Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishingplatforms. The new Wiley Online Library will be migrated over the weekend of February 24 & 25 and will be live onFebruary 26, 2018. For more information, please visit our migration page:http://www.wileyactual.com/WOLMigration/

Notice:

Log in / Register

Go to old article view

European Journal of Pain Explore this journal >

Original Article

Children with chronic stress-induced recurrent muscle pain haveenhanced startle reaction

First published:4 May 2017 Full publication history

DOI:10.1002/ejp.1057 View/save citation

Cited by (CrossRef):0 articles Check for updates

Funding sources None.

Conflicts of interest None declared.

G. Alfvén , S. Grillner, E. Andersson

View issue TOC Volume 21, Issue 9October 2017 Pages 1561–1570

Abstract

Background

Children with recurrent pain of negative chronic stress origin from different locations have acharacteristic pattern of tender points in the temporal, trapezoid, great pectoral and abdominalmuscles. We tested the hypothesis that the startle reaction is activated in these children and that someof the startle-activated muscles are related to the tender point pattern and the recurrent pain.

Methods

In children/adolescents, aged 10–17 years, 19 with recurrent psychosomatic pain (PAIN) and 23

Citation tools

Go To

controls (CON) we measured and analysed resting activity and acoustic startle response withelectromyography (EMG) for the muscles involved in the pattern of tender points and also the lumbarerector spinae.

Results

The PAIN group showed higher resting activity and higher acoustic startle response values than theCON group for all six muscles together regarding the mean amplitude in the initial 200 ms, and duringthe burst of activity, and longer burst duration and shorter burst latency. For PAIN versus CON, allseparate muscles showed generally higher values of EMG amplitudes and burst durations, and shorterlatencies for the burst onset in all measures; with significance or strong trends for several parametersand muscles.

Conclusion

For the first time in children with recurrent psychosomatic pain, increased resting activity andpotentiated startle response were demonstrated in the muscles involved in the stress tender pointpattern.

Significance

This study demonstrates in adolescents how recurrent pain of negative stress origin from the head,stomach, back and chest is related to increased startle reaction and increased muscular tension inthese regions. This study contributes to the understanding of the mechanisms underlying the globalburden of recurrent pain.

1 IntroductionRecurrent pain of ‘non-organic’ origin in childhood is frequent (Roth-Isigkeit et al., 2004), and oftenimpairs quality of life (Petersen, 2008). For a century, scientific studies have indicated that psychologicalproblems and stress are of major aetiological importance (Moro, 1913; Nylander, 1960; Apley, 1975), andthis has been corroborated in many recent studies (Bakoula et al., 2006; Murberg and Bru, 2007; Alfvenet al., 2008; Østerås et al., 2015).

Stress experience influences muscular tension. However, the tension and its possible relation to recurrentpsychosomatic pain have been little studied. In children with recurrent abdominal pain of ‘non-organic’origin Alfvén in 1993 described a specific pattern of tender points involving the temporal, the trapezius,the great pectoral and the rectus abdominal muscles (Alfvén, 1993a) (Fig. 1A), and for the first timedemonstrated statistically the co-occurrence of abdominal pain, headache and chest pains in this group ofchildren (Alfvén, 1993b). Alfvén suggested that the amygdala, the centre for the fear reaction (LeDouxet al., 1988), is also affected in stress (Alfvén, 1993c; p. 53), and this was corroborated later (McEwenand Gianaros, 2010). Based on the knowledge that the amygdala plays a pivotal role in fear eliciting thestartle reaction (Hitchcock and Davis, 1986), the possibility was raised that the pattern of muscle tension,tender points and pain is a manifestation of stress-induced activation and potentiation of the startlereaction (Alfvén, 1997, 1998). In all children examined since more than 20 years fulfilling the criteria forthe diagnosis of psychosomatic pain, see below, (Alfven, 2003), this pattern of tender points has beenfound.

