Brain (1990), 113, 813-820
THE BULBOCAVERNOSUS REFLEX
A SINGLE MOTOR NEURON STUDY
by DAVID B. VODUSEK and
MARTIN JANKO
(From the University Institute of Clinical Neurophysiology, University Medical Centre, Ljubljana,
Yugoslavia)
SUMMARY
INTRODUCTION
The contraction of bulbocavernosus muscle or anal sphincter on penile squeeze (the
bulbocavernosus reflex; Bors and French, 1952; Lapides and Bobbitt, 1956) indicates
integrity of the reflex arc containing the sensory and motor fibres of the pudendal nerve
and spinal segments SI - S 4 . Its electrophysiological correlate — using electrical impulses
for stimulation and EMG techniques for recording the response — was described by
Rushworth (1967) and later introduced into routine use in patients with suspected sacral
neurogenic lesions (Ertekin and Reel, 1976). The bulbocavernosus reflex is usually
included in textbooks under the 'exteroceptive' or cutaneomuscular reflexes and believed
to be polysynaptic, a view held also by workers who have studied its behaviour
more closely (Krane and Siroky, 1980; Varma et al., 1986). In our previous
electrophysiological studies we observed that the bulbocavernosus reflex shows little
habituation and therefore postulated that it might have an oligosynaptic central integration
(VoduSek et al., 1983). The single fibre EMG (SFEMG) technique allows selective
recording of single muscle fibre potentials, representing activity within single motor
units and therefore discharges from single motor neurons. Such recordings make
measurements of reflex latencies at the level of single motor unit responses possible.
From repeated latency measurements (i.e., of consecutive reflex discharges) the latency
variability of single motor unit reflex responses can be calculated and expressed, for
Correspondence to: Dr David B. VoduSek, Institute of Clinical Neurophysiology, University Medical Centre, ZaloJka
cesta 7, 61105 Ljubljana, Yugoslavia.
© Oxford University Press 1990
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
Reflex latency variability was established for single motor neuron discharges in the bulbocavernosus reflex,
as elicited by electrical stimuli to the dorsal penile nerve and recorded by a single fibre EMG electrode
in the bulbocavernosus muscle. Whereas many reflex responses had a rather large latency variability of
above 1000 /is (expressed as SD of mean latency) there was a group of motor neurons with a variability
of around 500 /is. Single motor neuron reflex responses with shorter latencies tended to show less variability. No habituation of single motor neuron reflex discharges was observed on prolonged regular repetitive
stimulation. Both absence of habituation and the relatively low latency variability of bulbocavernosus reflex
responses for single motor neurons suggest similarities between this reflex and the first component of the
blink reflex; we postulate that the shortest bulbocavernosus reflex pathway is oligosynaptic.
814
D. B. VODUSEK AND M. JANKO
instance, as the SD of the mean latency value. This latency variability has been correlated
with the complexity of the reflex arcs of the H reflex (Trontelj, 1973), blink reflex
(Trontelj and Trontelj, 1973), and the flexor reflex (Janko and Trontelj, 1983). The
aim of this work was to study reflex responses of single bulbocavernosus motoneurons
to electrical penile stimulation and to establish their latency variability, thus making
comparisons with other better known reflexes possible. A preliminary report has already
been presented (Janko and VoduSek, 1983).
SUBJECTS AND M E T H O D S
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
Eleven male volunteers, recruited from patients with premature ejaculation, all of whom gave informed
consent, participated in the study. They were aged 19—35 yrs, without evidence of neurological disease,
and were not diabetic and did not take drugs. In the first 6 subjects a Medelec MS6 EMG machine was
used for stimulation and recording; the measurements of reflex latency variability were mostly made offline from tape. The recorder HP 3960 had — at high speed (15 ips) — a jitter that did not exceed 1 —2 /is
over segments of 10 ms. The measuring unit for latency variability consisted of a custom built 'jittermeter'
(Trontelj et al., 1979) and a computer (HP 2100S). In the rest of the study a Medelec-Vickers Mystro
EMG system was used for stimulation, recording and analysis. The high-pass filters were set at 500 Hz
and the low-pass filters to 16 and 20 kHz, respectively, on the MS6 or Mystro machine. The bulbocavernosus
reflex was elicited electrically with ring electrodes (Disa 13L69) mounted on the penis. Direct responses
were elicited by electrical stimulation of the pudendal nerve (branch) in the perineum (VoduSek et al.,
1983) with a bipolar surface electrode (Medelec TP NSP53054). Both reflex and direct responses were
elicited with square wave pulses of 0.2 ms duration at a regular rate (0.5 - 2 Hz for the reflex and 2 - 5 Hz
for the direct responses). Recordings were made with SFEMG needle electrodes (Medelec 53032) from
single sites in bulbocavernosus muscles.
