Irzternatiortal Journal of Pwchophysio/ogy, 13 ( 1992) 7 1-77 0 1992 Elsevier Science Publishers B.V. All rights reserved 0167~8760/92/$05.00
71
INTPSY 00398
Richard A. Sherman, Vernice D. Griffin, Cecile B. Evans and Anita S. Grana Fitzsimons Army Medical Center. Aurora, CO (USA)
(Accepted 30 April 1992)
Key words: Phantom limb pain; Muscle tension; Mechanism
Previous studies of relationships between surface EMG of the residual limb and phantom pain have not shown which changed first. Thus, predictive relationships could not be demonstrated. 24 male (20) and female (4) amputees between the ages of 33 and 71 who reported either burning (3). cramping (Sk shocking-shooting-stabbing (61, or a combination of these descriptions of phantom pain (7) participated in one or two recording sessions. Raw surface EMG from the major muscles of the residual limb was recorded while subjects activated an event marker to indicate changes in pain. All eight subjects with cramping phantom pain reported changes in pain after the recording showed sharply demarcated increases in EMG. Subjects reporting either shockingshooting or burning pain did not show any consistent relationships between EMG dnd pain. Three of the four subjects reporting experiencing both shocking-shooting and cramping phantom pain simultaneously during recordings showed changes in EMG preceding changes in pain. Sensations of cramping phantom pain were preceded by increases in muscle tension in the residual limb in almost every instance for each of our subjects showing changes in cramping phantom pain. Thus. changes in muscle tension in the residual limb are likely to either be causes or close intermediaries for the cause of cramping phantom pain but not necessarily of other common descriptors.
INTRODUCTION In the past, the ‘cause-or-reaction’ quandary has prevented determining whether changes in muscle tension which are related to changes in phantom pain are causes of the pain or simply reactions to it. For example, our previous inlaboratory studies of relationships between sur-
Correspondence to: R.A. Sherman, Psychophysiology Service, Dept. of Clinical Investigation, Fitzsimons Army Medical Center, Aurora, CO 80045-5001. USA. The opinions and assertions contained in this manuscript are the private views of the authors and are not to be construed as official or as reflecting the views of the United States Departments of Veterans Affairs, Army, or Defense.
face electromyograms (EMGs) of the residual limb and cramping phantom pain show that they change together from day to day (Sherman and owever, these studies could not indicate whether pain or EMG changed first because (a) for most amputees, phantom pain intensity usually changes only a few times per day (Sherman and Sherman, 1985) and (b) recordings were only made for 9, 10 s intervals over a period of several minutes. Other studies have demonstrated the utility of surface electromyograyhic biofeedback to help amputees control their pain (Sherman, 1976, 1991). Many of the amputees who showed !ong term success from the treatment reported spasms in their residual limbs which appeared to be related to onset or intensity of their phantom pain. Spike activity was ob-
72
served on an oscilloscope for these subjects and it was found that they could reduce the intensity af their pain or prevent onset of brief episodes if they learned to control the spike activity through observing the oscilloscope (Sherman et al., 19791. A general relationship between overall tension in the residual limb and intensity of the phantom pain from week to week was also noted. Many of the subjects participating in large respondent surveys (Sherman and Sherman, 1983, Sherman et al., 1985) and home log studies (Sherman et al., 1988) reported that their pain changed when they contracted their residual limbs or did more work than usual. Thus, it is very likely that there is a causal relationship between change in residual limb muscle tension and change in phantom pain. Such a relationship is difficult to demonstrate because lengthy recordings would have to be made