The startle reflex is a well-known neuromuscular defence reaction elicited from the caudal reticularpontine nucleus (Yeomans and Frankland, 1995). It starts with the blink reflex – a rapid contraction of theorbicularis oculi muscle – and is followed by a forward thrusting of the head and a descending flexor wavereaction, extending through the trunk and the knees (Hunt and Landis, 1936) (Fig. 1B). It can be elicitedby many fear- and stress-provoking stimuli such as a strong acoustic signal (acoustic startle), but also byother signals such as unexpected visual stimuli. Stress also potentiates the startle reflex, i.e. makes itstronger and more easily elicited (Schmitz et al., 2011). It has become an often-used reflex todemonstrate stress and increased nervous sensibility in both animal and human studies (Grillon andBaas, 2003).

Hypotheses: Children with psychosomatic recurrent pain have due to stress a highermuscular resting activity and potentiated startle reflexes in muscles involved in the clinicalstress tender point pattern. These muscle points are co-located with areas where recurrentpsychosomatic pain is common.

Figure 1.Open in figure viewer

Stress tender points and typical startle reaction position. (A) From a photo of an adolescent with

long-standing recurrent psychosomatic pain. Typical stress tender point pattern shown as grey

dots, cf Figure B. (B) Hunt and Landis classical picture from 1936 (18) of a person in startle

after unexpected pistol shot near the ear with a typical crouch, cf figure A.

1.1 Aims

To investigate the muscular resting activity and the acoustic startle response with EMG in children withrecurrent psychosomatic pain, and specifically the muscles involved in the stress tender point pattern, andalso the lumbar erector spinae.

2 Patients and methods2.1 Subjects

The subjects were nineteen children, 11 girls and eight boys, mean age 12.7 (10–17) years, with pain of‘non-organic origin’ recurring at least once a week for more than 3 months and affecting activity of dailylife. Eighteen had had abdominal pain for in mean 37.2 (8–140) months all fulfilling diagnosis criteria offunctional abdominal pain. All reported recurrent headache, five backache and three shoulder pain. Allreported stress and anxiety problems mainly related to school and family and fulfilled six or seven ofseven criteria for the diagnosis of psychosomatic pain (Alfven, 2003), see below.

Eighteen children had nine of nine stress tender points and one had seven (Alfvén, 1993a). One child hadschool-stress-induced depression. All the children met the same paediatrician (G.A.), a specialist in painin children at a special ward in the Stockholm area, between 2011 and 2014, at several consultations untildiagnosis.

Matched controls were 23 children, 14 girls and nine boys, mean age 13.0 (10–17) years withoutrecurrent pain problems and in general good health. Fifteen were enrolled from four school classes, andeight from personal contacts. Socio-economic status was comparable in the two groups.

In both groups, no hearing difficulties, enuresis, ADHD, panic disorder or post-traumatic stress disorderwere reported.

Informed consent from child and parents was obtained. The study was approved by the regional ethiccommittee Stockholm, Sweden (reg. no. 2013/17-31/3).

Acoustic Startle Responses (ASR) were elicited with a 50-ms-duration stimulus of unexpected shortwhite noise (USWN) at 105 dB with an instantaneous rise time in both ears via headphones. The child laysupine in a calm milieu. The stimulus was controlled with a digital audio-stimulator. Eight ASR were givenwith varying time intervals, generally between 1.5 and 2.0 min, similarly for all subjects (Andersson et al.,1997; Blumenthal et al., 2005).

2.2 Electromyography

ASR was studied with electromyography (EMG) for six muscles: OR-orbicularis occuli (registered justbelow the lower eyelid), TE-temporalis (temple area), TR-trapezius (mid-portion between acromion andC7), PE-pectoralis (lateral mid-portion on the trunk), RA-rectus abdominis (mid-portion at umbilicus level)and ES-lumbar erector spinae (L3–L4 about 2–3 cm lateral to the spinal processes) (Fig. 2B). The erectorspinae muscle was included in the experiment as recurrent back pain is common in children withpsychosomatic pain, and we wanted to examine the possible involvement of this muscle in the startlereaction.

