| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departamento de Fisiología y Zoología, Facultad de Biología, Universidad de Sevilla, 41012 Sevilla, Spain
Submitted 22 April 2003; accepted in final form 6 June 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
; Delgado-García et al. 1986a,b; Dieringer and Precht 1986; Fuchs and Luschei 1970; Fuchs et al. 1988; Goldstein and Robinson 1986; Pastor et al. 1991; Robinson 1970; Skavenski and Robinson 1973). The abducens nucleus contains the motoneurons supplying the ipsilateral lateral rectus muscle and internuclear neurons that innervate the contralateral medial rectus motoneurons in the oculomotor nucleus (Baker and Highstein 1975; Graybiel and Hartweig 1974; Steiger and Büttner-Ennever 1978). The firing pattern and synaptology is similar in both types of abducens neurons (de la Cruz et al. 1994; Spencer and Sterling 1977). Firing of an abducens neuron commences as the eye reaches a defined position threshold in the orbit while moving toward the ipsilateral side. It has been shown that, beyond the threshold for unit recruitment, neuronal firing rate is proportional to both eye position during spontaneous fixations and eye velocity during saccades. Abducens neurons properties, such as sensitivities and recruitment threshold, are distributed along ranges that ensure fine gradation of force and stable generation of eye position to avoid eye drift and misalignment (Easter 1973).
Work on spinal motoneurons indicates that recruitment order depends on intrinsic properties like size and electrophysiological correlates like current threshold and input resistance (Gustaffson and Pinter 1984, 1985; Henneman et al. 1965). Recruitment also varies with the synaptic input organization and activity patterns of afferent neurons (Burke 1981b; Heckman and Binder 1993; Cope and Sokoloff 1999). When recruitment order remains unchanged under stimulation of different afferent sources, it is argued that the variability in intrinsic properties and the similarity in the synaptic organization could be causal factors (Binder 1989; Burke 1981a,b; Clark et al. 1993; Henneman et al. 1965). Thus several studies on spinal motoneurons have revealed a similar synaptic organization of monosynaptic and reciprocal Ia inputs and recurrent and polysynaptic reflex flexor inhibitions (see in Binder 1989; Burke 1981b). Likewise, variations of synaptic efficacy have been correlated with intrinsic factors such as conduction velocity, input resistance or tetanic tension demonstrating that the synaptic organization of several afferent systems scales according to the size principle (Burke 1968; Burke et al. 1976; Dum and Kennedy 1980; Friedman et al. 1981; Zajac and Faden 1985). By contrast, some reports suggest that recruitment pattern can be altered depending on the activated afferent source (Davies et al. 1993; Desmedt and Godaux 1981). Thus some descending pathways and segmental afferents are neutrally or even inversely organized in relation to the size principle (Kanda et al. 1977; Munson et al. 1982; Powers and Binder 1985; Powers et al. 1993; Westcott et al. 1995). Alterations in the recruitment order occur in stapedius motoneurons depending on sound laterality (Kobler et al. 1987) and in first dorsal interosseus motor units for different movements (Desmedt and Godaux 1981; Jones et al. 1994).
We took advantage of our model of tetanus neurotoxin (TeNT) reversible deafferentation to study the recruitment pattern of abducens neurons under two different forms of reduced innervation (González-Forero et al. 2001, 2002b) for different types of eye movement in the alert cat. We aimed to evaluate the degree to which recruitment properties are dependent on the source of afferent input. We have described that the low dose of TeNT produced disinhibition of firing by blocking a large fraction of inhibitory inputs, whereas the high dose of TeNT produces a reduction of both inhibitory and excitatory synaptic potentials leading to a depressed firing pattern (González-Forero et al. 2002a). Our present data indicate that 1) recruitment order in abducens neurons was similar under different spontaneous and reflexive motor responses, and 2) concurrent changes of neuronal sensitivities and recruitment pattern after TeNT treatment indicate synaptic input organization is a relevant factor for establishing the recruitment threshold and spacing.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Surgical preparation for chronic recordings
All surgical and experimental methods follow closely those described in González-Forero et al. (2002a, 2003). After a protective injection of atropine sulfate to reduce vagal reflexes (0.5 mg/kg, im), the animal was anesthetized with sodium pentobarbital (50 mg/kg, ip) and maintained deeply anesthetized with supplementary doses (3 mg/kg, ip) throughout the 3 h that surgery lasted. The animals were mounted in a stereotaxic frame and implanted intracranially with silver bipolar electrodes to stimulate both VIth nerves and the medial longitudinal fascicle. Eye coils, 22 mm diam, made up of Tefloninsulated multistranded stainless-steel wire (200-µm wire section), were implanted to record eye movements by means of the magnetic field search-coil technique (Fuchs and Robinson 1966). A square window (5 mm side) was drilled in the occipital bone to access with microelectrodes, and a dental acrylic chamber was constructed around the window. To immobilize the head during recording sessions, a restraining system was constructed with acrylic resin cemented to self-tapping screws attached to the skull. Postoperative care was provided daily throughout the experiment. During the initial 3 days, antibiotics (streptomycin and penicillin, 20.000 IU/kg/d, im), corticosteroids (dexametasone, 5 mg/kg/d, im), and analgesics (pirazolone, 0.1 g/kg/d, im) were administered. All the animals used showed no signs of distress in their normal feeding, grooming, and social behavior. Every 2 days, rinsing with sterile saline aseptically cleaned the recording chamber. The brain surface was instilled with drops of antibiotics (gentamicin sulfate) and corticosteroids (dexametasone). The cerebellar surface was protected with a silicone tissue, and the chamber was sealed with aseptic gauze and a cap.