Muscle fibre potentials belonging to single motor units were sought, uncontaminated by firing of potentials
belonging to other motor units. Repetitive reflex responses of particular single motor units were recognized
as such by their constancy of shape. They were either single muscle fibre potentials (recognized according
to established criteria; see Stilberg and Trontelj, 1979) or — even easier for identification — two or more
muscle fibre potentials constituting a complex potential with time-locked spikes appearing according to
the all-or-none principle with a small jitter between individual fibres (see Trontelj and Trontelj, 1973).
Selective and reliable recording of such responses was made easier by accepting only recordings without
interference from neighbouring motor units. The latter nearly always appeared at a different threshold,
which made 'clean' recordings easier. For reflex latency variability measurements, a series of consecutive
responses was sampled. As a rule two series of 50 responses were obtained for a particular motor unit.
Recordings from the same bulbocavernosus muscle were accepted as belonging to a different motor unit
if elicited at a different threshold value at a different needle electrode position. Up to 9 recordings from
different motor units were obtained in a single subject, counting sites in both bulbocavernosus muscles.
For both reflex and direct response analysis, latencies were measured automatically to a set point, adjusted
manually on the steep slope of the spike potential. The accuracy of measurement for the two systems is
stated as 1 /is for the jittermeter and 1/1000 of the time-base used for the Mystro. Mean latencies with
SDs and ranges were calculated. As a rule two measurements were obtained. The SD of mean latency
was taken as a measure of variability. If its values in repeated measurements differed by more than about
20% they were repeated; only recordings with a 'repeatable variability' were accepted. Recordings with
the lowest latency variability were then included and are reported. Measurements of latency variability
were also made for consecutive direct responses (muscle fibre potentials) from bulbocavernosus motor
units. Direct responses were obtained with strong stimulation (see Trontelj et al., 1986) of pudendal nerve
(branches). Two series of 50 consecutive responses were sampled for each individual potential and the
lower value is reported.
BULBOCAVERNOSUS REFLEX
815
RESULTS
25 ms
FKJ. 1. Bulbocavernosus reflex response of a single motor neuron, as characterized by
an individual muscle fibre action potential. Consecutive responses from above downwards.
Stimulus intensity is increased at upper arrow and decreased again to the previous level at
lower arrow. A discrete latency change can be seen on changing the stimulus intensity.
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
On increasing the strength of penile stimulation, responses from single motor units
(see Methods) could easily be isolated with proper SFEMG needle adjustments.
'Threshold' stimulation was defined arbitrarily as the strength that elicited a response
on about 90% of repetitive stimuli at a particular frequency of stimulation. Usually the
responses from motor units emerged rather suddenly on threshold stimulation and then
became regular (stimulus:response ratio 1) with very slight increases in strength. On
further increases in strength, responses from other motor units often also appeared.
On suprathreshold stimulation, individual responses could easily be followed for longer
periods of time. We estimate that in some instances about 1000 consecutive repetitions
of the same motor unit reflex response were obtained without any drop-outs.
The mean latencies of reflex responses of various motor units chosen for measurement
ranged from about 29 to 45 ms. In 2 subjects motor unit reflex discharges could also
be obtained and measured at a later latency interval between about 50 and 75 ms. The
latencies of consecutive reflex responses showed some gross and some subtle latency
changes. The gross changes consisted of a bimodal or polymodal distribution of response
latencies; the response 'jumped' between these discrete latency values on consecutive
discharges. Occasionally this was closely correlated with changes in the strength of
stimulus applied, shorter latencies being obtained with stronger stimuli (fig. 1). In addition
to this phenomenon a slight variability of latency could also be seen in case of responses
816
D. B. VODUSEK AND M. JANKO
with a 'constant' latency. If in these cases latencies of consecutive reflex responses were
plotted, they occasionally showed a somewhatfluctuatingpattern of changes; nevertheless
their frequency histograms showed a rather normal distribution.