Two surface electrodes (N-00-S; Medicotest, Denmark, with a 10 mm pick-up diameter) were placed witha 2 cm inter-electrode central distance and a ground electrode for each muscle. EMG signals wererecorded and analysed with an eight channel-system (Muscle tester model ME3000P8, MegaWinsoftware, Mega Electronics Ltd Kuopio Finland/Meditech Sverige AB Sweden). The signals were amplified375 times with a pre-amplifier, initially band-pass filtered (8–500 Hz), AD-converted with a sampling-frequency of 1000 Hz, transferred to a computer and later rectified. The sampled raw EMG signals werefurther high-pass filtered at 50 Hz to eliminate artefacts, e.g. due to movements (Andersson et al., 1997;Blumenthal et al., 2005). For each muscle, we calculated peak magnitude, mean amplitude (RMSaverage expressed in uV) over various intervals and some timing parameters (cf. below). Mean values(M) and standard deviation (SD) were calculated for each test parameter and muscle.

From the rectified EMG, we first recorded the lowest rest activity level (µV). This was followed by themaximal peak value (µV during 1 ms) and time (ms) to peak of at least 10 µV seen within 200 ms from

Figure 2.Open in figure viewer

Analyses and example of EMG response after a sound stimuli in a PAIN-subject. (A) Schematic-

image for burst of activity, with onset/offset (ms), peak value (µV), and average amplitude (µV)

during 50 ms, 100 ms and 200 ms (from OR-activity-start) and during burst . (B) EMG

response with y-scale up to 50 µV for all muscles, except OR-orbicularis occuli (up to 250 µV).

Other muscles were TE-temporalis, TR-trapezius, PE-pectoralis, RA-rectus abdominis and ES-

erector spinae.

the USWN of an activity lasting at least 10 ms (Fig. 2A). Further, the average amplitude level (RMS inµV) was assessed at three time intervals of 50 ms, 100 ms and 200 ms from start of OR-activity(≥10 µV). We noted also the main burst activity (Fig. 2A). In addition, we measured burst duration andtime to start of burst activity from USWN (ms). An examiner manually marked response onset and offset.Artefacts due to, for example, heartbeat, loose electrodes were carefully avoided.

Regarding the burst, in some muscles occasionally a top value of ≥10 µV within the 200 ms after theUSWN could not be found. In those cases, the average activity level was analysed only during a veryshort period (~6–10 ms, i.e. just a few ms around the maximum amplitude <10 µV). Then, only burstmean amplitude, and not burst duration or latency, was included in the analyses.

Finally, EMG was measured while repeated sound events (USWN) were given (with shorter timeintervals), when the examiner retold a severe personal stress-provoking event told by the child before therecording session. The stressful story had been with the child for a longer time.

The Health Questionnaire for Children and Young People with 52 questions on 10 items ( Kid Screen -52 )(Ravens-Sieberer et al., 2005) was posed and the raw scale scores were transformed by hand to Raschperson parameter estimates.

The Verbal Rating Scale for Stress 0–5 (Alfvén and Nilsson, 2014), validated and tested for reliability,with six verbal statements (from: I didn′t feel any stress at all, = 0, to I felt the worst stress I can thinkof, = 5), was used.

In the analysis, individuals were randomly matched regarding age and sex between the PAIN and theCON groups. This was done because statistically higher values were generally found for the youngerchildren (10–13 years) than for those aged 14–17 years in the CON group, but not in the PAIN group, fora third of the parameters tested for TR, PE and ES. No statistical differences were seen between boysand girls for the two groups.

The seven criteria for the diagnosis of psychosomatic pain (cf. above) are: (1) onset or aggravation ofchronic negative stress at the time of onset of recurrent pain, (2) pain in parallel with chronic negativestress, (3) feeling better or pain-free during periods of no or lessened chronic negative stress, (4) stress-induced acute pain, (5) most pain attacks related to acute stress, (6) the child was followed up for at least1 year, (7) the parents and child and their doctor agree on the diagnosis (Alfven, 2003).

