The Effects of Normal Human Variability and
Hand Activity on Sensory Testing with the Full
Semmes-Weinstein Monofilaments Kit
Nicola Massy-Westropp, OTR, CHT, MHealth
Department of Orthopaedic Surgery
Flinders Medical Centre and Repatriation General Hospital
Daw Park
Australia
The Semmes-Weinstein (S-W) monofilaments are a
handheld sensory assessment originally intended to
detect areas of skin that have abnormal sensation and
to record the degree of severity of the sensory distur-
bance occurring as a result of brain injury.1 Since
their development, the S-W monofilaments have
demonstrated their ability to detect abnormal func-
tion in peripheral nerves.2–4
Bell-Krotoski and Tomancik5 validated the forces
applied by the S-W monofilaments. They concluded
that if the monofilaments as supplied by the manu-
facturer were of correct length and diameter, the tests
are repeatable within a small, predictable range. Bell-
Krotoski and Tomancik add that the manufacturers
of the monofilaments “did not intend to provide spe-
cific measurable thresholds of force or stress but a
relative range of progressive pressures.”
All assessment instruments may provide different
results when tested on the same person a second
time. The first reason for a variation in results is
instrument-related error. The S-W monofilaments
may deliver a different range of pressures if a
monofilament is the wrong length, kinked, or even
This research was supported by North-Eastern Community
Hospital, Campbelltown, South Australia, and by ATL Ultra-
sound, South Australia.
Correspondence and reprint requests to Nicola Massy-Westropp,
OTR, CHT, 20 Barr Smith Ave, Myrtle Bank, Australia, 5064; e-
mail: <mwestropp@picknowl.com.au>.
48 JOURNAL OF HAND THERAPY
ABSTRACT:: The first part of this study aimed to determine the
effect of human variability on assessment using Semmes-Weinstein
monofilaments. Fifteen healthy subjects were tested using a full
monofilament kit of known calibration. When the subjects were
retested, they responded to a different filament 48% of the time. The
second part of the study aimed to determine the effects of vigorous
hand activity on sensation, as measured with Semmes-Weinstein
monofilaments. Forty randomly selected healthy subjects were
assessed before and after they performed a brief, standardized grip-
ping activity. After hand activity, the subjects responded to a dif-
ferent filament 53% of the time. In both parts of this study, the
changes in monofilament results were random in direction and
magnitude. The number of changes did vary between test sites, the
radial nerve sites being most variable and the ulnar nerve sites least
variable. It is concluded that the results of testing with the five
smallest monofilaments in the full Semmes-Weinstein kit were not
reliable in normal subjects and that they did not detect change in
sensory function in normal subjects.
J HAND THER. 2002;15:48–52.
fatigued after multiple uses.6 The second reason is
examiner-related error. The examiner may distract
the subject or record results incorrectly. A third rea-
son is human variation. Human variation is true
change that occurs independently of the test process.
To detect change, an instrument must possess sensi-
tive increments of measurement and be reliable
enough to ensure that variation interpreted as change
is not examiner or instrument error. The S-W
monofilaments have been manufactured to decrease
instrument error, and protocols for their use have
been developed to decrease examiner error.
The third source of variability in the S-W monofil-
aments is to be further investigated. The first aim of
this study is to determine the effects of normal
human variation on testing with the monofilaments.
Studies that have investigated the sensitivity to
change of the S-W monofilaments have varying
results. Gelberman et al.3 induced median nerve com-
pression in 12 healthy subjects, using increasing
amounts of external pressure over the carpal tunnel.
Eleven of the 12 subjects in this experiment com-
plained of paresthesia in the hand and then recorded
abnormal sensation when tested with S-W monofila-
ments. Braun et al.4 performed S-W monofilament
assessments on patients who had symptoms of carpal
tunnel syndrome. Results of the S-W monofilament
assessments were normal before the patients per-
formed 7 minutes of vigorous hand activity. Abnormal
results of S-W monofilament assessments occurred in
34 of 40 patients after they performed the activity.
These findings suggest that S-W monofilaments are
sensitive to fluctuations in median nerve function.