Extracellular recordings
Recording sessions started after 2 wk of recovery, at a rate of 2.5 h per session on alternate days. During the recording session, the animal was gently restrained with elastic bandages in the feline-restraining system. The assembly consisted of a Perspex box cushioned with spongy material and located within the eye movement-recording frame. The system rested on a servocontrolled table driven by a DC-motor in the vertical axis. To induce the vestibulo-ocular reflex, the table oscillated sinusoidally at the frequency of 0.1 Hz and amplitude of ±20°. To induce optokinetic nystagmus, a planetarium projected a rotating random pattern of dots onto a screen in front of the animal. The portion of dura mater accessible from the recording chamber was gently removed under local 2% lidocaine anesthesia. Extracellular recordings were carried out with glass micropipettes beveled to a resistance of 1–3 M
and filled with 2 M NaCl. Motoneurons and internuclear neurons were antidromically activated from the VIth nerve and medial longitudinal fascicle, respectively.
Data storage and analysis
The extracellular neuronal activity was amplified and filtered at a bandwidth of 10 Hz–10 kHz. Action potentials were fed into a window discriminator, and the resulting Schmitt-trigger pulses were stored at a time resolution of 10 µs in a computer using CED 1401 A/D card (Cambridge Electronics Design, Cambridge, UK). Computer programs were developed to display and select neuronal and eye movement data between cursors. Data consisted of epochs of firing occurring during stable eye fixations or during the vestibulo-ocular reflex. Interspike intervals were selected 
200 ms after saccades or fast phases of the nystagmus to avoid the postsaccadic slide in firing rate that follows saccadic bursts. Data for the calculation of saccadic velocity sensitivity were selected during saccadic discharges. The corresponding eye position and eye velocity were also stored to be correlated with the neuronal firing. We recorded the responses of 558 abducens neurons during spontaneous eye movements. Out of them, in the control situation, 127 were motoneurons and 50 were internuclear neurons. During the high- and low-dose treatments, the number of recorded motoneurons and interneurons was 123 and 58 and 132 and 68, respectively. Some of these neurons were also recorded during the optokinetic and vestibulo-ocular reflexes. The discharge parameters were calculated using the equation: FR = F0 + ksP + rsV; where ks is the eye position sensitivity (in spikes/s/°) obtained as the slope of the rate-to-position (FR and P) plot, the coefficient rs is the slope of the rate-to-velocity (V) plot, and F0 represents the firing at straight ahead gaze, thus the ordinate intercept. The position threshold for firing (in degrees) was extrapolated from the rate-position plot as the abscissa intercept and calculated as Th = –F0/ks (Robinson 1970; Skavenski and Robinson 1973). For calculations, we used the left eye (injected eye) for motoneurons and the right eye (noninjected eye) for internuclear neurons. All abducens neurons were recorded in the left nucleus. The firing during the slow phases of the vestibular nystagmus was described as proportional to both eye position (P) and eye velocity (V). The discharge characteristics followed the equation: FR = F0 + kvP + rvV; where kv and rv are obtained by multiple linear regression analysis as the sensitivities to eye position (in spikes/s/°) and velocity (in spikes/s/°/s) obtained as the slope of the rate-to-position plot, and F0 is an independent term that can be used to yield the calculated recruitment threshold during the vestibulo-ocular reflex as Th = –F0/
v.