2.3 Statistic

The Mann–Whitney U test was performed to detect significant differences between the PAIN and theCON groups. This was because the Kolmogorov–Smirnov test showed some non-normally distributedparameters. The significance level chosen was P < 0.05 marked with an *. Sometimes, a strong trendwas seen just above the p -value limit and this was marked with a (*).

3 ResultsA typical Acoustic Startle Response (ASR) EMG from the six recorded muscles from an adolescent girlwith psychosomatic pain is shown in Fig. 2B. Characteristically, the first response occurred in OR,followed by TE and then TR together with PE and RA, and finally ES (cf. Table 1).

Table 1. For the groups with psychosomatic pain (PAIN) and without pain (CON), meanvalues (±SD) are shown for each muscle regarding: occurrence of a burst (≥10 µV, 1 = yesand 2 = no); peak amplitude (µV); peak latency from the unexpected short white noise

(USWN, ms); average RMS amplitude (µV) for the fixed time intervals of 50, 100 and200 ms from the onset of activity for OR; average RMS amplitude during the burst (µV);duration of the burst (ms); latency for start of burst activity (≥10 µV) from the sound (ms)

Burst Peak Peak 50 ms 100 ms 200 ms Burst Burst Burst

≥10 µV Ampl Latency Ampl Ampl Ampl Ampl Duration Latency

1 = yes,2 = no µV ms µV µV µV µV ms ms

OR

8STARTLE– PAIN:

1,0 214 72 45 37 28 40 96 37

SD 0,0 94 10 18 20 23 16 38 5

8STARTLE– CON

1,0 187 73 40 28 18 32 83 40

SD 0,0 132 10 27 19 12 17 33 7

HIGHESTof 8- PAIN

1,0 329 72 55 48 59 49 120 37

SD 0,0 138 19 20 26 121 23 66 7

HIGHESTof 8- CON

1,0 272 76 54 40 24 40 96 38

SD 0,0 149 21 31 24 14 19 49 7

STORY-highest-PAIN

1,0 229 63 42 29 18 36 78 33

SD 0,0 111 11 24 17 10 18 39 7

STORY-highest-CON

1,0 200 64 40 24 15 33 68 40

SD 0,0 164 10 35 22 16 20 39 9

TE

8STARTLE– PAIN

1,04( ) 62 76 13 11 8 15 78 47

*

* * * * * * * *

SD 0,15 29 14 6 5 4 6 41 8

8STARTLE– CON

1,13 38 74 9 7 6 11 47 60

SD 0,20 31 9 7 5 4 6 35 14

HIGHESTof 8- PAIN

1,10 84 84 15 13 11 18 92 45

SD 0,20 57 39 7 7 6 8 58 8

HIGHESTof 8- CON

1,10 44 83 10 8 9 13 59 63

SD 0,30 42 41 8 6 14 10 40 27

STORY-highest-PAIN

1,00 63 78 13 10( ) 8 19 49( ) 58

SD 0,00 56 34 9 7 7 15 49 34

STORY-highest-CON

1,20 33 70 10 7 5 15 26 58

SD 0,40 22 14 12 7 4 15 19 16

TR

8STARTLE– PAIN

1,30 177 106 8 26 20( ) 25 75 74( )

SD 0,38 381 30 10 57 37 38 53 34

8STARTLE– CON

1,43 42 106 5 7 5 8 49 87

SD 0,42 62 24 5 8 4 8 34 24

HIGHESTof 8- PAIN

1,20 397 103 12 47 43 50 101 56

SD 0,40 931 40 14 95 95 83 86 17

HIGESTof 8- CON

1,30 98 111 12 15 9 13 72 79

SD 0,50 281 41 33 34 18 21 51 36

* * * * * *

* * * * *

* *

*

* * * * *

STORY-highest-PAIN

1,60 67 88 4 10 9 12 69 63( )

