• Excitability changes in afferents innervating the urethra, perineum and hindlimb were measured in decerebrated cats during micturition and in response to stimulation of lumbosacral afferents. Increases in excitability were interpreted as primary afferent depolarization (PAD) and decreases as primary afferent hyperpolarization.
• Excitability increases were observed in 11 of 19 urethral pudendal afferents during micturition. Four of these 11 afferents showed an excitability increase during voiding. Seven of these showed a biphasic change with a decrease in excitability when sphincter activity resumed at the end of the void. Three of 19 afferents showed an excitability decrease during micturition and no change was detected in five afferents.
• During micturition, the peak amplitude of urethral afferent-evoked excitatory postsynaptic potentials in seven of eight sphincter motoneurones was diminished to a mean of 36% of control values.
• Eighty per cent of hindlimb cutaneous afferents and 50% of dorsal penile/clitoral and superficial perineal nerve afferents in the sacral cord showed increased excitability during voiding. No excitability increases were measured in 13 hindlimb cutaneous fibres examined in the lumbar segments.
• PAD was observed in sacral urethral, perineal and hindlimb cutaneous afferents in response to electrical stimulation of other perineal, urethral, hindlimb cutaneous and group II muscle afferents.
• It is concluded that control of transmission from urethral afferents by the micturition circuitry is different to that by sensory transmission from hindlimb and perineal regions during micturition. We hypothesize that more than one population of sacral PAD-mediating interneurones is involved.

To determine whether primary afferent depolarization (PAD) could contribute to sphincter reflex modulation during micturition, Angel et al. (1994) examined the excitability of a sample of afferent fibres contained in the common sensory pudendal nerve, the nerve which has strong excitatory actions on the sphincter motoneurones (Bishop et al. 1959; Fedirchuk et al. 1992a). Angel et al. (1994) documented the presence of excitability increases (PAD) in a portion of sensory pudendal afferents but did not examine other hindlimb, perineal and urethral afferent fibres also terminating in the first sacral spinal segment (S1). Based on the fact that the actions of hindlimb and perineal afferents on several populations of sacral motoneurones were suppressed during micturition (Fedirchuk et al. 1994) and the observation that PAD circuits could be activated by the micturition circuitry (Angel et al. 1994), the hypothesis was then proffered that a more widespread PAD of a variety of sacral afferents could occur during micturition. This generalized, decreased sensory transmission would be related to the importance of minimizing inappropriate reflex excitation of sphincter motoneurones during voiding. If restricted to the sacral segments, the coincidental reduction in transmission to hindlimb motoneurones would not probably produce a significant motor perturbation since most of the hindlimb muscle innervation originates with motoneurones located more rostrally within the lumbar cord (Vanderhorst & Holstege, 1997).

When Angel et al. (1994) examined the excitability of sensory pudendal afferent fibres during micturition, the branch of the sensory pudendal nerve innervating the glans or clitoris (dorsal penile or clitoral nerves, respectively) was not separated from the small branch of the sensory pudendal nerve branch that innervates the urethra (Martin et al. 1974). Thus, the sample studied by Angel et al. (1994) could have included either urethral or dorsal penile/clitoral cutaneous afferents. The distinction between the urethral afferents and those in the dorsal penile or clitoral nerves is very important when one considers the differences in their actions on lower urinary tract reflexes. Stimulation of the receptive fields associated with the sensory pudendal nerve, or electrical stimulation of the dorsal penile or clitoral nerve, inhibits bladder reflexes (de Groat & Ryall, 1969; Lindstrom et al. 1983; Morrison, 1987; Kruse et al. 1992) and activates sphincter reflexes in spinal intact animals (Garry et al. 1959; Fedirchuk et al. 1992a). In contrast, distension of and flow through the urethra, or electrical stimulation of the branch of the sensory pudendal nerve innervating the urethra, can facilitate bladder contractions and evoke more complex changes in sphincter activity patterns (Barrington, 1914; Shefchyk & Buss, 1997, 1998; Mazieres et al. 1997). To summarize, the fact that some urethral afferents provide feedback that influences excitatory bladder pathways while others are linked with sphincter control suggested the possibility of a more complex organization and control of urethral afferent actions during continence and micturition.

