
Conversely, the short onset latency (26 ms) and the high sensitivity (19.4% of response increase per 1% g) of macular responses in RA 32 suggested the presence of a reinnervation of the utricle on the operated side. Consistent with this prediction was a much longer onset latency (50 ms) of very small remaining responses after the section of N.VIII on the side of the former RA nerve section. We did not succeed in measuring the sensitivity and the MAD in this frog a 2nd time immediately after VIIIth nerve section because of the weakness of the responses. Similarly, the high sensitivity (19.0% of response increase per 1% g) in RA frog 26 suggested the presence of a contribution from utricular afferent nerve inputs from the operated side in the responses recorded before N.VIII section, even though the onset latency (37 ms) was relatively long and the MAD orientation was with 157° far away from that in controls. After N.VIII section the onset latency was much longer (65 ms), the sensitivity was reduced (12.8% of response increase per 1% of g) and the orientation of the MAD (104°) was altered as well. Obviously, the macular abducens nerve responses recorded in this chronic RA frog consisted of response components that originated in part on the utricle on the operated and in part on the utricle on the intact side.
Macular response parameters recorded from the abducens nerve on the operated side of chronic RA frogs
Ipsilesional abducens nerve responses evoked by linear acceleration were recorded in 8 chronic RA frogs. In 4 of these animals we had recorded from both abducens nerves simultaneously. As shown by sensitivity measurements (see above) ipsilesional abducens nerve responses recorded in chronic RA frogs with and without utricular reinnervation were very similar and did not reflect the 2 groups to which these frogs belonged. Therefore we averaged the results measured for all 8 frogs and compared them with a similar population of control animals. As shown in Table 1, neither the mean onset latencies nor the response sensitivities of both populations were significantly different. The orientation of the MAD in the horizontal plane (54 ± 10°; n = 8) and the elevation of this MAD (3 ± 5.5°; n = 8) overlapped with control values.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
The vector orientation of linear acceleration for the activation of maximal abducens nerve responses in chronic RA frogs differed from the corresponding vector orientations in controls, particularly because of the presence of a vertical response component from the lagena. Such a lagenar contribution was absent in controls, prominent in chronic RA frogs without utricular reinnervation, and barely detectable in chronic RA frogs with utricular reinnervation. Thus the contralesional abducens nerve responded again to linear head acceleration 2 mo after RA nerve section but the spatial tuning of these responses was clearly modified. The responses of the ipsilesional abducens nerve of the same individuals, however, remained unaltered in their spatial characteristics. In essence, the vestibular command signals for ipsilesional and for contralesional abducens internuclear and motoneurons were no longer mirror image–like. As a consequence, compensatory maculo-ocular reflexes of chronic RA frogs were no longer bidirectionally as symmetrical in their spatial tuning as in controls.
Regeneration, reorganization, and reversibility
The regeneration of proximal vestibular and auditory afferent nerve fibers in the brain stem after a section of N.VIII between the ganglion of Scarpa and the brain stem was studied in frogs by Sperry (1945), Zakon (1983), and Newman and Honrubia (1992). Regeneration started to become detectable after about 3–5 wk after the nerve section. Thereafter, minor deviations from control data were reported in the frequency tuning curves of central auditory neurons (Zakon 1983), in the regional distribution of thick and thin vestibular nerve afferent fibers in the vestibular nuclear complex (Newman and Honrubia 1992), or in the strength of recovered vestibular head reflexes (Sperry 1945). Apart from these minor deviations, however, each of these reports emphasized the high degree of specificity of recovery in frequency tuning, reflex behavior, or afferent nerve projection patterns. Some of these early results provided strong support for the theory of chemoaffinity as an important mechanism for the development of precise representational maps (Sperry 1963). The precision of this representation for vestibular sense organs is not expressed in an organotypical map (as for somatosensory, visual, or auditory inputs) but yet in the fact that about 90% of the 2nd-order vestibular neurons receive a monosynaptic input from only one of the 3 semicircular canal nerves in frog (Straka et al. 1997) as in pigeon (Wilson and Felpel 1972) and cat (Kasahara and Uchino 1974) and in the presence of well-organized convergence patterns for afferent canal and macular signals onto 2nd-order vestibular neurons (Straka et al. 2002).