SD 0,50 187 49 4 25 16 26 72 51

STORY-highest-CON

1,70 8 135 2 2 2 3 13 131

SD 0,50 3 42 1 1 1 2 10 42

PE

8STARTLE– PAIN

1,31( ) 66( ) 102 6 10( ) 9 16( ) 64 81

SD 0,38 80 21 6 12 10 15 30 28

8STARTLE– CON

1,48 31 102 6 5 4 8 40 85

SD 0,35 36 24 6 6 4 7 30 20

HIGHESTof 8- PAIN

1,20 130 93 12 18( ) 19 28 87 62

SD 0,40 224 37 19 30 34 40 75 33

HIGESTof 8- CON

1,40 40 90 8 7 6 9 60 68

SD 0,50 65 29 12 10 8 10 56 26

STORY-highest-PAIN

1,70 115 137 4 24( ) 16 15 126 113

SD 0,50 368 54 9 86 52 41 133 79

STORY-highest-CON

1,70 8 100 2 2 2 4 13 98

SD 0,40 4 46 1 1 1 3 11 48

RA

8STARTLE– PAIN

1,55 28 95 3 5 5 9 42 77

* * * * *

* * * * * *

* *

* *

* *

SD 0,42 37 14 2 7 5 9 21 12

8STARTLE– CON

1,75 27 109 3 4 3 7 37 94

SD 0,30 48 15 7 6 4 8 24 20

HIGHESTof 8- PAIN

1,50 45 101 4 8 7 12 55 71

SD 0,50 68 13 5 13 10 15 25 16

HIGESTof 8- CON

1,70 78 98 10 11 7 13 54 74

SD 0,50 209 15 37 28 15 30 42 19

STORY-highest-PAIN

1,90( ) 21 105 1,3( ) 3( ) 3 5 87 88

SD 0,30 65 – 1,0 9 8 10 77 –

STORY-highest-CON

2,00 4 – 0,8 1 1 2 – –

SD 0,00 2 – 0,7 1 1 1 – –

ES

8STARTLE– PAIN

1,55 27( ) 129 2,0 2,9 3,8 8 39 109

SD 0,39 30 35 1,0 1,4 3,4 8 22 32

8STARTLE– CON

1,80 12 130 1,0 1,4 1,3 3 24 126

SD 0,27 17 30 0,7 1,4 1,3 4 18 26

HIGHESTof 8- PAIN

1,50 38 160 1,7 3,0 4,5 10 59 127

SD 0,50 44 39 0,7 2,6 4,3 10 27 44

HIGESTof 8-nCON

1,80 15 138 1,3 1,8 1,5 4 37 105

SD 0,40 28 29 0,9 2,8 1,7 6 19 12

* * *

* * * * *

* * * *

The level of resting activity in the PAIN group compared to the CON group was significantly higher for allsix recorded muscles together and also separately for a majority of the muscles (Fig. 3, all SDs between0.6 and 1.4). This means that there is a higher level of background activity in these muscles.

Mean values for EMG results are presented for all six muscles in: (1) all 8 USWN together, (2) the ASR, at the highestpeak activity in OR, for each muscle among the 8 USWN (with no further provocation), and finally (3) ASR, at the highestpeak activity in OR, for each muscle among all USWN when retelling a story with an unpleasant memory. A significantdifference between the two groups is marked with *(P < 0.05). A p-value just above this limit indicates a trend and ismarked with (*).

STORY-highest-PAIN

1,60 10 95 2,2 2,3 2,8 3,9 21 80

SD 0,50 8 65 1,9 1,8 3,1 2,6 27 39

STORY-highest-CON

2,00 4 – 0,9 0,9 0,8 1,5 – –

SD 0,20 1 – 0,8 0,8 0,7 0,8 – –

* * * * *

Figure 3.