TeNT injection
Following several control recording sessions, animals were anesthetized with sodium pentobarbital (50 mg/kg, ip) for the injection of TeNT (kindly provided by Dr. J. O. Dolly; Imperial College, London, UK). The left lateral rectus muscle was isolated under a dissecting microscope, and a total of 4 µl of the neurotoxin dilution was injected using a microsyringe. The injections were performed in different trajectories starting from the distal end on the lateral rectus muscle and directed toward the belly to ensure homogeneous spread of the toxin. The TeNT doses employed were 0.5 and 5 ng/kg dissolved in physiological saline. Control recordings were taken previous to TeNT injection. Then, three animals per group were injected.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
During spontaneous eye movements, abducens neurons were recruited into activity at a given eye position threshold. The neuron fired at constant frequencies for eye fixations and either fired a burst for saccades in the on-direction (Fig. 1A,
) or pauses briefly for saccades in the off-direction (Fig. 1A, 
), respectively. The minimum observed stationary firing that abducens neurons could attain was typically between 10 and 15 spikes/s (Fig. 1A, *). Sometimes after a brief burst following an on-directed saccade, the firing decayed abruptly into cut-off (Fig. 1A, 
). Abducens internuclear neurons followed the same pattern of firing (data not shown). Treatment modified the discharge pattern of abducens neurons as short as 2 days after injection of TeNT in the lateral rectus muscle, and the effects lasted for about 3 wk. The data were grouped then as 0–21 days for maximal effects and 22–60 for the recovery. Drastic effects were observed after the high-dose treatment (5 ng/kg). Neurons had a reduced relationship to both fixations and saccades as shown in Fig. 1B (, ) low-dose treatment (0.5 ng/kg) produced an increased firing due to disinhibition (see González-Forero et al. 2002a,b). Thus pauses were not observed during off-directed saccades (Fig. 1C, ).
|
Distribution of recruitment thresholds in abducens neurons
Threshold distribution demonstrated a bell-shaped function in control abducens neurons (Fig. 1D). Mean threshold in control motoneurons was –9.34 ± 9.59° (SD; n = 118), which was significantly different (P < 0.05, Student's t-test) than that of control internuclear neurons (–12.91 ± 7.64°, n = 50). The distribution of thresholds in both types of neurons was skewed (–1.4 and –1.2) to the left (up to –48.6 and –38.0° for motoneurons and interneurons, respectively). Treatment resulted in changes in the distribution of thresholds of both motoneurons and internuclear neurons. A high dose of TeNT produced a further shift in the distribution toward the offdirection (Mn: –24.74 ± 22.96°, n = 123; Int: –36.73 ± 34.66°, n = 57) than control (Fig. 1E; P < 0.05, ANOVA, Tukey test). In this case, the skew to the left was –2.3 and –2.1 and the left tail of the distribution reached more eccentric threshold values in the off motor field (Mn: –144.2 and Int: –190.1°; Fig. 1E). Approximately 30% of the population of recorded interneurons had calculated thresholds beyond eye positions more eccentric than –40°, that is, these units were active at all eye positions. To the contrary, after the low dose, the first neurons to recruit had thresholds of –25.4 and –24.9 for motoneurons and internuclear neurons, respectively. Only the distribution of thresholds of low-dose treated abducens motoneurons was normal (Fig. 1F; P > 0.05, Kolmogorov-Smirnov test). The mean threshold values were different from high-dose treated neurons (Mn: –4.31 ± 5.97°, n = 132; Int: –7.91 ± 7.41°, n = 68; P < 0.05, ANOVA, Tukey test). Pooling the data during the initial 21 days after low-dose treatment demonstrated differences in mean recruitment threshold of motoneurons but not interneurons with respect to control (Mn: –13.14 ± 6.27°, n = 57; Int: –10.08 ± 6.10°, n = 12; P < 0.05, ANOVA, Tukey test). Approximately 1.5–7.3% of the population of abducens neurons had thresholds lower than –20°, whereas in the high-dose treatment, the percentage rose between 44 and 59%. Thus although both doses reduced the mean recruitment threshold, opposed effects were observed on the recruitment rank, since it narrowed under low dose and expanded under high dose relative to control (Fig. 1, D–F).
Relationship of recruitment with static position and saccadic velocity sensitivity
We plotted the regression lines between firing rate and eye position for individual cells. It was noticed that the slope of the rate-position lines (ks) was higher as the cell recruited at more positive eye positions. An example is illustrated in Fig. 2A for 50 control motoneurons randomly selected throughout the recruitment rank, where it can be appreciated how the lines with steeper slopes also had higher abscissa intercepts. Regression lines obtained under the high dose demonstrated lower slopes or sensitivities. On the other hand, the low-dose-treated neurons demonstrated nonlinear rate-position plots (data not shown). When the data were plotted separately for eye fixations occurring after on- or after off-directed saccades, two different regression lines could be fitted. The line for offdirected movements showed low sensitivity and threshold as in the high dose data, while the line for on-directed movements resembled control data. Thus low-dose-treated neurons showed intermediate effects between the high dose and control data. In control neurons, plots of eye position sensitivity during spontaneous eye movements (ks) against the recruitment threshold demonstrated that, although the linear fit produced reasonably good correlation coefficients (r) of 0.76 and 0.71 for motoneurons and internuclear neurons, respectively, the data clearly departed from a linear scatter (Fig. 2B). Thus control motoneurons and internuclear neurons showed significant (P < 0.001) exponential relationships with high correlation coefficients (r = 0.84 and r = 0.82, respectively). Similar data scatters can be seen in previous work (Broussard et al. 1995; Dean 1996; Fuchs et al. 1988; Van Gisbergen and Van Opstal 1989). A relationship was found for the control velocity sensitivity during saccades (rs; P < 0.002), but correlation coefficients were less notable (r = 0.58 for both motoneurons and internuclear neurons; Fig. 2C).