Although these studies demonstrate abnormalities
in S-W monofilament minikit results with patient
activity, other authors have found no sensory
changes with patient activity. The relationships
between sensory testing with S-W monofilaments
and activity have been explored in 20 normal sub-
jects. McGough and Zurwasky7 revealed that after 5
minutes of vigorous hand activity, the sensory status
of 20 normal subjects (as measured with the monofil-
aments) did not change. Gillenson et al.8 evaluated
the effect of wrist flexion and extension compared
with the normal test position of a neutral wrist on
results of monofilament testing of 31 normal subjects.
No consistent changes occurred in normal subjects.
All these studies used the S-W monofilaments
minikit. This kit has five filaments from the range of
filaments included in the full kit. The minikit fila-
ments assess the same overall range of pressures as
the full 20-filament kit. It is hypothesized that the the
20-filament kit is more sensitive to small changes
than the minikit, because of the small measurement
increments between each filament. Because studies of
the effects of activity on testing with the minikit have
conflicting results, a second aim of this study is to
determine the effects of vigorous hand activity on
testing with the full S-W monofilaments kit.
METHOD
Variability Study
Bell-Krotoski9 suggests a method of hand screening
using S-W monofilaments that is designed to save
time and to be easily repeated for serial testing. Three
sites on the hand are selected for median nerve test-
ing, three for ulnar nerve testing, and one for radial
nerve testing. The sites for median nerve testing are
the pulp of the thumb, the pulp of the index finger,
and the volar proximal phalanx of the index finger.
This test was modified to exclude the hypothenar
eminence site for brevity, as the aim of the study is
not detection of pathology in a specific nerve. The
radial nerve site was repeated close to the second
metacarpal for ease in location.
A full S-W monofilaments kit was assessed to
determine its calibration, using methods similar to
those of Bell-Krotoski et al.10 The filaments demon-
strated less than 8% variability between each appli-
cation and no overlap in the pressures delivered by
the filaments. During the assessment of calibration,
the filaments did not demonstrate any fatigue over a
minimum of 10 trials each.
Following ethical approval, a convenience sample
of 10 adult volunteers was sought from the Univer-
sity of South Australia. Subjects were questioned to
rule out any neck or upper limb injury, condition, or
symptoms in the past year. The screen test recom-
mended by Bell-Krotoski9 and modifications as
described above were applied. Each site was tested
five times in the manner recommended by Stone,11
by one examiner. Each hand was tested. The order of
the sites tested and the timing of application were
varied. Right and left hands were tested alternately.
If a filament slipped on the skin, the application was
repeated. If a response was noted on any of the five
applications, the site was scored as positive. Any sub-
ject failing to record normal sensation at least one
time10 at any site was excluded from the study.
Activity Study
Following ethical approval, 40 randomly selected
students were questioned in the same way as the sam-
ple above. The subjects were tested using the method
described above, before hand activity, immediately
after hand activity, 5 minutes after hand activity, and
10 minutes after hand activity. Only the dominant
hand was assessed. With the dominant hand, the sub-
jects used pruning shears that had a spring opening to
cut a dowel rod 6 mm in diameter. They cut for 5 min-
utes at a rate of four cuts every 5 seconds.
Data Analysis
The monofilaments provide increasing increments
of pressure. The scale by which they increase is irreg-
ular, however, and the values for each filament are a
range rather than a discrete value.12 For these rea-
sons, the filaments were treated statistically as adja-
cent categories.
Subjects were analyzed as a group and then sepa-
rated for analysis by gender. Gender-related variabil-
ity in sensation has been reported by Doeland13 and
Bell-Krotoski.9
For both parts of the study, agreements were com-
pared with disagreements that occurred between test
and retest. Subjects were compared for the number of
agreements recorded between sites and within either
median or ulnar nerve innervation zones (since there
were more than one site in these). Kappa scores were
calculated for each site. Left and right hands and
dominant and nondominant hands were considered
separately for the human variability study.
RESULTS
Human Variability
Thirty hands of 15 adult volunteers were tested.
The volunteers included seven men and eight
women from 27 to 38 years of age. Test and retest
occurred within 10 minutes.
January–March 2002 49
 TABLE 1. The Semmes-Weinstein Monofilaments
Screen Test: Number of Agreed Responses on
Repeated Testing
Agreements
Nerve and Site
No. % responded to a different filament 19% of the time.
