Journal ofP.rvchosomaric Printed m Great Britain.
Research. Vol
36. No 2. pp
137-148.
1992. 0
0022-3999192 $5.00+.00 199: Pergamon Press plc
PSYCHOPHYSIOLOGICAL MECHANISMS OF CHRONIC HEADACHES: INVESTIGATION USING PAIN INDUCTION AND PAIN REDUCTION PROCEDURES P. R. MARTIN, G. V. MARIE and P. R. NATHAN (Received
3 January
1991; accepted
26 June 1991)
Abstract-Psychophysiological investigations of tension headaches and migraines have produced inconsistent and inconclusive findings. The study reported here aimed to further explore headache mechanisms by repeating earlier research with two main variations. The first pertained to recording and analytic procedures: a different approach was taken to measuring and interpreting cephalic vascular activity. The second was in the experimental conditions used: pain levels were manipulated so that the psychophysiology of headaches could be studied under conditions of pain increasing and decreasing. The findings provided no support for the role of muscle tension as a headache mechanism. Tentative support was obtained for vascular models of headaches. Headaches were associated with distension of the superficial temporal artery between pressure pulses. The data suggested that local vasomotor activity might play a role in dilating the temporal artery but that the effect seemed driven by elevated blood pressure and heart rate.
INTRODUCTION
TRADITIONAL beliefs concerning the pain mechanisms of the two common types of headaches are that tension headaches are caused by ‘sustained contraction of skeletal muscles’ and migraines are caused by ‘cranial arterial distension and dilatation’ Ill. Since the mid-1970s, over 100 studies have published data relevant to these hypothesized mechanisms [2] and many of the findings have been inconsistent with the classic conceptions (see Refs [3, 41 for reviews). For example, studies comparing tension headache sufferers with controls in terms of pericranial EMG have often failed to find significant differences between the resting levels of the two groups [5, 61 while others have [71. Several studies have found no significant differences in stress response between these two groups [8, 91 although some have 1101. Most studies have failed to differentiate the groups with respect to recovery from stress [ll, 121 but others have [131. Comparisons of migraineurs and controls at rest have resulted in a number of significant findings including abnormalities of temporal blood flow [ 14, 151, elevated heart rate 116, 111 and elevated diastolic blood pressure [161. Other studies have failed to find abnormalities on cardiovascular variables [9, 171. Distension of temporal arteries in migraineurs relative to controls has been demonstrated in response to noise [ 121, drinking ice water [ 181, and mental arithmetic [ 161, but other studies have been unable to differentiate migraineurs and controls in terms of vascular reactivity to stress [8, 111. One study reported significantly more vasoconstriction measured from the earlobe in a migraine group during post-stress recovery t51, but another study found no differences between migraineurs and controls on four cardiovascular
Department of Psychology, The University of Western Australia. Address for correspondence: Dr Paul R. Martin, Department of Psychology, Western Australia, Nedlands 6009, W.A., Australia. 137