|
We were next interested in testing whether these relationships also held for TeNT-treated abducens neurons. Thus under the high-dose effects, abducens motoneurons demonstrated a significant (P < 0.0001) exponential correlation between ks and recruitment threshold (r = 0.74; Fig. 3A; and ). We noticed that the ks parameter decreased from a control value of 6.25 ± 3.05 spikes/s/° (n = 118; Fig. 3B) to a mean value of 1.47 ± 1.20 spikes/s/° (n = 53) during the initial 21 days of post-treatment (). From 21 to 60 days post-treatment (), sensitivity was not different from control (5.01 ± 2.17 spikes/s/°, n = 70; P < 0.05, Kruskal-Wallis ANOVA, Dunn's method for post hoc comparisons). Exponential relationships between neuronal sensitivity and position threshold were also found significant (P < 0.0001) for both the short term (<21 days, Fig. 3A, ) and long term (21–60 days; Fig. 3A, ) groups of high-dose treated motoneurons as independently considered (r = 0.71 and r = 0.63, respectively).
|
In a similar manner, internuclear neurons treated with the high dose demonstrated a significant (P < 0.0001) exponential correlation (r = 0.73) of ks with recruitment threshold. It is noteworthy to indicate that the effects of the high dose injection of TeNT in the lateral rectus muscle also reached the internuclear neurons of the abducens nucleus (see in González-Forero et al. 2003). Thus we noticed that the ks parameter of internuclear neurons decreased from a control value of 7.0 ± 2.69 spikes/s/° (n = 50; Fig. 3D) to a mean value of 1.47 ± 1.20 spikes/s/° (n = 32) in the early phase (
21 days, ), and later (22–60 days, ) sensitivity returned to control values (6.23 ± 2.85 spikes/s/°, n = 26; P < 0.05, Kruskal-Wallis ANOVA, Dunn's method for post hoc comparisons). Last, a low dose of TeNT did not modify the overall ks sensitivity of motoneurons (5.58 ± 1.79 spikes/s/°, n = 132; Fig. 3B, 
) or internuclear neurons (5.98 ± 1.47 spikes/s/°, n = 68; Fig. 3D, ) compared with their corresponding controls (P > 0.05, Mann-Whitney Rank sum test) and thus were treated accordingly as an homogeneous group. Again an exponential relationship best fitted the relationship between ks and recruitment threshold in low-dose treated motoneurons (r = 0.64) and internuclear neurons (r = 0.55; Fig. 3, A and C; – – –).
These above described results indicated that during the TeNT effects of different doses, or later after recovery, a sensitivity-based recruitment order existed in abducens neurons that was correlated to functional parameters. Moreover, we found, and later in the discussion section it will be shown, that the main relationship between ks and threshold not only scaled as a result of neuronal sensitivity changes but also shifted toward the left compared with control as a consequence of recruitment changes induced by TeNT. Thus disinhibition caused by low-dose treatment produced an initial shift to the left but the high-dose deafferentation further shifted the main relationship to the left.
Were either of these two parameters (ks and threshold) related to biophysical properties of the recorded neuron? To pursue this question we performed linear regression analysis between the ks or the recruitment threshold versus the latency period for antidromic activation in the control sample. Results were only significant (P < 0.005) for control motoneurons. Thus negative correlations were found between ks (r = –0.47) and latency or recruitment threshold (r = –0.41; n = 50) and latency. Neither internuclear neurons nor abducens neurons after treatment showed significant relationships between antidromic latency and either ks or threshold (data not shown). In conclusion, weak linear relationships were found only in control motoneurons, indicating that the higher the threshold or the ks, the shorter the antidromic latency.
Relationship of recruitment with dynamic eye position and velocity sensitivities
Firing of motoneurons and internuclear neurons during vestibular stimulation followed the same principles exposed above (Fig. 4A). Scatterplots of the firing rate versus eye position or eye velocity showed circulation for different directions of movement (Fig. 4, B and D). The partial regression plots, after the subtraction of the complementary component (e.g., the velocity component rv · V for the rate-to-position plot) revealed that scatterplots collapsed into lines whose slopes were the corresponding sensitivities, as illustrated in Fig. 4, C and E. After treatment with TeNT, abducens neurons demonstrated similar changes as those described above for spontaneous eye movements. After the low-dose injection, firing was disinhibited, whereas after the high-dose injection, firing was depressed (data not shown).