The startle response for the PAIN group had significantly higher mean amplitudes (µV) than the CON forall six muscles and the eight startle responses together, both in the fixed time period of 200 ms from thestart of OR-activity (12.3 ± 11.3 vs. 6.3 ± 3.4), and during the entire burst (18.8 ± 12.8 vs. 11.7 ± 5.7,Fig. 4A). Higher mean amplitudes were generally noted in both parameters for all individual muscles, withmost showing significance or strong trends (TE,PE,TR,ES, Table 1).

Open in figure viewer

Mean EMG amplitude (µ V ) during rest. Values are shown for all six muscles together and

separate muscles: OR-orbicularis occuli, TE- temporalis, TR-trapezius, PE-pectoralis, RA-rectus

abdominis and ES-erector spinae. PAIN versus CON showed significantly higher values (dark

vs. light grey bars) for all muscles together and for majority of each muscle. (Strong trend for PE

* P = 0.052).

Figure 4.Open in figure viewer

Average for all eight startle sound events together (no other stress provocation), including all six

muscles, for: (A) Mean amplitude (µV) during 200 ms from OR-activity-start and during burst ,

(B) Burst duration and - latency (ms) from sound to muscle-activity-start (≥10 µV), (C) Average

amount of bursts: –mean of 1.0 = all muscles burst (peak ≥ 10 µV) and 2.0 = no muscle had

burst (peak ≥ 10 µV) in any of the eight stimuli. PAIN versus CON showed significant differences

(star). (OR, orbicularis occuli).

Significantly, longer duration and a shorter latency for burst (ms) for PAIN (64.9 ± 27.5 and 64.0 ± 15.9)than for CON (48.8 ± 18.0 and 74.4 ± 9.7, respectively), were noted for all six muscles and all eight ASRtogether (Fig. 4B). The same pattern was generally seen for all individual muscles, but this was significantonly for TE, PE and RA (with a trend for TR) in one or both of these parameters (Table 1).

Calculating the average amount of burst activity, we have used an index where the value of 1.0 indicatesthat all muscles had a burst (peak at least 10 µV) whereas 2.0 indicates that no muscle expressed a burst(peak at least 10 µV) for any of the eight sound stimuli. PAIN, compared to CON, had significantly largernumber of muscles with a burst (1.28 ± 0.26 vs. 1.42 ± 0.20), Fig. 4C. A similar pattern was seen for allsix individual muscles, with strong trends for TE and PE (Table 1). In OR all persons in both groupsalways showed a burst (with a peak of 10 µV or more).

For a retold stressful story , the PAIN group had significantly higher values than the CON group in theASR when all six muscles were included regarding both the peak amplitude (90.3 ± 122.0 vs.45.5 ± 34.7 µV, Fig. 5A) and mean amplitude within 200 ms from the start of activity in OR (9.5 ± 15.2 vs.4.4 ± 3.5 µV, Fig. 5B). For all individual muscles, higher mean values were seen in PAIN than in CON,where the discrepancy was significant for TE and ES in both parameters and for TR and PE within200 ms (Table 1).

Figure 5.Open in figure viewer

Generally, in all measures all separate muscles showed higher values of EMG amplitude parameters andburst duration and shorter latencies for burst onset in PAIN than in CON. This was so for (1) the mean ofeight ASR, (2) the highest ASR for OR of all eight unprovoked stimuli and (3) the highest ASR for OR ofall story-provoked stimuli (Table 1). Several parameters showed significant differences or strong trends,especially for TE, TR, PE and ES, and sometimes for RA (Table 1).

Similar patterns were found, with a few exceptions, when comparing the highest startle response in thestress-provoked situation with the highest response of the initial eight startle sounds without provocation(Table 1). In both situations, higher mean values for PAIN than for CON were most often seen regardingsome amplitude parameters and burst duration. For both conditions, several parameters showedsignificant differences or strong trends for muscles TE, PE and ES, and this was also true in the provokedsituation in TR and RA. In contrast to the PAIN group, no person in the CON group showed a burst with apeak of at least 10 µV for either RA or ES during the story retelling.