Radial:
Dorsal MP joint of index finger 8/30 27
Median:
Fingertip of thumb (volar) 13/30 43
Volar proximal phalanx , index finger 17/30 57
Fingertip of index finger 19/30 63
Ulnar:
Volar proximal phalanx , small finger 19/30 63
Fingertip of small finger 17/30 57
TOTAL AGREEMENTS
93/180 –
OVERALL PERCENTAGE – 52
Both hands were tested, resulting in 30 tests of each
site. All subjects responded to the five smallest fila-
ments in the S-W monofilaments full kit in the screen
test. The range of agreements scored by each subject
between test and retest was 17% to 75%. Male and
female subjects were represented throughout this
range.
The number of agreements between test and retest
were 93 of 180 total tests (52% agreements).
Disagreements between test and retest occurred in
random directions, i.e., subjects detected both small-
er and larger filaments during the retest. There was
no relation between dominance or side tested and a
greater number of agreements. Kappa scores for each
site ranged from 0.09 (radial site) to 0.21 (index and
small proximal phalanges), which indicated poor
test–retest reliability.14
Of the 87 disagreements, 82 occurred between adja-
cent filaments. Five of the 87 disagreements were two
to three filaments apart. No one filament demonstrat-
ed a greater number of agreements or disagreements.
The number of agreements differed between the
various nerve innervations. Disagreements within
the ulnar nerve sites occurred only between adjacent
categories. In the median nerve and radial sensory
nerve sites, five disagreements were two or three
monofilaments apart. Table 1 shows the agreements
between test and retest for each site on the screen test.
Activity Study
Nineteen male and 21 female subjects between 17
and 56 years of age participated in this part of the
study. Hand activity did not produce any gender-
related variability in responses to the S-W monofila-
ments; however, the test time in relation to the activ-
ity and the test site affected the number of variable
responses.
The subjects responded to a different filament after
hand activity 53% of the time. The changes were ran-
50 JOURNAL OF HAND THERAPY
dom in direction and were usually to an adjacent fila-
ment. Between the post-test and the post-5 test, sub-
jects responded to a different filament 36% of the time.
Between the post-5 and the post-10 tests, subjects
Results for the radial nerve test site varied the most
in both the variability and the activity studies, with
agreements recorded 25% of the time only. The medi-
an nerve sites recorded fewer agreements than the
ulnar nerve sites in both the reliability and the activ-
ity studies. Figure 1 shows the number of changes in
responses between the radial, median, and ulnar
nerve sites.
DISCUSSION
The results of this study indicate that the 15 sub-
jects in the variability study and the 40 subjects in the
activity study had normal light-touch sensation. The
results suggest that it is normal to have variability in
light-touch perception among the five smallest fila-
ments. Because of the nature of the S-W monofila-
ments, this study could not provide a range in mil-
ligrams of human variability; however, in normal
subjects, variability between one and three of the five
smallest filaments can occur. This occurrence leads to
the question, “Why have five filaments in the normal
range?”
Theories to explain the variability of detection of
the smallest filaments are suggested below. Test and
retest variations may have been due to examiner- and
protocol-related errors. The exact spot in the hand
that the monofilament contacted may have varied;
that is, the monofilament may have touched a slight-
ly different site on the hand on the second test than
on the first.
Sensory receptor density can vary between indi-
vidual subjects and between different ages in indi-
vidual subjects.15 Perhaps receptor density also
varies between various zones of the hand, allowing
different thresholds to be detected in the same area.
FIGURE 1. Percentage changes in each nerve distribution
between tests. Black column indicates radial nerve site; gray,
median nerve site; white, ulnar nerve site.
It is not clear from this study that marking each site
with a pen would increase reliability, nor would
marking be practicable between therapy visits in a
clinical setting.
Human variability may account for “lapses” in
detection. For example, temporary decreases in
peripheral circulation may cause some receptors to
“sleep”. Doeland et al.13 suggest that decreased
peripheral circulation causes age-related sensory
diminution. This same phenomenon may occur tem-
porarily in all persons.
Bell-Krotoski9 states that the aim of sensory testing
is not to record sensation “100% of the time.” This
also explains why Bell-Krotoski16 and Stone11 suggest
that each site is tested at least three times and that
one response was to be taken as positive. This study
allowed for less than 100% detection by testing of
each site five times and recording one response as
positive. In an alternative method of interpretation17
found in the literature, researchers recorded a posi-
tive response only after three consecutive positive
responses from the subject. The reliability or variabil-
ity associated with this method was not stated.