Regeneration of distal vestibular nerve afferent fibers was studied in frogs after RA nerve section by Hernandez et al. (1998). About 4 to 6 wk after RA nerve section the number of axons in the anterior vertical canal nerve branch had recovered to control values again. In our experiments utricular reinnervation might have been delayed by a similar time span, although the distance from the site of the nerve section to the utricular macula (in our experiments) was shorter than that to the crista of the anterior vertical canal (in the experiments of Hernandez et al. 1998). However, the latter authors had reapproximated the sectioned nerve ends to promote regeneration, whereas in our experiments the nerve ends were turned away from each other in an attempt to delay reinnervation. Because of the shorter distance we expected that a reinnervation of the utricular macula occurs earlier after RA nerve section than a reinnervation of the horizontal or anterior vertical canal organs. In both groups of RA frogs (with and without utricular reinnervation) we had also measured the onset latencies and the response sensitivities of the contralesional abducens nerve for horizontal angular acceleration (unpublished data). As a result, we found that the onset latencies and response sensitivities were very similar in both groups of frogs but differed significantly from control values. The latencies were longer and the sensitivities were lower, suggesting the absence of a functional reinnervation of the canal organs. Indeed, similarly long latencies and low sensitivities were reported for the responses of the contralesional abducens nerve from chronic UL frogs (Agosti et al. 1986), in which a reinnervation was excluded after all labyrinthine organs had been removed during surgery.
Head and body posture normalized in UL and in RA frogs with the same time course (Goto et al. 2002). Two months after RA nerve section only small differences were detected between the postural recovery of RA frogs with and RA frogs without utricular reinnervation. Apparently, postural normalization progressed faster than nerve regeneration and macular reinnervation. Therefore we had to develop response criteria that allowed a subdivision between both groups of RA frogs. Significantly longer latencies and lower sensitivities of contralesional abducens nerve responses with respect to control data presented reliable criteria for such a separation. The changes seen in these response criteria after a section of N.VIII on the operated side of chronic RA frogs were fully compatible with this classification. The latter results also demonstrated that the responses to horizontal linear acceleration originated in the contralesional utricle in chronic RA frogs without utricular reinnervation. These utricular input signals were mediated by excitatory commissural fibers to vestibular neurons on the ipsilesional side more effectively in chronic than in acute RA frogs (see Fig. 9, C and D). This increased efficacy of commissural excitation on the operated side was a direct consequence of the synaptic reorganization that was described after RA nerve lesion in earlier reports (Goto et al. 2000, 2001). The changes in the synaptic commissural input organization are cooperative and include an expansion of excitatory and a weakening of inhibitory commissural inputs.
The delay in the onset latencies of contralesional abducens nerve responses of chronic RA frogs without utricular reinnervation reflected more the recruitment of spontaneously inactive central vestibular neurons (about 50%) on either side of the brain stem than the additional conduction time across the brain stem by commissural fibers. This interpretation is consistent with the fact that the onset latencies of macular- and of canalrelated responses strongly decreased in RA frogs as in controls with an increase in the magnitude of acceleration. Hence, the longer latencies indicate that commissural utricular inputs of central vestibular neurons on the ipsilesional side of RA frogs were weaker than afferent utricular inputs evoked by the same linear acceleration in controls. These weaker inputs recruited fewer ipsilesional vestibular neurons than in controls and activated weaker and more delayed contralesional abducens nerve responses than in controls. A similar weakness is also present in dynamic vestibular head reflexes and necessitates the recruitment of a sequence of head catch-up saccades during a displacement in the light (Dieringer 1989). However, the postural reflexes of these frogs are quite normal at that stage, mainly because of parallel spinal synaptic changes with a faster time course (Straka and Dieringer 1995).