The highest OR-activity of all eight initial ASR was in average number 3.0 ± 2.0 for the PAIN group and3.8 ± 2.2 for the CON group, with no significant difference between the groups.

The Verbal Rating Scale for Stress showed significantly higher values for the personal stress-provokingstory for PAIN (mean 3.1) than for CON (2.1).

Kid Screen in the PAIN group showed significantly diminished physical activity (16.7 vs. 19.9) and a lessgood family situation (25.6 vs. 28.6), but no difference was found for the other eight items. The questionsof the screening were posed immediately before the investigation, which could have made the responsessuboptimal.

Retold stressful story – amplitudes – in highest response – average for all six muscles together:

In highest response (of all startle responses while retelling unpleasant-story) (A) Peak

amplitude (µV) within 200 ms from sound, and (B) Mean amplitude (µV) during 200 ms from

start of OR-activity. In both comparisons, group with psychosomatic pain-PAIN had significantly

higher values (star) than controls-CON (dark vs. light grey bars). (OR, orbicularis occuli).

4 DiscussionThe experience of stress is subjective and not measurable with objective methods. Reliable reports ofpersonal stress and stress-provoking pain experience, prerequisites for psychosomatic diagnosis, have tobe caught in a confident and long-term contact with a person knowledgeable regarding stress and pain.These demands were met in this study, which makes the diagnosis reliable. Reliable methods wereapplied for studying Acoustic Startle Response (ASR) in a well-defined group of children with stress,anxiety and recurrent pain. Results from several amplitude and timing parameters as well as muscleswere measured and separately presented, which previously have not been done.

For the first time, it has been demonstrated that children with psychosomatic recurrent pain haveaugmented resting activity in the pain affected muscles and potentiated, earlier and longer activity in thestartle response in muscles involved in the stress tender point pattern (STP) and the lumbar erectorspinae. These effects can be explained by altered descending neural activity from higher brain centres,such as the amygdala via the pontine nucleus in chronic stress (Hitchcock and Davis, 1986). Recurrentpsychosomatic pains such as headache, stomach pain, shoulder and chest pain share location with thestress tender point pattern (Alfvén, 1993a,d). We suggest that the increased resting activity in itself andalso the earlier elicited and potentiated startle reaction can promote the stress tender point pattern andalso induce recurrent psychosomatic pain. To our knowledge, there are no earlier clinical studies

concerning startle and recurrent pain.

An increased excitability in spinal cord α-motor neurons may be due to supraspinal projections, but alsodue to elevated γ-motor activity and muscle-spindle input (Ia, II) and other afferents (III, IV) contributing tochronic hypertonicity and pain (Johansson and Sojka, 1991; Djupsjöbacka et al., 1994, 1995a,b;Knutson, 2000). Here, stimulation of local chemo-nociceptors by various substances (i.e. bradykinin,prostaglandins, arachidonic acid, potassium and lactic acid) may contribute.

In chronic muscular pain, it has been shown that mental stressors can trigger long-lasting hypersensitivityof nociceptors in response to a subsequent exposure to a low concentration of inflammatory mediators(Reichling et al., 2013). Of special interest is a study in rat showing that unexpected short white noise of105 dB, the same as used in this study, elicits hyperalgesic priming in the masseter muscle, known to bepart of the startle reaction. This will lower the pain threshold (Reichling et al., 2011). Such lowering of painthreshold can via CNS feedback loops enhance the level of muscle activity and promote startle reactions.

Further research is desirable regarding how stress can result in increased resting activity and potentiatedstartle reaction.

The elevated resting activity in the different muscles studied will increase muscular tension seen in thePAIN group strengthens the hypothesis presented earlier, suggesting that psychosomatic pain in childrenis promoted by muscular tension (Alfvén, 1997).