from those amputees who show frequent changes in phantom pain intensity. Surface electromyography ( shown to be a reliable, noni sure relative intrasubject strength of muscle contraction in the residual Limb. Andreassi (1980) and Basmajian and DeLuca (198.5) have reviewed the literature relating the surface EMG signal to overall force applied by t e muscle and contraction intensity. The consensus of opinion is that the surface EMC signal is consistently and reliably proportional to the force and contraction intensity for each subject. This study attempted to overcome the iimitations of earlier recording studies by utilizing subjects who reported that they went through periods during which their phantom pain frequently changed significantly every few minutes during
TABLE I Characteristics of participating amputees and cf their recordings
‘Types of phantom pain’ refers to the descriptions of phantom pain usually experienced rather than to descriptions specifically reported during recordings. S#
1 2 3 4 5
Age
61 3x 65 60 44
6 7
33 47
8
39
9
62
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
44 74 67 58 58 s9
62 41 64 63 71 41 48 70 63
Sex
M M M M F M
M M M M M M M M M F M F F M M M M M
Nwnfw
Tjpes of
M4scle
of Rcdings
phantom puin
recorded
2 2 43 24 5 31 1 6
1 1 1 1 1 2 1 2 1 1 1 1 2 1 1 1 1
i
i
1 1 22 25 1 2
1 1 I 2 I 1
stab burn, cramp. stab tingle cramp shock, burn. cramp shoot. cramp burn, tingle, sharp no pain shock burn burn, shock cramp, burn, shoot cramp burn shock, shoot cramp. shoot cramp shock cramp burn, cramp. shoot cramp shock, burn cramp shock
Location of arnputatiorr
YrS since amp
above knee (AK) Below knee (BK) AK BK at knee bilateral BK BK BK AK AK at shoulder BR at hip BK BK above elbow (AE) AE bilat toes AK BE BE BE bilateral AK at pelvis
41 12 46 3 i 12 25 7
i
from vast lat gastroc add magnus ant tib vast lat vas med gastroc vas med vast lat vast lat post trap gastroc glut med per lat lat soleus tricep tricep ant foot vast lat tricep ulnaris tricep rect fern glut max
II
Pain did not
EMG/no
EMG
* EMG 1 EM6
8.
36
4
IO
did not change.
7.3+7.1 4.5 -t 9.7
during recordings.
Seconds Bet EMG
and Pn’ while the WX
confidence
EMG/no
EMG/no
EMG/few
chng
chng
chngs
EMG/changes
chng
chng
(standard
deviation)
#
to the type of change in EMG most common for
one session during which they
O-3-8.8
-3.3-17.8
2.6-3.2
in the cofumn titled ‘Mean & SD. in the next column to the right.
and Change in pain is reported
limits around the mean are reported
that subject. The average number of seconds between change in EMG
Times pain chngd with common reft’ refers to the number of times changes in pain were temporally closely related
change’ means that EMG
refers to the descriptions of phantom pain actually reported
Pain Description’
3 16
2.9 + 0.4
recording sessions. When subjects were recorded at more than
before pain
before pain
before pain
close to same time
the same description of phantom pain, the data for both sessions are combined.
‘Phantom
”
change
change
descriptors of phantom pain at different
EMG/no
EMG/no
‘EMG/no
reported
* Subject 22 reported different
No pain
I H
22 *
crmp/brn/shk
Stabbing
fh
20
crmp/shk
12
crmp/brn/shk
If
brn/shk
3 ir 5
crmp/shk
shk/brn/stb
hrn/crmp/stb
Combinations
24
1x
EMG/chngs - 0.2-0.8
0.4-0.9
0.7-2.3
EMG/few
10
i .!I
1.1-2.4
0.6
9
0.3 + 0.9
0.7 f 0.5
0.3- 1.2 0.9-2.6
recording
IS 27
16
1.5+ 1.0
1.8-F I.1
1.2+0.5
f.7+0.8
0.8-3.2
No pain during
shooting
before pain
19 12
2.0 f 2.2 0.8 + 0.9
Iimits
ccmfidence
99%
3
EMG
EMG/no
10
n
16 Close to same time
23 7 change
12
22 *
14
s 19
19
21
20
21
5
22
bet EM.2 & pit
common relt
6
before pain
# seconds
mean & S.D.