The University
of
138
P. R. MARTINet al.
variables measured during post-stress recovery [ 191. Several studies have reported vascular abnormalities associated with tension headaches 120, 211, and other studies have reported EMG abnormalities associated with migraine [7, 221. It again needs emphasizing, however, that other studies have failed to find such differences [8, 91. While inconsistencies in the literature preclude drawing any conclusions with respect to headache mechanisms, the authors have argued elsewhere [3, 231 that vascular hypotheses look more promising than muscular hypotheses. A key problem for vascular models, however, lies in measuring the hypothesized mechanism of ‘cranial arterial distension and dilatation’. This has usually been accomplished via the indirect method of photoplethysmography . Photoplethysmography is based on the principle that light reflected from the tissue under the transducer is a function of local blood flow I241. A host of factors other than blood flow may affect the signal recorded by the polygraph, however, such as the precise position of the transducer relative to the artery, how the transducer is attached to the skin, skin characteristics, and technical parameters of the recording apparatus. Researchers have responded to these problems by using relative rather than absolute measures in their analyses 1251. Pulse amplitude is used as the unit of measurement for example, which consists of the amount of displacement in the polygraph signal caused by the passage of the pressure pulse along the artery. Also, some researchers analyse changes in pulse amplitude from baseline recordings rather than absolute pulse amplitudes. Some experts on plethysmography have argued that a consequence of the methodology being an indirect measure and one which is affected by a number of variables is that plethysmography results cannot be compared between laboratories, individuals or even with individuals on different occasions [241. If this argument is accepted then the approach has serious limitations for investigating headache mechanisms as most vascular hypotheses pertain to differences between headache sufferers and nonheadache controls, or between individuals when suffering from headaches and when headache-free. Also, using pulse amplitude as the unit of measurement results in a confounding of two variables: distension of the artery between pressure pulses and distension of the artery during the passage of pressure pulses. The relationship between arterial diameter and pulse amplitude depends on when distension occurs: distension during the passage of the pulse but not between pulses would result in an increase in pulse amplitude; equal distension during and between pulses would result in no change in pulse amplitude; and distension between pulses but not during passage of the pulses would result in a decrease in pulse amplitude. Another problem with the approach commonly used is that it provides no information about mechanisms of change in vascular activity. Headache researchers have frequently interpreted observed changes in pulse amplitude as indicating peripheral sympathetic vasomotor activity. Such an inference is not justified as observed changes may arise from alterations elsewhere in the cardiovascular system such as blood pressure and heart rate. Put simply. an artery may be distended as a result of relaxation of local smooth muscle or elevated blood pressure. It is for this reason that experts in the measurement of peripheral vascular activity recommend that these other cardiovascular variables should also be measured 1241. The approach taken here was to use a photoelectric transducer and to sample the amplified output rapidly with a laboratory computer. This would enable indirect
139
Mechanisms of chronic headaches measurement
of the maximum,
minimum
and average
distension
of the artery
to be
stored for later analysis. Rather than calculating difference scores between the maximum and minimum values which would approximate the usual pulse amplitude measure, the measures would be analysed separately to avoid confounding. This study also included continuous measures of blood pressure and heart rate so that the mechanisms of vascular changes could be explored. A final variation in this study was in the experimental conditions employed.
Instead of comparing headache sufferers and controls before, during and after stress, this study aimed to manipulate head pain while monitoring psychophysiological processes that might med;ate the pain. METHOD Subjects
Headache sub_jects were recruited via newspaper articles. Inclusion criteria for the study were: (a) headache frequency of at least one a week; (b) headache chronicity of at least 12 months; (c) diagnosed by family doctor as suffering from tension headache, migraine or combined headache; and average
(d) aged 19-59 yr. Control subjects were recruited by asking headache subjects to bring along a friend of same sex and similar age (? 3 yr) who does not experience regular headaches (an average frequency of less than one per month). This procedure was followed because it results in controls who are matched for age and sex, but also tends to result in matching on other variables such as socioeconomic status and education. Headaches were diagnosed as in our previous studies [26, 271. The approach used was a slightly modified version of one developed by Philips 1281 and is based on three migrainous features similar to those used by many neurologists [291: (a) unilateral onset of headache; (b) accompanied by nausea and/or vomiting; and (c) sensory prodromata. Subjects responded to questions about each of these features using a five-point scale (never, rarely, sometimes, usually, always). The diagnosis ‘migraine’ was applied if patients responded ‘always’ or ‘usually’ to two or three of the three questions (or if they responded ‘always’ or ‘usually’ to ‘accompanied by vomiting’); ‘tension’ was applied if they responded ‘never’ or ‘rarely’ to two or three of the three questions; and ‘mixed’ was applied in all other cases. A total of 5.5 headache and I4 control subjects satisfied the inclusion criteria. Subject characteristics are presented in Table I. TABLE
N
Group
Migraine Tension Mixed Control *Standard
19 23 I3 I4 deviations
I.--SUBJECT CHARACTERISTICS
Sex ratio (M : F) 2 : I7 9 : 14 4:9 4 : 10 included
Mean age
(yr) 39.3 37.6 39.1 33.1
(10.6)* (IO. I) (9.5) (9.5)
Mean headache chronicity (yr) 16.6 (11.7) 16.6 (11.1) 20.9 (12.1) -
in parentheses.