|
The plot of motoneuronal kv or rv against the recruitment threshold produced significant (P < 0.001) exponential fits in control motoneurons (Fig. 5, A and C; ——) with high correlation coefficients (r = 0.85, and r = 0.67, respectively). Motoneurons treated with the high dose also showed a significant (P < 0.0001) exponential correlation between kv or rv and recruitment threshold (r = 0.67 for both; Fig. 5, A and C; and ; – – –). Treatment reduced temporarily both sensitivities: the kv parameter decreased from a control value of 6.23 ± 2.90 spikes/s/° (n = 55; Fig. 5B, 
) to a mean value of 2.41 ± 2.24 spikes/s/° (n = 33) during the initial 21 days of post-treatment (). Thereafter and 60 days of post-treatment (), sensitivity was not different from controls (5.83 ± 3.34 spikes/s/°, n = 37; P < 0.05, Kruskal-Wallis ANOVA, Dunn's method for post hoc comparisons). The rv parameter decreased from a control value of 1.2 ± 0.46 spikes/s/°/s (n = 55; Fig. 5D, ) to a mean value of 0.62 ± 0.49 spikes/s/°/s (n = 33; 21 days, ), and sensitivity resumed to normal (1.21 ± 0.51 spikes/s/°/s, n = 37; , P < 0.05, Kruskal-Wallis ANOVA, Dunn's method for post hoc comparisons). The low dose of TeNT did not modify the kv or rv sensitivities of motoneurons throughout the experimental course (kv: 6.73 ± 2.88 spikes/s/°, rv: 1.46 ± 0.62 spikes/s/°/s; n = 64; Fig. 5, B and D, ; P > 0.05, Student's t-test). An exponential relationship (P < 0.005) best fitted the relationship between kv and recruitment threshold in motoneurons (r = 0.43; Fig. 5A, ), but no significant relationships could be obtained from the scatter of rv under the low dose (Fig. 5C, ). Thus recruitment threshold during the vestibulo-ocular reflex was exponentially related to sensitivity in a similar fashion to the pattern observed during spontaneous eye movements.
|
Additionally, a subset of 24 control and 19 high-dose motoneurons were also recorded during ramp optokinetic stimulation that allowed calculating an eye position sensitivity (ko) as the eye moved in slow phases of the nystagmus at constant velocity. The ko was not different from the corresponding ks or kv (P > 0.05, ANOVA) both in control (ko = 5.3 ± 1.8 spikes/s/°) and after high-dose treatment (ko = 3.93 ± 2.56 spikes/s/°). Multiple regression of the three k sensitivities yielded significant (P < 0.01) correlations (r = 0.87 and 0.77 for control and high-dose treatment, respectively).
Effect of different modality of eye movements on recruitment
Do abducens neurons behave in similar ways while performing different types of eye movement? To solve this question, we correlated the data obtained under spontaneous eye movements with those resulting from vestibular stimulation. If sensitivities were similar it would indicate that the combination of afferents used by one cell to produce one or another type of eye movement was similar or at least the combined weight of different inputs was similar. Under these premises, it could also be expected that the recruitment order under different modalities would be similar. Thus good positive correlations were obtained by plotting kv against ks, indicating that the abducens neurons tend to fire with similar sensitivity irrespective of the inducing source of eye movements. Figure 6A illustrates for abducens motoneurons during control (r = 0.82; ) and high-dose treatment (r = 0.87; 
) that the ks and kv sensitivities were correlated. Good relationships were also found for interneurons (r = 0.75 and 0.82, respectively; Table 1).
|
|
To determine the recruitment under different modalities, we correlated the threshold value obtained under both spontaneous movements and vestibular stimulation. Correlation coefficients for motoneurons in control, high dose, and low dose were 0.79, 0.81, and 0.69, respectively. Results for internuclear neurons were r = 0.74, 0.75, and 0.83 for control, high, and low doses, respectively. In control (n = 18) and high-dose (n = 12) treated motoneurons, it was possible to obtain a multiple regression analysis of thresholds (P < 0.005) calculated from control, vestibular, and optokinetic data. In both cases, correlation of coefficient was 0.96. These data indicated that, at the populational level, abducens motoneurons and internuclear neurons recruited in a similar manner under two different eye movement modalities. However, recruitment threshold was approximately 10° more negative during the high dose for both motoneurons and internuclear neurons (P < 0.05, ANOVA, Dunn's method). Correlation between the two vestibular sensitivities (kv vs. rv) was good in control motoneurons (r = 0.72), indicating that motoneuronal sensitivity for two different signals tended to vary in a similar way (Fig. 6B, ). In high-dose-treated motoneurons correlation was good (r = 0.79; Fig. 6B, ). However, motoneurons under the low dose had smaller correlation coefficients (<0.48) for all regressions performed (Table 1). Finally attempts to correlate the saccadic sensitivity (rs) with other sensitivities produced low correlation coefficients between 0.38 and 0.7 (Table 1).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
TeNT-induced changes in the firing pattern and discharge characteristics
Changes in firing pattern and discharge characteristics could be explained by the progressive deafferentation produced by increasing doses of TeNT according to our previous findings showing that the low dose decreases inhibitory synaptic transmission, whereas the high dose affects both inhibition and excitation (González-Forero et al. 2002b). As illustrated in Figs. 2A and 7A for control motoneurons, recruitment range was wide for low-sensitivity units and narrow for high-sensitivity neurons, which indicates that small eye position variations recruit more units in the high-threshold region than in low-threshold positions (Fig. 1D; i.e., the more tension that is needed, the finer the recruitment gradation). Thus while firing rate modulation appears as a fundamental mechanism to account for low-force generation, both firing rate modulation and unit recruitment would serve at an intermediate force output. Nevertheless, since recruitment of abducens units occurs mostly in the nasal hemifield, firing rate modulation would be the only factor responsible for force gradation in the temporal hemifield. Related schemes of force gradation have been observed and modeled (Fuglevand et al. 1993; Heckman and Binder 1991a; Milner-Brown et al. 1973). It has been suggested that in small hand muscles, force gradation is achieved by rate coding, whereas large muscles operate under a broad recruitment scheme (Kukulka and Clamann 1981). The extraocular muscles, which generate a precise gradation of force using fast muscle fibers, would fit within the first class since most units are recruited by mid-motor range (Goldberg et al. 1976; Grantyn and Grantyn 1978).