The gate-control theory of Melzack and Wall18 may
also explain human variability in detecting sensation
with the smallest filaments. This theory states that
responses carried by the large A-alpha fibers, which
conduct light touch, can be modulated at the level of
the spinal cord. This means that some of these
impulses may be conducted but never consciously
received by the brain. This may occur even if a sub-
ject is concentrating on the test but lapses into other
thoughts or receives another internal signal, such as
the need to shift off a pressure point on the seat.
During testing, the 2.83 filament, which is considered
best for detecting normal sensation, is somehow
“tuned out” less often than smaller filaments.
These theories do not explain the differences in the
number of agreements between the radial, median,
and ulnar sites. These differences were constant
throughout the variability and activity studies, the
radial site being least reliable.
This might be explained by the difference in the
type of skin, the skin at the radial nerve site being
glabrous and that at the other sites being
nonglabrous. Nonglabrous skin has a quickly adapt-
ing mechanoreceptor (the hair follicle) different from
that of glabrous skin (Pacinian and Meissner cells).
Nonglabrous skin also has a slowly adapting
mechanoreceptor (the Merkel cell) different from that
of glabrous skin (Ruffini organs).19 Perhaps the relia-
bility of these receptors is more variable in glabrous
skin. This observation does not account for the find-
ing that the results at the ulnar nerve sites were
somewhat more reliable than the results at the medi-
an nerve sites, since both are glabrous skin.
Changes in detection of filaments cannot be attrib-
uted to learning effects through the test processes.
Subjects did not become more sensitive or the results
less variable as testing progressed. The results reflect
neither a learning effect nor an effect of mental
fatigue during testing.
Reliability or variability? The very nature of the S-W
monofilaments instrument may create variability
when the test is reapplied. The S-W monofilaments in
this test kit deliver a range of forces up to 8% of their
total force. Each S-W monofilament delivers a force
within a small range of forces.5 It is possible that the
filament delivers the upper limit of its range of force
on the first test and the lower limit of force on the sec-
ond. The subject may not be able to detect force at the
lower limit of this particular filament.
The design of this study precluded testing of these
theories; however, the findings do support the origi-
nal purpose of the S-W monofilaments, to define nor-
mal and abnormal areas of cutaneous sensation.5,20
The aim of the S-W monofilaments test is not to quan-
tify levels of normality, only normality and abnor-
mality. It appears that human variability does not
allow for reliable quantification of normality but dis-
tinguishes between normal and abnormal on repeat-
ed testing.
The findings in this study did not support the find-
ings of studies suggesting that female subjects are
able to detect more subtle levels of touch-pressure
than male subjects.10,13 The authors of these studies
did not report on the variability for male and female
female subjects but rather their sensation thresholds
on one testing.
The participants in the human variability study
and the activity study had normal sensation. It
remains unknown whether subjects with abnormal
sensation would provide more or less reliable
responses. The sampling method for the variability
study did not provide a sample large or variable
enough to be generalized to the population at large.
The random sampling method used in the activity
study reduced the likelihood of bias in some senses ,
but the sample comprised of university students,
illustrated in the low mean age.
CONCLUSIONS
Quantification of “normal” sensation using the five
smallest filaments in the full S-W monofilaments kit
did not produce the same results on repeat testing.
Clinicians are encouraged to explore the purpose of
S-W monofilaments testing with regard to choosing
which “normal” filaments will be used.
Scores indicated very poor agreement between the
smallest five filaments on test and retest. On testing,
changes would have to be of greater magnitude or
consistent direction or would have to cross from nor-
mal to abnormal categories to be attributed to the
experiment.
Test results were most variable at the radial senso-
January–March 2002 51
ry nerve site and least variable at the ulnar sites in
both the variability and activity studies. Reasons for
this were not clear.
The ulnar and median sites demonstrated the same
percentage of agreements, but the ulnar nerve site
demonstrated disagreements only between adjacent
filaments.
The effects of variability require further research
among “borderline” subjects. If the results for one
subject can vary between the 3.22 filament and the
3.61 filament, can they vary between a “normal” and
an “abnormal” interpretation?
Acknowledgments
The author thanks Dr. Bernie Hughes (University of South
Australia school of Pharmacy) for assistance in the monofil-
ament calibration process, Dr.Brenton Dansie for statistical
advice, and Associate Professor Esther May (University of
South Australia School of Occupational Therapy) for loan
of equipment.