Children with functional abdominal pain have been shown to have enhanced startle response, butindividual muscles were not studied Bakker et al., 2010). The magnitude of the startle response inorbicularis occuli muscle in healthy adults can increase depending on kind of pain stimuli and threat levels(Bustan et al., 2015; Horn-Hofmann and Lautenbacher, 2015).

High oxytocin levels can attenuate the startle reflex (Ellenbogen et al., 2014), and conversely low levelsas shown in children with psychosomatic pain (Alfvén, 2004) could have an opposite effect. Cortisol canpotentiate the startle reflex (Lee et al., 1994). Children with psychosomatic pain can have increasedcortisol levels (Törnhage and Alfvén, 2015), and thus high cortisol can be another contributing factor.

The Verbal Rating Scale for Stress demonstrated, as expected, higher stress values for the personalstress-provoking story in the PAIN group. To our surprise, the highest peak activity was generally notgreater for each muscle in the stress-provoked situation (with retold stressful story) compared to highestpeak activity among the eight initial unprovoked startle events (Table 1). One reason could be ahabituation effect (Kofler et al., 2001) since the final stress-provoked startle sounds were given withshorter time intervals and after the initial eight unprovoked startle events.

4.1 Possible confounding factors

The startle can be increased in children with enuresis (Ornitz et al., 1999), but here no child reported bed-wetting. In adolescent girls, but not boys, baseline startle decreases over the course of puberty (Schmitzet al., 2014), which tallies partly with our findings. The prepulse inhibition (PPI) of the startle reaction isimmature in children below 10 years (Gebhardt et al., 2012), but should be of no significance in the agegroup studied. The startle is influenced by genetic factors (Zhang et al., 2011) and early life experiencessuch as maltreatment, but these factors were not considered in the present design.

5 ConclusionsFor the first time, increased muscular resting activity and a potentiation of the startle reaction have beendemonstrated in children with recurrent psychosomatic pain in muscles involved in the stress tender point

References

Browse Publications

Browse by Subject

Resources

Help & Support

Cookies & Privacy

Terms of Service

About Us

pattern. It is expected that these findings stimulate further studies and improve clinical practice.

AcknowledgementsWe wish to thank all following persons for invaluable help with different aspects of the study: schoolnurses Olga Apatisidou and Karin Dahl-Klingvall regarding the children, Tim Crosfield for languagerevisions. For contribution with development of test equipment regarding recordings of simultaneous EMGand stable sound stimuli (for acoustic startle response, ASR), we want to acknowledge engineer KlasAndersson, engineer and sound expert Uno Persson (PC-Persson), manager Leif Häggkvist (MeditechSverige AB), professor and engineer Leif Kari, research engineer Kent Lindgren, engineer and PhD UlfCarlsson, the three latter persons at KTH Royal Institute of Technology, Stockholm, Sweden.

Author contributionsDr Alfvén in his clinical research discovered what is termed the pattern of tender points in children withrecurrent psychosomatic pain. He proposed that its neuro-anatomy can be explained by the startle reflex.Dr Alfvén recruited the patients and the school-children in the present control group and had a major partin the design and execution of the study. He is responsible for the Kid Screen and Verbal Rating of StressScale. He drafted the initial manuscript and approved the final version as submitted. Dr Grillner is thesenior advisor for this study with whom Dr Alfvén discussed his hypothesis and who suggested examiningthe startle reflex in an EMG-study. He brought in and inspired Dr Andersson, an EMG-expert, as a partnerin the study. He attended and contributed at several initial meetings regarding the study design, reviewedand revised the manuscript and approved the final version as submitted. Dr Andersson took part in thedesign and execution of the study and recruited some of the controls. She is responsible for the EMGanalyses, product development, tests, analysis and presentation of the EMG-method and data. She haswritten and revised portions in the initial manuscript and approved the final manuscript as submitted.

Related content

Wiley Job Network

Advertisers & Agents

Powered by Wiley Online Library Copyright © 1999 - 2018 John Wiley & Sons, Inc. All Rights Reserved