Apparutus Physiological activity was recorded with a Grass model 7D polygraph and a Cromemco Z-2D computer. Electromyographic (EMG) activity was monitored with three electrodes placed over the frontalis muscle two and a half centimetres above the eyebrows. Raw EMG up to a frequency of 500 HZ was sampled every millisecond, enabling average EMG levels to be calculated. Electrocardiographic (ECG) activity was also monitored every millisecond, electrodes being placed at the top of the sternum. at the fifth left intercostal space, and at the left elbow (ground). The radial artery pulse was continuously monitored by an electromechanical surface transducer [301 whose output was also sampled at 1 KHZ, enabling calculation of pulse transit time (radial R-wave to pulse interval, RPI). Cerebral arterial activity
was monitored with a photoelectric pulse sensor (Grass model PPS) placed over the zygomatico-facial branch of the superficial temporal artery. Absolute SBP and DBP measurements were taken during the session with a Copal model U-215 electronic sphygmomanometer. All sessions
were carried
out in a temperature-controlled
(2O”C), sound-attenuated,
psychophysiology
140
P. R. MARTIN et al
laboratory. The laboratory was divided door and a one-way mirror.
into subject
and experimenter
areas by a wall incorporating
a
A total of IO variables were recorded: eight physiological (six continuous variables using the polygraph and two periodic measures); and two subjective, PM~~, P~/il; crnd P,~Y. P,,x corresponds to the peak of the temporal arterial pulse signal and ih an index of the maximum distension of the artery; PU~N corresponds to the pre-pulse minimum level of the pulse signal and is an index of the minimum distension of the artery; PAv is an average of the samples taken of the pulse signal. Radial RPI. This is the interval between the occurrence of the ECG R-wave and the arrival of the associated pressure pulse at a point in the radial artery. Radial RPI is inversely correlated with systolic blood pressure [311, and to a lesser extent with diastolic blood pressure. Interbeat interval. Interbeat interval (IBI) is the time in millisec between successive ECG R-waves and is therefore an inverse of heart rate. Individual IBIS were averaged over 30.set intervals. Forehead EMG. Averages were computed based on samples of raw EMG taken throughout the session. Each of the above variables was averaged over 30.set intervals for the purpose of statistical analysis. S~stofic and diastolic blood pressure. SBP and DBP measurements via the inflating cuff were taken at the beginning and end of each phase. Headache intensify and srress. Subjects rated their headache intensity and perceived stress on scales from 0 to IO0 at the end of every second trial. Procedure Subjects were told that the experiment was a study of the development of pain problems and ways of dealing with them without drugs but rather by using the individual’s own resources. They were asked to sit in a comfortable armchair and electrodes and transducers were attached at the beginning of the session. The session was divided into a 6-min adaptation period followed by four IO-min phases. The session was further subdivided into trial and intertrial intervals each 30 set in duration. The beginning and end of trials were indicated by red and green lights, respectively. Subjects were asked to remain as still as possible during trial periods and to restrict their movements during the intertrial intervals. Phase I. This phase consisted of a baseline condition during which subjects were told to relax with their eyes open. Phase If. This phase consisted of the pain induction procedure for which a paradigm developed by Schweiger and Parducci [32J was employed. Subjects were told that an electric current stimulator would deliver a low voltage current through their foreheads. They were informed that the current would be too mild to be felt on the skin, but it had produced mild headaches in past studies, headaches which disappeared as soon as the current was turned OK. They were told that the procedure would be terminated if at any time they indicated that it was too uncomfortable. Subjects were also requested to carry out