Reorganization of synaptic inputs in the ipsilesional vestibular nuclei consisted of an expansion of excitatory signals from intact afferent as well as from commissural fibers (Goto et al. 2000, 2001). Accordingly, in RA frogs the signals from intact posterior vertical canal and lagenar nerves expanded in addition to the excitatory commissural inputs. The utricular response component of the contralesional abducens nerve originated in hair cells located laterally with respect to the striola (see Figs. 8D and 9, C and D), as determined by the orientation of the MAD and the fact that these signals were mediated by excitatory commissural fibers to ipsilesional vestibular nucleus neurons. The vertical macular inputs that contributed the elevation or depression component of the MAD of contralesional abducens nerve responses in chronic RA frogs originated in the lagena (Fig. 9D). These responses could have originated in hair cells located either on the intact or on the operated side and in each of these 2 organs on the one or on the other side of the striola. Because of these various possibilities we could not attribute unambiguously a particular vertical response component of the contralesional abducens nerve to a particular field of hair cells on the ipsi- or contralesional lagena. An expansion of afferent lagenar signals, however, is more likely than an expansion of commissural lagenar signals, given that disfacilitated 2nd-order vestibular neurons prefer afferent nerve inputs from intact fibers over excitatory commissural inputs (Goto et al. 2001).
Reversibility of synaptic reorganization after a reinnervation of the vestibular sense organs in chronic RA frogs is suggested by results from in vitro experiments (Goto et al. 2002). The evoked field potentials recorded in the ipsilesional vestibular nuclei of RA frogs after electric stimulation of the RA nerve on the intact side (commissural input) or of the posterior vertical canal nerve on the operated side (afferent nerve input) increased in amplitude postoperatively over the 1st 2 mo by about 100%, but then declined sharply toward control values over the next 2 mo. Over the latter postoperative period at least reinnervation of the utricular macula took place because UL on the operated side of these RA frogs resulted in postural deficits that increased in magnitude with longer time intervals (Goto et al. 2002). This normalization of the amplitudes of afferent and commissural field potentials suggests a return toward the original organization. In the in vivo experiments described here, the vector orientations of chronic RA frogs with utricular reinnervation still deviated from control vectors but the vertical macular response component was already strongly diminished compared with the response vectors of chronic RA frogs without utricular reinnervation (Fig. 7, Table 2). This reduction in the vertical macular response component may be considered an initial step toward normalization. In essence, the available in vitro as well as in vivo results support the view that the synaptic reorganization after a peripheral nerve section can be reversed with the result that the original map returns. A similar reversibility of the synaptic reorganization after a regeneration of the crushed median nerve was observed in the cortical somatosensory map of monkeys (Wall et al. 1983). Thus reversibility of the reorganization in the case of nerve regeneration is another common feature of the activity-related basic reaction pattern that is activated by nerve injury.
Emergence of inappropriate motor responses
Impressed by the successful spontaneous recovery of a normal posture after UL some of the earlier investigators argued that vestibular compensation represents a goal-directed recovery process that is induced by some recognized "error" in the system and that is directed to its elimination (Flohr et al. 1985). However, the type of motor learning that is used to increase or decrease the gain of the vestibulo-ocular reflex in a context specific manner does not appear to be involved in vestibular compensation. Maioli and Precht (1985), for instance, noted that motor learning was still possible in chronic UL cats to modify the vestibulo-ocular reflex gain but that this capability was apparently not used to improve their poor vestibulo-ocular reflex performance. An alternative view proposes that the activity-related postlesional synaptic reorganization after UL or RA nerve section represents a basic reaction pattern that is concerned more about cellular reactivation than about behavioral recovery (Goto et al. 2002). The beneficial consequences of this cellular reactivation would seem to consist of a reduced asymmetry in the resting activities of bilateral vestibular nucleus neurons and in the improved responsiveness of those cells that were silenced in consequence of the loss of afferent nerve inputs. The undesired consequences of the substitution of missing utricular nerve inputs by utricular commissural and lagenar commissural and/or afferent signals concern the spatial tuning of the output of these cells. The original spatial response tuning became reoriented with the result that the motor commands for contralesional abducens internuclear and motoneurons became spatially inappropriate.
A massive reorganization is known to occur in somatosensory maps of the brain stem and cortex of mammals after the denervation or amputation of a hand or an arm (see Kaas 2000). Undesired consequences of the postlesional reorganization of somatosensory or auditory maps reported by patients include phantom sensations, phantom pain or tinnitus (Flor et al. 1998; Mühlnickel et al. 1998). Thus the results reported here for the vestibular reorganization in the brain stem of frogs are fully compatible with the notion that silencing of an afferent nerve input activates a basic central reaction pattern that is common between sensory modalities and vertebrate species (Goto et al. 2001). From the results of this study we can add the properties "reversibility of the reorganization" and "emergence of undesired responses" to the list of common activity-related postlesional reactions.