The purpose of the present study was to test several of the hypotheses arising from our earlier work. In light of the different actions of the urethral and dorsal penile/clitoral branches of the sensory pudendal nerve, an examination of the excitability of afferents contained in the urethral branch of the sensory pudendal nerve separate from the dorsal penile/clitoral branch was undertaken to determine whether there were differences in modulation of transmission from these pathways during micturition. The examination of the patterns of segmental afferent-evoked PAD in urethral, perineal and hindlimb cutaneous afferents using electrical stimulation of segmental afferents was undertaken to determine the pattern of PAD evoked in sacral urethral, perineal and hindlimb afferents and clarify whether a subset of sacral afferents was subject to group II muscle afferent-evoked PAD, as suggested by Angel et al. (1994), or whether group II muscle afferent-evoked PAD was widespread in the sacral cord and found in afferents innervating a variety of target tissues. The determination of the patterns of segmentally evoked PAD in various sacral afferents, as well as excitability changes during micturition, will facilitate the process of identifying the interneurone population(s) involved in modulating sensory feedback during sacral reflex function.

This work has been presented previously in abstract form (Buss & Shefchyk, 1997; Shefchyk & Buss, 1997).

A variety of perineal and hindlimb peripheral nerves were dissected and cut on the left side of the animal. In the present study, four branches of the pudendal nerve were identified: the motor branch to the external urethral sphincter muscle (EUS), the motor branch to the external anal sphincter (EAS) and two parts of the sensory branch of the pudendal nerve. In the cat, the large main sensory component of the pudendal nerve travels in a large fascicle that clearly separates from the motor branches (Martin et al. 1974). From this main sensory portion a small branch innervating the urethra (abbreviated to urSPud) can be separated from the large fascicle commonly referred to as the dorsal penile or clitoral nerve (abbreviated to cutSPud). The dorsal penile/clitoral cutaneous branch of the pudendal nerve (cutSPud) conveys sensory information from the skin of the glans penis (Cooper, 1972) or the labia, vulva, clitoris and skin surrounding the vaginal orifice (Cueva-Rolón et al. 1994). The urethral branch of the sensory pudendal nerve conveys information from flow and perhaps distension-sensitive receptors in the urethral mucosa (Talaat, 1937; Todd, 1964).

In addition, the superficial perineal nerve (SFP), and hindlimb nerves including the caudal cutaneous femoralis (CCF), caudal and lateral cutaneous sural (CCS, LCS), and superficial peroneal (SP) were identified. During the examination of segmentally evoked excitability changes, the two branches of the sural nerve were examined separately since there were some differences in patterns for the two branches. In relation to excitability changes during micturition, no differences were observed and the data from the two branches of the sural nerve were pooled (abbreviated to Sural). The nerves innervating the following hindlimb muscles were also cut: posterior biceps-semitendinosus (PBSt), semimembranosus- anterior biceps (SmAB), triceps surae (GS), tibialis anterior (TA), peroneus longus (PerL), extensor digitorum longus (EDL) and flexor digitorum longus-hallucis longus (FDHL, but also including branches to interosseous and popliteus) and tibialis posterior (Tib). The saphenous (Saph) and quadriceps (Q) nerves on the ventral side of the hindlimb were cut and placed in cuff-style stimulating/recording electrodes. Any remaining nerves branching from the sciatic trunk were cut and ligated. All nerve stimulus strengths were expressed in multiples of the threshold current (T) that produced the first detectable volley at the cord dorsum (digitized at 5 kHz) following single pulses. The proximal portions of the EUS and EAS nerves were either stimulated to antidromically identify motoneurones or placed on bipolar recording electrodes to monitor sphincter efferent activity. Electroneurograms were filtered (high-pass cut-off, 300 Hz; low-pass, 3000 Hz) and amplified 1000-10 000 times. The conduction velocities of the fibres examined in the current study were calculated using the time from the onset of the intra-spinal stimulus to the unit firing in the peripheral nerve and the distance between the spinal recording site and bipolar electrode on the peripheral nerve.