|
We describe a homogeneous response in recruitment properties of abducens neurons to TeNT applications, whereas on the basis of muscle histochemistry (Spencer and Porter 1981, 1988) and retrograde tracing (Büttner-Ennever et al. 2001) at least two types of motoneurons—twitch and nontwitch— could exist. More recently, transneuronal rabies virus tracing, in primates, suggests differences in the afferent inputs to those two likely types of motoneurons (Büttner-Ennever et al. 2002). There are several reasons that led us to suspect that these reported nontwitch motoneurons, postulated to have a tonic firing, are not, at least to a large extent, present in our sample. First the tonic firing present after TeNT treatment is due to deafferentation since it occurred progressively through a time course and it also happened in both motoneurons and internuclear neurons. Second, the nontwitch population of motoneurons lies marginal to the abducens nucleus, which was not a preferential site of recording, and finally, all our recorded abducens neurons showed a phasic signal proportional to eye velocity.
Does progressive synaptic elimination produces changes in the pattern of recruitment? The answer is illustrated in Fig. 7A, where the exponential fits that adjusted the data are redrawn for motoneurons in control and after the two doses of TeNT. It can be observed that following TeNT, in both cases, the treated curves show reduced ranges of ks and progressive displacements toward the left (the off-motor field) as the dose increases. It is concluded, then, that the exponential pattern of recruitment applies to all populations studied here, indicating that recruitment is not linear through the motor range. Recruitment thresholds become more negative as the dose increased presumably due to the increasing deafferentation, which in turn could involve greater input resistance and effective synaptic currents from TeNT-resistant afferent inputs. Therefore it can be argued that neuronal sensitivity and threshold are linked to the number or density of afferents. Thus the low dose showed intermediate characteristics between the control and the high dose, whereas the high-dose treated group showed reduced sensitivity and threshold. This result indicates a link between recruitment threshold and sensitivity, and at least in our data, they seem mainly affected by the loss of synaptic inputs. Our findings are in agreement with Powers and Rymer (1988) who found a reduction in the firing modulation and recruitment force of medial gastrocnemius motoneurons in response to alterations of the synaptic drive by dorsal hemisection.
Intrinsic correlates to recruitment order
In spinal motoneurons, it has been shown that differences in excitability correlated to the motoneuronal size are responsible for the orderly recruitment (Henneman et al. 1965). The so-called size principle has been widely discussed in both physiological and morphological grounds (Burke 1981a,b; Ulfhake and Kellerth 1982). Input conductance and axonal conduction velocity have been positively correlated to cell size, indicating that smaller neurons will recruit before than thick-axoned ones (Kernell and Zwaagstra 1981). Similarly, in cat abducens motoneurons, the input resistance has been inversely correlated with the axonal conduction velocity (Grantyn and Grantyn 1978), indicating again a relationship between excitability and size. Our results, like those of others, indicate that although the correlation between ks and threshold is good, the expected negative relationship of these two parameters to conduction velocity are weak in control and even nonsignificant in the experimental groups. This could indicate that synaptic inputs responsible for eye position signals are distributed among control abducens motoneurons in relation to the size principle, but following the TeNT blockade this organization is lost, remaining only a sensitivity-based scheme for recruitment. However, since afferent synaptic activity significantly affects input resistance and other integrative properties (Destexhe and Pare 1999; Korogod et al. 2000), it could be expected at least certain co-variation degree between extrinsic and intrinsic determinants. Previous works in monkey and cat showed similar moderate relationships in abducens motoneurons (Delgado-García et al. 1986a; Fuchs et al. 1988) but not in the goldfish extraocular motoneurons (Pastor et al. 1991). However, some differences existed between the studies in cat and monkey in regard to abducens internuclear neurons. Both studies support a relationship between ks and threshold, like our present results, but differed in their relationships to latency. Our data, for a large sample of feline interneurons (n = 41), indicate, as in the study by Fuchs et al. (1988), that there is no good relationship between threshold and latency despite a relationship between ks (or other sensitivities) and threshold. It is interesting to note that, although conduction velocity has been found to be a good predictor for the recruitment order sequence in heterogeneous and homogeneous slow muscles (Bawa et al. 1984), it fails to predict the recruitment order in the fast-twitch (type F) motor-units of mixed muscles (Tansey and Botterman 1996; Zajac and Faden 1985; but see Cope and Clark 1991). Furthermore, in our study, the short length of cranial nerves would also contribute to generate prediction errors in the recruitment sequence. In summary, we conclude that a recruitment order based solely in size-related factors does not produce a rule to rank motoneurons as a function of neuronal sensitivity. The use of TeNT does not modify this view since it has been shown that TeNT does not alter the excitability of either spinal (Kanda and Takano 1983; Wiegand and Wellhöner 1979), cranial motoneurons in vivo (González-Forero et al. 2002a), or cultured neurons (Dimpfel 1979).