Peripheral nerve lesion is a convenient but not a necessary procedure to provoke an activity-related synaptic reorganization. In fact, there is ample evidence from different sensory systems that a change in activity or use after behavioral training, classical conditioning, or prolonged natural sensory stimulation can result in the remodeling of sensory maps (e.g., Elbert et al. 1995; Xerri et al. 1994). Activity-related changes in the synaptic organization of the vestibular nucleus descendens were reported in space-flown neonatal fish (Ibsch et al. 2000). Also, the increase in efficacy of utricular inputs after the plugging of all 3 canals on one side in the monkey can be assumed to result from such an activity-related reorganization (Angelaki et al. 2002).
Spontaneous, unassisted recovery after a partial or complete unilateral loss of labyrinthine function (vestibular compensation) therefore appears to result from changes, some of which are more beneficial (e.g., postural recovery) than others (reflex dynamics; see above). Because the underlying mechanism, however, is activity-related, rehabilitative training can shape the ongoing reorganization. The greater success of recovery for static than for dynamic reflex performance might very well be explained by this basic mechanism. Whereas alerting reafferent signals indicating deficiencies in head–body balance and posture are available permanently during the recovery period, dynamic vestibular reflex signals, however, will be available only during natural activities. It is therefore conceivable that spontaneous recovery of dynamic reflexes is limited and that the improved functional recovery observed, for example, in vestibular patients after specific vestibular exercises (Strupp et al. 1998) is related to a central reorganization that is shaped by visual, vestibular, and proprioceptive inputs during physiotherapy.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: N. Dieringer, Department of Physiology, Pettenkoferstr. 12, 80336 Munich, Germany (E-mail: dieringer{at}phyl.med.uni-muenchen.de).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Angelaki DE, Newlands SD, and Dickman JD. Inactivation of semicircular canals causes adaptive increases in otolith-driven tilt responses. J Neurophysiol 87: 1635-1640, 2002.
Blanks RH and Precht W. Functional characterization of primary vestibular afferents in the frog. Exp Brain Res 25: 369-390, 1976.[ISI][Medline]
Dieringer N. Alterations in neural and behavioral response properties after vestibular lesions. In: Vestibular Compensation: Facts, Theories and Clinical Perspectives, edited by Lacour M, Toupet M, Denise P, and Christen Y. Paris: Elsevier, 1989, p. 83-94.
Dieringer N. "Vestibular compensation": neural plasticity and its relations to functional recovery after labyrinthine lesions in frogs and other vertebrates. Prog Neurobiol 46: 97-129, 1995.[ISI][Medline]
Dieringer N and Precht W. Functional recovery following peripheral vestibular lesions: due to—in spite of—in parallel with— or without synaptic reorganization. In: Adaptive Processes in Visual and Oculomotor Systems, edited by Keller EL and Zee DS. Oxford, UK: Pergamon Press, 1986, p. 383-390.
Elbert T, Pantev C, Wienbruch C, Rockstroh B, and Taub E. Increased cortical representation of the fingers of the left hand in string players. Science 270: 305-307, 1995.
Estes MS, Blanks RH, and Markham CH. Physiologic characteristics of vestibular first-order canal neurons in the cat. I. Response plane determination and resting discharge characteristics. J Neurophysiol 38: 1232-1249, 1975.
Flohr H, Abeln W, and Lüneburg U. Neurotransmitter and neuromodulator systems involved in vestibular compensation. In: Adaptive Mechanisms in Gaze Control, edited by Berthoz A and Melvill Jones G. Amsterdam: Elsevier Science, 1985, vol. 1, p. 269-277.
Flor H, Elbert T, Mühlnickel W, Pantev C, Wienbruch C, and Taub E. Cortical reorganization and phantom phenomena in congenital and traumatic upper-extremity amputees. Exp Brain Res 119: 205-212, 1998.[ISI][Medline]