A laminectomy was done to remove the seventh lumbar (L7) to third lumbar (L3) vertebrae and the animal was placed in a Transvertex recording frame. A precollicular-postmamillary decerebration was performed with a spatula. The cerebral cortices, along with all the neural tissue rostral to the brainstem transection, were removed and then the halothane anaesthesia was discontinued. The mean carotid or femoral arterial blood pressure was recorded throughout the experiment and maintained above 80-100 mmHg. Dextran was administered to replace blood loss and a 5 % glucose and bicarbonate solution was infused at 5 ml h⁻¹ throughout the experiment. The animals were given the neuromuscular blocker gallamine triethiodide (5 mg every 45 min) and were artificially ventilated such that end-tidal CO₂ was maintained between 3 and 5 % for the remainder of the experiment. A bilateral pneumothorax was made to minimize thoracic movement artifacts. The nerves were mounted for stimulation or recording on bipolar Ag-AgCl electrodes within a mineral oil pool formed from the skin of the back and hindlimb. Body and pool temperatures were maintained near 37°C using radiant heat. The dura over the exposed spinal cord was cut and patches were made in the pia mater to permit insertion of a microelectrode. At the termination of the experiment the paralysed decerebrate animals were killed by withdrawal of the artificial ventilation and an intravenous injection of a concentrated potassium chloride solution.

Figure 1 Excitability changes in two urethral pudendal afferents during PMC- (A) and distension-evoked (B) voiding

Figure 5 Excitability changes in a superficial perineal afferent (A; conduction velocity, 57 m s−1) and a caudal cutaneous femoralis afferent (B; conduction velocity, 40 m s−1) during distension-evoked micturition

Figure 6 Excitability changes in a hindlimb caudal cutaneous femoralis afferent (A; conduction velocity, 119 m s−1) and a superficial perineal afferent (B; conduction velocity, 57 m s−1) in S1 during micturition evoked by PMC stimulation

Results are given as means ±s.d.

Along with the abrupt biphasic response at the time of the micturition reflex, Fig. 1B shows one of two urethral afferents that displayed a progressive decrease in excitability over several minutes during bladder filling prior to micturition. This same pattern was seen in both afferents during repeated filling and voiding cycles. In both of these afferents, the step-like decreases in the intra-spinal stimulus current seen during the filling period were coincident with progressive reductions in EUS electroneurogram activity. While this pattern of change in excitability and EUS activity was seen in only two animals in the present study, similar progressive decreases in EUS activity (or motoneurone membrane potential) during the later stages of slow bladder filling have been reported before (Fedirchuk & Shefchyk, 1993; see Fig. 1 of Shimoda et al. 1992). Superimposed on the gradual intra-spinal current threshold decrease was sometimes an additional small decrease in threshold at the time of the bladder contraction. In both afferents, the intra-spinal current threshold increased to above control levels immediately after the void as was characteristic of the biphasic response described earlier. In these cases, control levels were measured minutes earlier before the gradual decrease in current threshold manifested itself.

In addition to the biphasic excitability changes observed, four of 19 urethral afferent fibres displayed a monophasic increase in excitability when the EUS activity was suppressed and the bladder contracted (not shown). Three of the 19 afferents displayed a decrease in excitability or PAH during the reflex void and the remaining five afferents showed no measurable change in excitability. The excitability changes in the urethral afferents during micturition are summarized in Table 1.

Table 1 Summary of excitability changes in afferents during micturition

Figure 2 PAD evoked in two urethral afferents (conduction velocity: A, 45 m s−1; B, 42 m s−1) terminating in S1 produced by trains of stimuli at varying strengths (indicated by filled bars above the traces) to cutaneous (A) and hindlimb muscle (more ...)