Synaptic organization as factor that influences recruitment
According to the size principle, the governing rule for rank ordering is input resistance. If synaptic weights are similar, those cells having more input resistance would recruit earlier (Gustafsson and Pinter 1984), but the synaptic currents are not necessarily the same in all abducens neurons. In the study of Broussard et al. (1995), it was found that excitatory and inhibitory responses after vestibular electrical stimulation were positively correlated with the k and r sensitivities. Moreover, intracellular studies reveal different amplitude of synaptic potentials across abducens neurons (Baker and Highstein 1975; de la Cruz et al. 1994; Pastor et al. 1997). In computational grounds, it has been predicted that a varying synaptic weight of premotor neurons (abducens internuclear neurons) simulates an adequate recruitment of medial rectus motoneurons (Dean 1997). Our results indicate that synaptic input plays a key role in determining a pattern of recruitment since the high dose led to a recruitment pattern that shifts and expands in range with respect to control. TeNT not only blocks synaptic inhibition, but also excitation, as shown both in vivo (Curtis et al. 1976; Kanda and Takano 1983) and in vitro (Gobbi et al. 1993). TeNT produced a dose-dependent reduction in the density of synaptophysin-immunoreactive terminals in the abducens nucleus (González-Forero et al. 2002b). The main effect of TeNT in the spinal cord is the blockade of inhibitory inputs to the motoneuron, which produces disinhibition and tetanic paralysis (Brooks et al. 1957; Curtis and De Groat 1968). Thus we propose that the more deafferented the cell, the more regular its firing (González-Forero et al. 2002a), and also, the earlier it will recruit in conjunction with lower sensitivities along the exponential relationship that links sensitivities and threshold to the synaptic inputs received.
It has been argued that variations in the relative distribution of synaptic input to low- and high-threshold neurons will serve to alter the ultimate magnitude of threshold differences in a pool, that is, the recruitment gain (Kernell and Hultborn 1990). Therefore excitatory systems organized like the Ia monosynaptic input on spinal motoneurons, which produces greater effective synaptic currents in low-threshold motoneurons, would expand the intrinsically determined rank and could aid to preserve the recruitment sequence (Heckman and Binder 1993). On the contrary, excitatory projections opposedly organized like rubro-, vestibulo-, or corticospinal pathways may not only increase the recruitment gain but also disrupt the recruitment sequence (Binder et al. 1998; Heckman and Binder 1993; Wescott et al. 1995). Although the effects of the gradient of inhibition on the recruitment rank are not so clear in spinal motoneuron pools, (Heckman and Binder 1991b, 1993), we would expect that recruitment rank was compressed or expanded when afferent inhibition was organized in a similar or in an opposite scheme to the Ia monosynaptic input, respectively. Our results are consistent with this suggestion, since recruitment gain was altered in opposing directions under different forms of selective deafferentation. Since excitatory and inhibitory afferences to the abducens nucleus are organized in a way that induces greater responses in higher-sensitivity neurons (higher threshold neurons; Broussard et al. 1995), selective disinhibition would decrease the recruitment gain and additional excitatory deafferentation would expand the rank as found here.
A second factor controlling recruitment order could also be the discharge characteristics of premotor neurons. For instance, Iwamoto et al. (1990) described a relationship between eye position sensitivity and recruitment threshold in eye movement-related second order vestibular neurons. Our results show that abducens internuclear neurons (a premotor neuron) also show recruitment order that could influence the recruitment properties of medial rectus motoneurons, as postulated by Dean (1997).