With regard to reflex latencies and their variabilities the following observations were
made. (1) Latencies could shorten by 'jumps' on increasing the stimulus strength (fig.l).
(2) A slight but significant decrease in latency was also observed on increasing stimulus
strength in responses otherwise appearing at a 'constant' latency value (measurements
made in 7 motor unit responses showed a mean shortening of the latency for 715 /is,
as well as a mean decrease in reflex jitter for 317 /is; see fig. 2). (3) Some motor units
occasionally discharged more than once in a particular reflex response; both a polymodal
FIG. 2. Bulbocavemosus reflex response of a single motor neuron on
threshold (above) and slightly suprathreshold (below) electrical stimulation.
Consecutive responses always from above downwards. At threshold
stimulation occasional failure of the stimulus to elicit a reflex response can
be seen, as well as a greater variability of reflex latencies.
-
10 ms
latency distribution as well as multiple discharges were occasionally observed for the
same motor unit (fig. 3). (4) Response latencies from 4 motor units were measured
separately at both an 'early' and 'late' discrete latency value which were 3—5 ms apart.
Reflex responses showed less reflex latency variability at the 'early' latency value (444
± 44 us vs 1012 ± 138 /ts). (5) Putting all measured minimal reflex latency variabilities
of individual motor units together, a clear grouping could be seen. Motor units from
the 'late' latency interval (between 50 and 75 ms) had a larger latency variability as
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
,—Jl,
BULBOCAVERNOSUS REFLEX
Fio. 3. Bulbocavemosus reflex response of a single motor neuron represented by a
complex potential of 2 single muscle fibre action potentials belonging to this particular
motor unit. Responses to consecutive stimuli from above downwards. The signal is delayed
for 30 ms for clearer recognition. Double and triple discharges on individual electrical
stimuli can be seen occasionally. Single discharges may occur at variable discrete latency
values, which roughly correspond to latencies of individual discharges within a multiple
discharge.
compared with those from the 'early' latency interval (between 29 and 45 ms). Even
among the latter, those with short latencies tended to have less latency variability (fig.
4). The reflex latency variability of motor units from the 'late' latency interval was
between 1706 and 10901 /is (7 measured motor units); for the motor units from the
'early' latency interval it was between 387 and 2151 /ts (49 measured motor units).
(6) There was a whole group of motor units among those appearing at the early latency
interval which had a reflex latency variability around 500 /is (fig. 4).
To estimate the contribution of peripheral (also technical) factors to the reflex latency
variability, the variability of latencies of direct responses in the bulbocavemosus muscle
on pudendal nerve (branch) stimulation was also measured (fig. 4). In some measurements
low jitter values (5 us and below) were obtained and were discarded as they were believed
to be due to direct muscle stimulation (see Trontelj et al., 1986). The latency variability
of direct responses ranged from 8 to 50 /is (mean 26,9 ± 10 /ts; 33 measured motor
units).
DISCUSSION
Latency variability of reflex responses recorded with surface electrodes reflects the
variability of those motor units activated at the beginning of the reflex responses (Lloyd
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
ms
817
818
D. B. VODUSEK AND M. JANKO
60-,
^
*
40-
*
*
*
* * * * *
*
E
c
20H
1000
1500
Jitter (jis)
T11I I I | I I I I I I I I F 1
2000
2500
FIG. 4. Reflex latency variability of single motor neuron responses with latencies in the 'early' latency interval (stars)
as well as variability of direct responses (filled circles, lower left).