chngd with
6
A EMG
pain chngd
relutionship
# times pain
17
change
change
times
Phantom purn changed dwing the recording most common
24
EMG/no
EMG/no
recording
change drrrintg
1-3
6
4
Sitbject
bmve~~~ phantom limb pain arrd mlcscle tension ijl the residual limb
Shocking-
Burning
Cramping
descriptioir
Phuntom puitt
TLmporul reiutionships
TABLE
2:
74
waking hours. Other subjects who relatively stable intensities of phantom li were recorded to provide a comparison normal
t t
4
female amputees. w
quently changed significantly participated in either one or two agreeing to participate
the residual and intact limbs onto a digital chart recorder and, frequently, a digital tape recorder. The signals were filtered and amplified by CoulV digital chart recorder. To cstab-
Fig. i. Samples of a chart recording showing surface EMG
from the residua
c~~rnpi~~ ~h~~~~rn pair intensity made by a Sr
ges
i
t never cone-
f 8 bifatecal amputee
can naIs when phantom pain e of them. This is imporany major reaction in the
n to pain is being recorded as
to anything including
were very distinct from move-
rting changes in cramping phancreases in the residual limb’s EMG o 9 s before the button was pressed cate onset of, or increase in, phantom pain = I.3, SD. = I.4). With the exception of etween
‘pain
signifying’ button-press
re-
sponses foIlowing increased EMG activity were Zonger than the reaction times evoked by light flashes. The fight evoked reaction times were significantly shorter than the times between G and the button press indicating n (independent Student’s c = 3.6 Qt Degrees of Free one taiIed P = 0. Most of the vari in time was between subjects. One subject (number four) who reported onty cramping phantom pain during the recording, did not report any changes in pain intensity during the recording and did not show any major ffuctuations in EMG other then a few related to movement. Three subjects reported a combination of cramping and other descriptors of pain during the recording. One of these (number 16) did not report any changes in pain during the recording and no major changes in EMG were recorded. Another (number 12) did show a predictive relationship between changes in EMG and pain for all ten changes in pain reported. The third subject (number 20) reporting mixed descriptions of phantom pain during the recording, did show the predictive relationship between change in EMG and change in pain for three of the four changes in pain but the pain was described mostly as burning during these changes. As can be seen from inspection of TabIe II, there are no consistent relationships between change in EMG and change in phantom pain for other descriptions of phantom pain. Only one subject with shocking phantom pain (number 18) showed even an occasional relationship between EMG and pain (seven of 27 instances). The abrupt increases in EMG typically found among people with cramping phantom pain did not occur among subjects who either did not experience any phantom pain or changes in their phantom pain during their recordings (including the subject whose cramping phantom pain did not change during the recording). These changes occurred occasionally among five subjects whose pain changed during the recording who did not report cramping phantom pain. One of these subjects, who normally reported combined burning and shocking pain, had two such changes during a pain free period. One subject reporting
burning pain showed three such changes at virtually the same time as changes in phantom pain and one such change unrelated to phantom pain. Two amputees reporting shocking pain showed one and two such changes, respectively, during pain free recordings. A third person with shocking pain showed four such changes at almost the same time pain changed and five changes just before pain changed.
same efwere in progress. Patients reporte fects while at home. The predictive nature of t cramping phantom pain relationship is sufficiently well establishe to warrant development of treatments based on reducing overall leve muscle tension and spike E G activity in the residual limb. These treatm for patients who report era but not necessarily for tho sensations.