REFERENCES
1. Weinstein S. Fifty years of somatosensory research: from the
Semmes-Weinstein monofilaments to the Weinstein Enhanced
Sensory Test. J Hand Ther. 1993;6:11–22.
2. McGill M, Molyneaux L, Spencer R, Fan Heng L, Yue DK.
Possible sources of discrepancies in the use of the Semmes-
Weinstein monofilament: impact on prevalence of insensate foot
and workload requirements. Diabetes Care. 1999;22:598–607.
3. Gelberman RH, Szabo RM, Williamson RV, Dimick MP.
Sensibility testing in peripheral-nerve compression syn-
dromes: an experimental study in humans. J Bone Joint Surg.
1983;65A:632–8.
4. Braun RM, Davidson K, Doehr S. Provocative testing in the
diagnosis of dynamic carpal tunnel syndrome. J Hand Surg.
1989;14A:195–7.
5. Bell-Krotoski J, Tomancik E. The repeatability of testing with
Semmes-Weinstein monofilaments. J Hand Surg. 1987;12A:155–61.
52 JOURNAL OF HAND THERAPY
6. Booth J, Young JM. Differences in the performance of commer-
cially available 10-g monofilaments. Diabetes Care. 2000;23:984.
7. McGough CE, Zurwasky ML. Effect of exercise on volumetric
and sensory status of the asymptomatic hand. J Hand Ther.
1991;4:177–80.
8. Gillenson S, Parets N, Bear-Lehman J, Stanton D. The effect of
wrist position on testing light touch sensation using the
Semmes-Weinstein pressure aesthesiometer. J Hand Ther.
1998;11:27–31.
9. Bell-Krotoski JA. Sensibility testing: current concepts. In: Hunter
JM, Mackin EJ, Callahan AD (eds). Rehabilitation of the Hand:
Surgery and Therapy. 4th ed. St. Louis, Mo.: Mosby, 1995:109–28.
10. Bell-Krotoski J, Fess EE, Figarola JH, Hiltz D. Threshold detec-
tion and Semmes-Weinstein monofilaments. J Hand Ther 1995;
8:155–62.
11. Stone J. Sensibility. In: Casanova J (ed). Clinical Assessment
Recommendations. 2nd ed. Chicago, Ill.: ASHT, 1992:71–84.
12. Dellon AL, Mackinnon SE, Brandt KE. The markings of the
Semmes-Weinstein monofilaments. J Hand Surg. 1993;18A:756–7.
13. Doeland HJ, Nauta JB, van Zandbergen JB, et al. The relation-
ship of cold and warmth cutaneous sensation to age and gen-
der. Muscle Nerve. 1989;12:712–5.
14. Fleiss JL. The Design and Analysis of Clinical Experiments. 3rd
ed. New York: Wiley, 1986.
15. Weinstein S, Drozdenko R, Weinstein C. Evaluation of sensory
measures in neuropathy, tendon and nerve surgery in the hand.
In: Hunter JM, Schneider L, Mackin EJ (eds). Rehabilitation of
the Hand: Surgery and Therapy. 4th ed. St. Louis, Mo.: Mosby,
1995:63–76.
16. Hunter JM, Mackin EJ, Callahan AD (eds). Rehabilitation of the
Hand: Surgery and Therapy. 4th ed. St. Louis, Mo.: Mosby, 1995.
17. Malchaire J, Rodriguez Diaz LS, Piette A, Goncalves Amaral F,
de Schaetzen D. Neurological and functional effects of short-
term exposure to hand-arm vibration. Int Arch Occup Environ
Health. 1998;71:270–6.
18. Lankford LL. Reflex sympathetic dystrophy. In: Rehabilitation
of the Hand: Surgery and Therapy. In: Hunter JM, Mackin EJ,
Callahan AD (eds). Rehabilitation of the Hand: Surgery and
Therapy. 4th ed. St. Louis, Mo.: Mosby, 1995:794–5.
19. Dellon AL. Somatosensory Testing and Rehabilitation. Bethesda,
Md.: American Occupational Therapy Association, 1997.
20. MacDermid JC, Kramer JF, Roth JH. Decision making in detected
abnormal Semmes-Weinstein monofilament thresholds in carpal
tunnel syndrome. J Hand Ther. 1994;7(3):158–62.