and Mclntyre, 1955; Rail and Hunt, 1956) and correlates with the complexity of their
reflex arc (Penders and Delwaide, 1973). Recording with the single fibre EMG (SFEMG)
technique allows for selective observation of single motor neuron reflex discharges in
humans (by detecting single muscle fibre potentials, which are marking individual motor
unit behaviour). Trontelj (with coworkers) pioneered this research, measuring the reflex
latency variability of the H reflex (Trontelj, 1968), the first and second components
of the blink reflex (Trontelj and Trontelj, 1973) and the flexor reflex (Janko and Trontelj,
1983). Reflex latency variability must of course involve both 'central' factors (variability
of synaptic delay, etc.) as well as 'peripheral' factors (variability at stimulation site,
the neuromuscular transmission, etc.). The peripheral factors, however, can be estimated
by measuring the latency variability of electrically elicited direct responses in the same
muscle. Thus Trontelj and Trontelj (1973) were able to show that the latency variability
of reflex motor neuron responses in the first component of the blink reflex was 25.5
times the latency variability of their direct responses. Our measurements give a value
of about 20 times greater variability for the most stable bulbocavernosus reflex responses,
as compared with latency variability of direct responses in the bulbocavernosus muscle.
We can also compare the bulbocavernosus reflex latency variability with the latency
variability of the monosynaptic H reflex. Its latency variability (expressed as SD of
mean latency) was shown to be ~ 150 fis (Trontelj, 1973), which makes the
bulbocavernosus latency variability ~ 3 times greater. This relatively 'small' variability
of bulbocavernosus reflex responses in the shorter latency range is quite unlike the typical
large latency variability of polysynaptic reflex responses, for instance for single motor
neuron responses in the second component of the blink reflex (Trontelj and Trontelj,
1973) or the flexor reflex (Janko and Trontelj, 1983). Such 'large' latency variability
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
FT I T TI | I I I I I I I I 1 | T 1 T
500
BULBOCAVERNOSUS REFLEX
819
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
(i.e., above 10 times H reflex latency variability) was recorded for bulbocavemosus
reflex responses in the 'late' latency interval. The latency variability for these late reflex
responses is actually underestimated in our study as responses with obviously large latency
variability were not selected for measurement. Some motor neuron responses from the
'early' latency range in the bulbocavemosus reflex also showed such larger latency
variability (and tended to have relatively longer mean latencies as the more stable
responses). Such motor neurons also often showed a bimodal or polymodal distribution
of reflex response latencies. A likely explanation of polymodal latency distribution of
reflex responses as well as of multiple discharges of the same motor neuron in a reflex
response is conduction through alternative interneuron pathways including a different
number of interneurons. This would also explain the relatively larger latency variability
of the motor neuron response appearing at the later latency value in a response with
a bimodal latency distribution. Similarly, the relatively larger latency variability of
bulbocavemosus reflex responses with longer latencies is probably reflecting additional
interneurons in their reflex arcs. There was also a correlation between the stimulus
strength and the reflex latency variability, the latter being decreased by an increase in
stimulus strength. This could be accounted for by an increase in steepness of the excitatory
postsynaptic potentials, caused by spatial facilitation (depolarization of more afferents
on stronger stimulation).
An interesting comparison can be made with the reported latency variability of responses
in the levator ani obtained by intramuscular electrical stimulation (Trontelj et al., 1974):
they observed direct, recurrent, and reflex responses. The latter had latencies between
about 27 and 85 ms, and their latency variabilities were grouped around the values 155 /ts,
550 /ts, 1200 /is and 2300 /ts. These latency variability values (and therefore the
complexities of the respective reflex arcs) were compared by the authors with the H
reflex, and the first and the second components of the blink reflex, respectively. In
our study of the bulbocavemosus reflex latency variability we clearly could not obtain
values close to 'H reflex' latency variability as no muscle afferents were stimulated.
However, the other values correspond remarkably well; this further strengthens our
opinion that identical reflex responses can actually be obtained in any perineal or pelvic
floor muscle on stimulation of sacral somatic afferents (Vodusek et al., 1983).