Urethral afferents showed PAD evoked by hindlimb muscle nerve stimulation at 5T, but not at 2T, presumably reflecting a predominant group II muscle afferent-evoked PAD (Jack, 1978; Edgley & Jankowska, 1987; Jankowska & Riddell, 1994). In a total of 68 tests in 17 urethral afferents in which both 2T and 5T stimulus strengths were tested, 5T but not 2T strengths evoked an excitability increase 97 % of the time. Figure 2B shows the change in intra-spinal stimulation current evoked by electrical stimulation of gastrocnemius muscle afferents in one of the two afferents in which PAD was evoked with < 2T stimulation. It is noteworthy that, in both these cases, it was stimulation of the gastrocnemius afferents that evoked the PAD and that a considerable proportion of group II muscle afferents in the gastrocnemius nerve were activated at strengths < 2T (Jack, 1978). A summary of the incidence of PAD measured in urethral afferents evoked by stimulation of various segmental nerves at 5T is shown in Fig. 3. Many of the nerves produced PAD in at least 50 % of the afferents tested and the mean decrease in intra-spinal stimulation current ranged from 7 to 19 %.

Figure 3 Summary of the incidence of excitability changes evoked in urethral afferents by stimulation of various segmental afferents at 5T

The most commonly observed pattern of modulation of urethral-evoked EPSPs during micturition is shown in Fig. 4. In this EUS motoneurone, during the void and initial period of membrane hyperpolarization (Fig. 4Ab and Bb), the peak amplitude of the urethral-evoked EPSP was diminished to 10 % of its prevoid control amplitude. In the remaining part of the void when the membrane was still hyperpolarized, the EPSP recovered to 25 % of the control value (Fig. 4Ac and Bc). This pattern of modulation of urethral-evoked EPSPs was observed in six sphincter motoneurones that hyperpolarized by 4-10 mV during micturition. On average, in these motoneurones the peak amplitude of the urethral-evoked EPSPs was diminished to 36 ± 19 % of the prevoid value. In one EAS motoneurone which depolarized by several millivolts during the void, the EPSP was reduced to 67 % of the control amplitude. In another EAS motoneurone which did not show a change in membrane potential during voiding, the amplitude of the urethral-evoked EPSP was unchanged. While not measured directly in the current protocol, it is quite likely that there was a conductance change in the motoneurones that hyperpolarized or depolarized during voiding (see Fedirchuk & Shefchyk, 1993; Fedirchuk et al. 1994).

Figure 4 Modulation of urethral pudendal-evoked excitatory postsynaptic potentials recorded in an EUS motoneurone during distension-evoked micturition

While the direction of the excitability changes (e.g. increase) observed during distension-evoked voids (Fig. 5) and PMC-evoked voiding (Fig. 6) were similar, the decrease in the intra-spinal current threshold during PMC-evoked voids was on average 8 % larger than that observed during distension-evoked voiding. In eight afferents it was possible to compare the evoked excitability changes during both distension- and PMC-evoked voiding. In three afferents, the decrease in the intra-spinal stimulus current was greatest during PMC-evoked voiding (mean = 12 ± 5 %) compared with the mean change of 8 ± 2 % measured during distension-evoked voiding. During PMC stimulation there was evidence of PMC stimulus-locked inhibition of the EUS activity but, as evident in Fig. 6A and B, the largest excitability changes occurred during the evoked void and were not sustained throughout the time of brainstem stimulation. Thus the PAD seen during PMC stimulation was not probably due solely to a stimulus-locked activation of descending systems but also reflected the activation of the micturition circuitry for a period of time during the stimulation. The other five afferents did not display an excitability change during either distension- or PMC-evoked voiding.

In sharp contrast to the afferents examined in S1, 13 afferents (10 Sural, 2 CCF and 1 SFP) sampled between mid L5 and rostral L7 failed to display a change in excitability (not shown) during either distension- or PMC-evoked voiding (see Table 1). Only one Sural afferent recorded from L7 displayed a 5 % increase in the intra-spinal current beyond control values (i.e.PAH). All the fibres tested in the lumbar segments received PAD in response to conditioning stimulation of other segmental cutaneous or muscle afferents.