a perceptual task during this phase consisting of matching the brightness of two outer strips to a middle strip. The experimenter adjusted the middle strip to diRerent levels of brightness and subjects had to match these levels by adjusting a dial which controlled the brightness of the outer strips. Subjects were told that this task might either prevent headaches, reduce their intensity, or have no effect at all. At the beginning of this phase, subject5 were told that the experimenter would switch on the electric current stimulator. Leads were connected between the junction box into which the forehead EMG electrodes were attached and the current atimulatol-. Authenticity was added by operating switches which resulted in an illumination of a red light and a tone being emitted from the apparatus, but no current was passed through the electrodes. At the end of the phase, the switches on the current stimulator were turned off and the leads connecting the subject to the stimulator were disconnected. Following Schweiger and Parducci [321, this paradigm was administered in a ‘stressful’ context. Hence, the inevitably stressful appearance of a psychophysiology laboratory was further accentuated by placing on view anxiety-inducing equipment such as dentistry drills. Schweiger and Parducci reported that more than two thirds of a sample of 34 college students reported mild headaches in response to this paradigm. They argued against the reports simply reflecting compliance with perceived demand characteristics both on the grounds that their instructions did not suggest that the experimenter wanted them to report headaches, and the claims of subjects at the post-experimental interview that their headaches were ‘real’. Phase III. This phase was a recovery period. As in phase I. subjects were told to relax with their eyes open.
Mechanisms
of chronic
headaches
141
Phase IV. This phase was designed to result in pain reduction and used the elation condition of a mood induction procedure reviewed by Clark [331, Subjects were played a 7-min section of music (a jolly Overture of ‘Coppelia’ by Delibes) accompanied by instructions to try and get into the happy mood suggested by the music. They were told that the music would not automatically get them into a happy mood. and that they would have to try hard to get into the mood. They were instructed to use any method they wished providing it didn’t involve movement. This procedure was used because in a previous study of ours on the relationship between headaches and mood it resulted in a significant reduction in headache intensity [261.
RESULTS
P&n manipulation Mean headache intensity scores were calculated for each of the four phases and entered into a 4 x 4 repeated measures ANOVA. The analysis revealed that pain had been manipulated successfully as the main effect for Phases was significant (F = 4.04, df = 3,186, p
Research. Vol
36. No 2. pp
137-148.
1992. 0
0022-3999192 $5.00+.00 199: Pergamon Press plc
PSYCHOPHYSIOLOGICAL MECHANISMS OF CHRONIC HEADACHES: INVESTIGATION USING PAIN INDUCTION AND PAIN REDUCTION PROCEDURES P. R. MARTIN, G. V. MARIE and P. R. NATHAN (Received
3 January
1991; accepted
26 June 1991)
Abstract-Psychophysiological investigations of tension headaches and migraines have produced inconsistent and inconclusive findings. The study reported here aimed to further explore headache mechanisms by repeating earlier research with two main variations. The first pertained to recording and analytic procedures: a different approach was taken to measuring and interpreting cephalic vascular activity. The second was in the experimental conditions used: pain levels were manipulated so that the psychophysiology of headaches could be studied under conditions of pain increasing and decreasing. The findings provided no support for the role of muscle tension as a headache mechanism. Tentative support was obtained for vascular models of headaches. Headaches were associated with distension of the superficial temporal artery between pressure pulses. The data suggested that local vasomotor activity might play a role in dilating the temporal artery but that the effect seemed driven by elevated blood pressure and heart rate.