Many hundred consecutive reflex discharges of individual motor units on regular
repetitive weak suprathreshold stimulation in this study show absence of habituation
in the bulbocavemosus reflex pathway. Absence of habituation is well known for the
first component of the blink reflex and was taken by Rushworth (1967) as an argument
against the polysynaptic nature of that reflex pathway. Trontelj and Trontelj (1973)
convincingly argued that the reflex arc of the first component of the blink reflex is
oligosynaptic. As already discussed, the early bulbocavemosus reflex responses similarly
show (1) a comparable ratio of reflex versus direct response latency variability; (2) a
comparable ratio of H reflex versus bulbocavemosus reflex latency variability; and (3)
absence of habituation. We therefore postulate that the first component of the
bulbocavemosus reflex, which is by common assumption usually referred to as a
polysynaptic cutaneomuscular reflex, in fact has an oligosynaptic reflex pathway.
820
D. B. VODUSEK AND M. JANKO
REFERENCES
BORS E, FRENCH JD (1952) Management of paroxysmal hypertension following injuries to cervical and
upper thoracic segments of spinal cord. AMA Archives of Surgery, Chicago, 64, 803—812.
ERTEKIN C, REEL F (1976) Bulbocavernosus reflex in normal men and in patients with neurogenic bladder
and/or impotence. Journal of the Neurological Sciences, 28, 1 — 15.
JANKOM, TRONTEU JV (1983) Flexion withdrawal reflex as recorded from single human biceps femoris
motor neurones. Pain, 15, 167 — 176.
JANKO M, VODUSEK DB (1983) The study of the bulbocavernosus reflex by single fibre EMG.
Electroencephalography and Clinical Neurophysiology, 56, SI06.
KRANE RJ, SlROKY MB (1980) Studies on sacral-evoked potentials. Journal of Urology, 124, 872-876.
LAPIDESJ, BOBBITTJM (1956) Diagnostic value of bulbocavernous reflex. Journal of the American Medical
Association, 162, 971-972.
LLOYD DPC, MCINTYRE AK (1955) Monosynaptic reflex responses of individual motoneurons. Journal
and Psychiatry, 36, 951-959.
TRONTELJ JV, JANKO M, CODEC C, RAKOVEC S, TRONTELJ M (1974) Electrical stimulation for urinary
incontinence: a neurophysiological study. Urologia Internationalis, 29, 213—20.
TRONTELJ JK, MIHELIN M, PLETERSEKT, ANTONI L (1979) Jittermeter - a microcomputer-based system
for single fibre electromyography. International Journal of Bio-Medical Computing, 10, 451 —459.
TRONTELJ JV, MIHELIN M, FERNANDEZ JM, STALBERG E (1986) Axonal stimulation for end-plate jitter
studies. Journal of Neurology, Neurosurgery and Psychiatry, 49, 677—685.
TRONTELJ M, TRONTELJ JV (1973) First component of human blink reflex studied on single facial
motoneurones. Brain Research, Amsterdam, 53, 214—217.
VARMA JS, SMITH AN, MCINNES A (1986) Electrophysiological observations on the human pudendo-anal
reflex. Journal of Neurology, Neurosurgery and Psychiatry, 49, 1411 -1416.
VODUSEK DB, JANKO M, LOKAR J (1983) Direct and reflex responses in perineal muscles on electrical
stimulation. Journal of Neurology, Neurosurgery and Psychiatry, 46, 67—71.
(Received March 27, 1989. Revised June 20, 1989. Accepted July 17, 1989)
Downloaded from http://brain.oxfordjournals.org/ by guest on September 18, 2016
of General Physiology, 38, 771-787.
PENDERS CA, DELWAIDE PJ (1973) Physiologic approach to the human blink reflex. In: New Developments
in Electromyography and Clinical Neurophysiology, Volume 3. Edited by J. E. Desmedt. Basel: Karger,
pp. 649-657.
RALL W, HUNT CC (1956) Analysis of reflex variability in terms of partially correlated excitability
fluctuation in a population of motoneurons. Journal of General Physiology, 39, 397—422.
RUSHWORTH G (1967) Diagnostic value of the electromyographic study of reflex activity in man.
Electroencephalography and Clinical Neurophysiology, Supplement 25, 65—73.
STALBERG E, TRONTEU JV (1979) Single Fibre Electromyography. Old Woking: Mirvalle Press.
TRONTELJ JV (1968) H-reflex of single motoneurones in man. Nature, London, 220, 1043-1044.
TRONTELJ JV (1973) A study of the H-reflex by single fibre EMG. Journal of Neurology, Neurosurgery