The cutaneous afferents associated with the different hindlimb nerves examined appeared similar to one another in terms of their sources of PAD although some differences may not have been evident due to the small sample size for some afferents. Figure 7 shows typical changes in the intra-spinal current threshold in three sural afferents and a CCF afferent evoked by stimulation of either cutaneous and urethral (Fig. 7A) or muscle nerves (Fig. 7B-D) at various stimulus strengths. The incidence and magnitude of the excitability increases in the two branches of the sural nerve (lateral and caudal sural) and the CCF afferents with conditioning stimulation of various segmental afferents are summarized in Fig. 8A. The data represented in this figure include not only the 28 afferents in which both cutaneous and muscle afferents were tested but also additional afferents in which only cutaneous afferents were tested. The mean decrease in intra-spinal current threshold with conditioning stimulation (5T) of hindlimb cutaneous afferents was 16 ± 8 % (range, 5-32 %), 13 ± 8 % (range, 5-38 %) for perineal afferents (cutSPud and SFP) and 10 ± 7 % (range, 5-26 %) with stimulation of the urethral nerve.

Figure 7 Cutaneous and muscle afferent-evoked PAD in four hindlimb cutaneous afferents terminating in S1

Figure 8 Incidence of PAD and the magnitude of the intra-spinal current threshold changes evoked in various hindlimb cutaneous afferents with stimulation of cutaneous (A) or muscle (B) afferents

With regard to the muscle afferent-evoked excitability changes, in 30 of 31 tests when muscle nerve stimulation was tested at both 5T and 2T, only 5T strengths evoked excitability changes. Conditioning stimulation at 2T (near maximum group I strength) was examined 84 times in 28 afferents with an average of three different muscle nerves tested on each afferent. A decrease in the intra-spinal current threshold of ≥ 5 % was found in only one of 28 fibres (a CCF afferent). However, while this single CCF afferent showed an increase in excitability with 2T PBSt stimulation, strengths < 2T produced no changes.

Figure 7B shows that stimulation of either a flexor (PBSt) or an extensor muscle nerve (Q, GS) at > 2T evoked a decrease in intra-spinal current threshold and that in almost every case more than one muscle nerve was effective in a given afferent. Although only tested in a few fibres, combined simultaneous conditioning stimulation of two effective muscle nerves at 5T did not increase the magnitude of the evoked PAD beyond that observed with conditioning of a single muscle nerve. There was also no evidence of a reduction of the excitability increase with simultaneous stimulation of two nerves using a 20 ms condition-test interval and both threshold and subthreshold conditioning stimulation strengths. In 13 of the 27 afferents that received PAD with 5T (but not with 2T) stimulation strengths, additional increases in the strength of the conditioning stimuli above 5T (e.g. 10T) resulted in a further 4 % decrease in the intra-spinal current threshold (as evident in Fig. 7B with Q stimulation and Fig. 7D with PBSt stimulation).

The PAD evoked with 5T conditioning nerve stimulation in the pooled sample of hindlimb afferents was examined to reveal the muscle nerves most effective in evoking PAD, presumably by group II muscle afferent activation. The histogram in Fig. 8B shows that stimulation of afferents of either flexor (PBSt) or extensor (GS, Q) muscle nerves was effective; about 75 % of all fibres received PAD from PBSt nerve stimulation, about 25 % from GS stimulation, and just over 80 % of the fibres received PAD from 5T Q stimulation (note, however, the small sample size). In addition, the mean decrease in intra-spinal current threshold evoked by muscle afferents at 5T was 9 ± 4 % as compared with that evoked by conditioning stimulation of cutaneous afferents (16 ± 8 %).