DISCUSSION Changes in reported sensations of cral.rping phantom pain were preceded by increases in muscle tension in the residual limb in almost every instance for each subject reporting such changes. However, only eight amputees who reported changes in cramping phantom pain during their recordings, participated in the study. It is possible that exceptions to this relationship might turn up if a larger group was recorded. The predictive relationship between change in EMG and change in cramping pain shown by our small sample can not prove that cramping phantom pain is actually directly caused by increased muscle tension. It certainly increases the likelihood that changes in muscle tension are eit underlying cause or an intermediary between the cause and sensation. We feel that it is more likely to be a direct cause rather than an intermediary iti% ave emonstrated that biofeedback of muscle tension and EMG spike activity from the resic?ual limb in which the aim is to decrease overall tension level and spike activity usually leads to long term reduction and/or disappearance of phantom pain among the vast majority of subjects receiving the treatment (Sherman et al.. 1979; Sherman et al., 19911. IMost of the patients in these studies quickly learned to recognize the relationship between EMG spikes from the residual limb displayed on an oscilloscope and rapid subsequent occurrence of cramping phantom pain. When the patients were able to prevent the spikes from occurring, they did not experience cramping phantom pain. We observed this directly in the laboratory while recordings
studies have been supported by studies in other animals (Laughlin and Armstrong, 1985). is possible that th relationship betwee people with phanto amputees reportin out simultaneous tionship between pain. Only five amputees only three had bur
few subjects having sol ticipated for us to fin between changes in EMG and changes in their pain. This may be the case since one subject with burning and one with shocking pain did show consistent EMG change-pain change relationships. that the length of time between activity and button-press indicating change in pain was consistently longer when the reaction time to a light flash, does not prove that the change in EMG came before the change in pain because one would expect the beginning of potentially less abrupt changes in internal state to be less easy to recognize, and thus, less rapidly responded to, than an abrupt, expected, light
77
blink. However, the 5-fold difference (from an average of 0.25 s to 1.3 s) is too much to account for through this difference. The difference in variability from a coefficient of variation (S.D. x f 36 to 107 would be expected with e from interpreting the start of an brupt stimulus such as a light to interpreting the start n unanticipated, potentially irregularly gradu ange in pain. It was difficult to perform in-laboratory recordings of a putees during episodes containent, significant changes in phantom pain because such episodes are unpredictable. Without these individuals. there would emonstrate a temporal ges in EMG and intenpain in the laboratory setting. atients having any severe pain problem (a) who have localized muscle tension abnormalities, (b) whose pain magnitude changes significantly every few moments, and (c) can pret when they are likely to experience rapidly nging pain well enough to schedule in-laboratory recordings, significantly limits the value of the techniques described above. Ambulatory recordings of muscle tension, movement, and logs of pain can provide very similar data for a much wider variety of people over a much longer time . The only disadvantage is that subjects ot to log changes in pain intensity as quickly as they do in the laboratory. When a few minutes variability in knowing when pain changes can be tolerated, we recommend the use of ambulatory monitors to resolve ‘cause-vs.-reaction’ quandaries.
ACKNOWLEDGEMENTS ER
AND
DISCLAIM-
This study was entirely supported by the US Army’s Health Services Command and the Department of Veterans Affairs. However, the opinions and assertions contained in this manuscript are the private views of the authors and are not
to be construed as official or as reflecting the views of the United States Departments of Veterans Affairs, Army, or Defense. We gratefully acknowledge the assistance of Crystal J. Sherman, MS in recording several of the subjects and in reviewing this manuscript. We also acknowledge the recommendations of one of the anonymous reviewers who suggested several ways to clari& the presentation of the results and to statistically support their significance.
REFERENCES J. ( 1989) Psychophysiology: Human behavior and psychophysiotogica~ response, Lawrence Eiibaum Associ-
Andreassi,
ates, Hillsdale. Basmajian, J-V. and DeLuca, C.J. (1985) Muscles ALi!-e:Their firncfions rereajed, 5th ed., Williams and Wilkins, Baltimore. Laughlin, M.H. and Armstrong, R.B. (1985) Muscle blood
Gormly, J. (1979) Treatment phantom limb pain with muscular relaxation training disrupt the pain-anxiety-tension cycle. Pain, 6: 47-55.
of to
Sherman, R.A. and Sherman, C.J. (1985! A comparicnn of phantom sensations among amputees whose amputations were of civilian and military origins. Pain, 21: 91-97. Sherman, R.A., Sherman, C.J. and Parker, L. (1984) Chroak phantom and stump pain among American veterans: iesuits of a survey. Pain, 18: 83-95.