Eye movement modality and recruitment order
Is neuronal threshold similar for different eye movements? Our results support that view for two types of eye movement: spontaneous and reflexively induced, either vestibular or optokinetic. However, the recruitment reversal found in the motor units of the interosseus muscle when executing movements of different direction but not of different speed should be noted. This finding indicated that motor commands are patterned in terms of movement directions rather than muscles themselves (Desmedt and Godaux 1981). Alterations in the recruitment sequence might be due to synaptic actions that do not follow the size principle scheme, for instance, the cutaneous input (Dum and Kennedy 1980; Kanda et al. 1977; see, however Clark et al. 1993; Riek and Bawa 1992). Likewise, other descending and segmental inputs do not scale according to the size-principle in cat spinal motoneurons (Dum and Kennedy 1980; Kanda et al. 1977; Powers and Binder 1985; Powers et al. 1993; Wescott et al. 1995). It has been argued that the inputs from polysynaptic descending and segmental inputs would be mediated through common interneurons whose synaptic distribution on spinal motoneurons would be different to that of the Ia input (Binder 1989; Heckman and Binder 1993; Powers and Binder 1985). Variations in the recruitment order in human soleus motoneuron pool have been also observed under voluntary or reflex activation (Davies et al. 1993) as well as in toe extensor motor units during phasic or tonic voluntary contractions (Grimby and Hannerz 1977). According to this, the possibility for reversal of the recruitment sequence in the oculomotor system could be found in saccadic versus sustained firings. Nevertheless, our data indicate that rs values were also exponentially related with eye position threshold. It is not surprising that abducens neurons have invariant threshold since 1) the lateral and medial rectus muscles have a single movement direction by contrast to vertical and oblique muscles (Simpson and Graf 1985), and 2) the signals conveyed to abducens nucleus for different eye movements arise from the same set of premotor structures (Baker et al. 1981). As suggested by Grantyn and Grantyn (1978), this organization would differ from that seen in spinal motoneurons having a more diverse number, origin and distribution of inputs, as well as more complex integrative tasks.
TeNT effects on the total output of abducens neurons
We looked at the consequences on the motor output of the synaptic deafferentation produced by TeNT on abducens neurons. Our data plotting ks versus the recruitment threshold produced better correlation coefficients for exponential than for linear relationships. This relationship, first described as linear by Goldstein and Robinson (1986), can be observed in other work (Broussard et al. 1995; Fuchs et al. 1988; Van Gisbergen and Van Opstal 1989), in Fig. 1 of Van Gisbergen and Van Opstal (1989) compiling the findings of four studies, and in models (Dean 1996). Although still exponential, the relationship shifted along the ordinate and abscissa axes after TeNT. Our data demonstrate that an exponential function describes the ks to threshold relationship both for control and partially deafferented neurons.
The total active force produced by the eye muscles is a transformation of the total firing for any given eye position into mechanical force (Dean 1996). To determine how the pooled data generate the motor output that ultimately is translated into force, we calculated the total motor output received by the lateral rectus muscle or the medial rectus motoneuronal pool under the three different conditions studied here. Motor output was calculated as the total number of spikes produced by our sampled population in steps of 1° as the sum of the firing rate expected for each neuron according to its actual relationship between FR and eye position. Figure 7B compares the normalized total innervation of motoneurons under different treatments. In all cases, it can be seen that the shape of our innervation function agrees well with that experimentally determined by electromyography or by the length-tension curve during eye fixations in humans (Collins 1975). It was noticed that, under both doses of TeNT, the motor output was initially larger than control, but around –6°, the control output surpassed the force generated by the high-dose treated pool (depressed), and finally at around +10°, it surpassed the force generated by the low-dose-treated group (disinhibited). As suggested in this and previous works, maximal force gradation is mainly achieved by firing rate modulation. Thus it would be expected that TeNT induce 1) motor output increases at low and intermediate force levels (off-hemifield) due principally to recruitment threshold reductions, and 2) diminished motor output at high force levels (on-hemifield) due exclusively to low firing rate modulation after high-dose TeNT and to the reduction in the number of high sensitivity neurons in the low-dose-treated group (Fig. 7A). These suggestions were confirmed in the motor predictions calculated in Fig. 7B. Thus it is shown that our two deafferented situations produce less force in the on-motor hemifield than the control. If motor output were translated directly to force and treated motor output was compared against the hypothetical antagonistic control output (with an opposite direction of activation), it would explain the reported eye deviation toward the on field under the low dose and toward the off-field under the high dose (González-Forero et al. 2002a).
A final consequence arises from comparisons between the data of motoneurons and internuclear neurons. We calculated the total output of internuclear neurons and plotted together with that of abducens motoneurons. The summed firing for our population of internuclear neurons resembled more that calculated by Dean (1997) for the sampled population of internuclear neurons in monkeys recorded by Gamlin et al. (1989) than those recorded by Fuchs et al. (1988). We found for control (plotted in Fig. 7C) and treated data that internuclear neurons generated a higher output than abducens motoneurons for any given eye position. This result is due to lower recruitment thresholds and higher sensitivities found in internuclear neurons (Dean 1997; Delgado-García et al. 1986b), which might rescale after the extra synapse with the medial rectus motoneurons in the route of signals conveyed by the abducens nucleus toward the yoked pair of horizontal muscles.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: Á. M. Pastor, Departamento de Fisiología y Zoología, Facultad de Biología, Avda. Reina Mercedes 6, 41012 Sevilla, Spain (E-mail: ampastor{at}us.es).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION |
|---|