While many of the afferents in perineal and hindlimb cutaneous nerves underwent PAD around the time of micturition, it was the patterns of excitability changes in urethral afferents that appeared unique in several aspects. While excitability changes in the afferents may act in many ways to change reflex function, both through facilitation and depression of transmission, we are proposing one simple scenario based on the physiology of the micturition sacral reflex system, which can be tested in future experiments. It is known that urethral afferents facilitate bladder reflexes as well as evoke sphincter reflexes (see Introduction) and, thus, variations in patterns of excitability in subgroups of these afferents might be expected. The portion of urethral afferents that underwent PAH or displayed no change in excitability during voiding may be afferents linked to excitatory bladder pathways. A decrease in transmission from this subgroup of urethral afferents destined to facilitate bladder activation might be counterproductive to the micturition reflex. In other urethral afferents, a biphasic increase (PAD) then decrease (PAH) in excitability was linked to the period during which sphincter activity changed in association with micturition. When the EUS activity decreased as the bladder contracted and the void occurred, a decrease (PAD) in transmission from those urethral afferents that activate excitatory sphincter reflexes would be appropriate. The abrupt inhibition of the PAD, or occurrence of PAH, as the void ended could function to increase transmission from urethral afferents when feedback from the urethra might initiate bursts of EUS activity and facilitate urethral emptying. Unlike the presence of PAD that was loosely associated with the micturition reflex seen in other sacral afferents (perineal and hindlimb cutaneous), the pattern of PAD and PAH suggests the involvement of a selective inhibitory control over the interneurones mediating the PAD of the urethral afferents destined to influence sphincter pathways. This suggestion is supported by the intracellular documentation of the depression of urethral afferent-evoked EPSPs in sphincter motoneurones during micturition. Examination of the modulation of urethra afferent-evoked postsynaptic potentials in a limited number of sphincter motoneurones during micturition revealed a time course of EPSP depression that is similar to the time course of PAD in many urethral afferents; the maximal PAD and EPSP depression occurred at the same time in the void. When the motoneurone membrane potential repolarized to prevoid levels (and often displayed rebound firing) immediately after the void, the EPSPs were no longer diminished in amplitude. In one or two examples, the EPSPs appeared to be potentiated (not shown) or reached spike threshold (Fig. 4A) as the EUS electroneurographic activity returned immediately after the bladder contraction. It was also during this period that PAH was observed in some urethral afferent fibres. However, since the urethral-evoked EPSPs are the result of a population of urethral fibres acting through interneurones interposed in polysynaptic pathways (Shefchyk & Buss, 1997), any modulation of the EPSP size and EUS motoneurone recruitment cannot be attributed solely to excitability changes in the urethral afferents. Membrane conductance increases in motoneurones (see Fedirchuk et al. 1994; present results in the one EAS motoneurone that showed no membrane potential change during micturition), as well as postsynaptic inhibition of excitatory interneurones interposed between the afferents and the various target motoneurones (Fedirchuk et al. 1992a; Shefchyk & Buss, 1997) could also contribute to the diminished likelihood that sensory input would recruit the sphincter motoneurones. Interestingly, these inhibitory mechanisms, whether acting presynaptically or at postsynaptic targets, appear to be similarly controlled by the micturition circuitry.

Previous neuroanatomical studies have described the terminations of afferents in the common sensory pudendal nerve (combined urethral and dorsal penile/clitoral; Thor et al. 1989) or only the clitoral afferents (Kawatani et al. 1994) and thus the sites of terminations of the urethral pudendal afferents, to our knowledge, have not been documented. Such information could shed light upon potential diverging urethral afferent pathways which could be under differential control. A more detailed examination of spinal reflex pathways activated by the urethral afferents has only recently begun (Buss & Shefchyk, 1997; Mazieres et al. 1997; Shefchyk & Buss, 1998).

Based on our results and those of Angel et al. (1994), it is now hypothesized that a common population of sacral interneurones mediate sacral segmental-evoked PAD of hindlimb and perineal cutaneous afferents. The predominant segmental excitatory inputs to the interneurones in this pathway appear to be hindlimb group II muscle afferents and cutaneous afferents associated with the hindlimb, perineum and urethra. The PAD-mediating interneurones probably participate in a system of negative feedback control of afferent transmission (Riddell et al. 1995).

Recently, Jankowska & Riddell (1994, 1995) described a population of PAD-mediating interneurones located in the sacral spinal cord that received strong excitatory segmental afferent input from group II muscle (PBSt, SmAB, GS, FDHL) and cutaneous (perineal and hindlimb) afferents. The pattern of segmental input to the interneurones they examined is very similar to the sources of PAD of the afferents found in the present study (see Fig. 8). While Jankowska & Riddell (1995) focused on the role of these interneurones in modulating transmission from hindlimb group II muscle afferents, based on the present results we propose that the interneurones they described may also mediate PAD of cutaneous and urethral afferents terminating in the sacral segments. Interestingly, R. R. Buss & S. J. Shefchyk (unpublished observations) have observed depression of group II muscle afferent-evoked monosynaptic field potentials in the sacral spinal cord during micturition and a similar depression of both cutaneous and group II muscle afferent transmission has been observed during fictive locomotion (Perreault et al. 1994). We hypothesize that the interneurones described by Jankowska & Riddell (1994, 1995) may be the group of PAD-mediating neurones that are recruited to depress sacral sensory transmission during a variety of motor tasks involving sacral reflex pathways including micturition (see discussion of decreased sensory transmission during movements in Milne et al. 1988; Collins et al. 1998).

Figure 9 Schematic diagram of the proposed circuitry controlling sacral urethral, perineal and hindlimb cutaneous afferent transmission

Recently, Blok et al. (1997) described a group of GABAergic neurones in the sacral dorsal grey commissure which receive a direct projection from the pontine micturition centre. They suggest that these interneurones mediate inhibition of the sphincter motoneurones during micturition. Since PAD within the spinal cord has been shown to be mediated by GABAergic interneurones (for review see Davidoff & Hackman, 1984), the GABAergic neurones described by Blok et al. (1997), which are in proximity to the sacral afferent terminals within the cord, might produce PAD in the sacral afferents as well as mediate inhibition of urethral sphincter motoneurones (Ramirez-Leon & Ulfake, 1993; however, see Shefchyk et al. 1998). Interneurones capable of mediating both PAD and postsynaptic inhibition of motoneurones have been described for spinal hindlimb motor circuits (Solodkin et al. 1984).

PMC stimulation evokes co-ordinated micturition along with EUS motoneurone membrane hyperpolarization (Fig. 1; and Fedirchuk & Shefchyk, 1993) and PAD of sacral afferents (Figs 1, 2 and 6). Both intracellular and extracellular data indicate that there may be PMC stimulus-locked action as well as the effects associated with the evoked micturition. The PMC connection to these GABAergic neurones described by Blok et al. (1997) may be responsible for the rapid onset, stimulus-locked effects sometimes observed in the EUS efferents (Fedirchuk & Shefchyk, 1993; Fedirchuk et al. 1994) as well as activating the micturition circuitry and associated synaptic events during voiding. This possibility awaits further testing during experiments designed to examine the actions of these interneurones during both PMC- and distension-evoked reflex micturition.

As summarized in Fig. 9, we hypothesize that the interneurones mediating segmental-evoked PAD as well as micturition-related depolarization of hindlimb and perineal cutaneous afferents are separate from those mediating the depolarization of urethral afferents during voiding. This proposition is derived from the observations that the interneurones involved in PAD of urethral afferents are inhibited at the time of the return of EUS activity as compared with the extended and variable timing of the PAD observed in other afferents during and immediately after the void. Although simplified by the omission of bladder afferent and parasympathetic pathways and descending pathways, one can see in Fig. 9 the growing complexity of the neural organization involved in the control of a variety of sacral reflex pathways during micturition. Here we consider the multiple sets of spinal interneurones, including at least two populations of interneurones mediating PAD and one mediating postsynaptic inhibition of the sphincter motoneurones that should now be considered in any conceptualization of striated sphincter control. One of the PAD-mediating interneurone populations (Jankowska & Riddell, 1994, 1995) may be less specific in terms of target afferents (i.e. cutaneous and group II muscle) and in terms of the motor tasks that recruit them (i.e. micturition or locomotion). The interneurones mediating urethral PAD are hypothesized to be accessed selectively by the micturition circuitry during voiding and are responsible for modulating transmission from the portion of urethral afferents that activate excitatory sphincter reflexes. The identification and understanding of the control of these various sacral interneurones may provide new avenues for the development of approaches to restore appropriate sphincter activity during both continence and micturition.

The authors with to thank Shannon Deschamps for excellent technical assistance during these experiments and Drs B. Fedirchuk, J. Quevedo and T. Gordon for their helpful comments regarding earlier versions of the manuscript. This work was funded by the Medical Research Council (MRC) of Canada; R. R. B holds an MRC studentship.