INTRODUCTION
TRADITIONAL beliefs concerning the pain mechanisms of the two common types of headaches are that tension headaches are caused by ‘sustained contraction of skeletal muscles’ and migraines are caused by ‘cranial arterial distension and dilatation’ Ill. Since the mid-1970s, over 100 studies have published data relevant to these hypothesized mechanisms [2] and many of the findings have been inconsistent with the classic conceptions (see Refs [3, 41 for reviews). For example, studies comparing tension headache sufferers with controls in terms of pericranial EMG have often failed to find significant differences between the resting levels of the two groups [5, 61 while others have [71. Several studies have found no significant differences in stress response between these two groups [8, 91 although some have 1101. Most studies have failed to differentiate the groups with respect to recovery from stress [ll, 121 but others have [131. Comparisons of migraineurs and controls at rest have resulted in a number of significant findings including abnormalities of temporal blood flow [ 14, 151, elevated heart rate 116, 111 and elevated diastolic blood pressure [161. Other studies have failed to find abnormalities on cardiovascular variables [9, 171. Distension of temporal arteries in migraineurs relative to controls has been demonstrated in response to noise [ 121, drinking ice water [ 181, and mental arithmetic [ 161, but other studies have been unable to differentiate migraineurs and controls in terms of vascular reactivity to stress [8, 111. One study reported significantly more vasoconstriction measured from the earlobe in a migraine group during post-stress recovery t51, but another study found no differences between migraineurs and controls on four cardiovascular
Department of Psychology, The University of Western Australia. Address for correspondence: Dr Paul R. Martin, Department of Psychology, Western Australia, Nedlands 6009, W.A., Australia. 137
The University
of
138
P. R. MARTINet al.
variables measured during post-stress recovery [ 191. Several studies have reported vascular abnormalities associated with tension headaches 120, 211, and other studies have reported EMG abnormalities associated with migraine [7, 221. It again needs emphasizing, however, that other studies have failed to find such differences [8, 91. While inconsistencies in the literature preclude drawing any conclusions with respect to headache mechanisms, the authors have argued elsewhere [3, 231 that vascular hypotheses look more promising than muscular hypotheses. A key problem for vascular models, however, lies in measuring the hypothesized mechanism of ‘cranial arterial distension and dilatation’. This has usually been accomplished via the indirect method of photoplethysmography . Photoplethysmography is based on the principle that light reflected from the tissue under the transducer is a function of local blood flow I241. A host of factors other than blood flow may affect the signal recorded by the polygraph, however, such as the precise position of the transducer relative to the artery, how the transducer is attached to the skin, skin characteristics, and technical parameters of the recording apparatus. Researchers have responded to these problems by using relative rather than absolute measures in their analyses 1251. Pulse amplitude is used as the unit of measurement for example, which consists of the amount of displacement in the polygraph signal caused by the passage of the pressure pulse along the artery. Also, some researchers analyse changes in pulse amplitude from baseline recordings rather than absolute pulse amplitudes. Some experts on plethysmography have argued that a consequence of the methodology being an indirect measure and one which is affected by a number of variables is that plethysmography results cannot be compared between laboratories, individuals or even with individuals on different occasions [241. If this argument is accepted then the approach has serious limitations for investigating headache mechanisms as most vascular hypotheses pertain to differences between headache sufferers and nonheadache controls, or between individuals when suffering from headaches and when headache-free. Also, using pulse amplitude as the unit of measurement results in a confounding of two variables: distension of the artery between pressure pulses and distension of the artery during the passage of pressure pulses. The relationship between arterial diameter and pulse amplitude depends on when distension occurs: distension during the passage of the pulse but not between pulses would result in an increase in pulse amplitude; equal distension during and between pulses would result in no change in pulse amplitude; and distension between pulses but not during passage of the pulses would result in a decrease in pulse amplitude. Another problem with the approach commonly used is that it provides no information about mechanisms of change in vascular activity. Headache researchers have frequently interpreted observed changes in pulse amplitude as indicating peripheral sympathetic vasomotor activity. Such an inference is not justified as observed changes may arise from alterations elsewhere in the cardiovascular system such as blood pressure and heart rate. Put simply. an artery may be distended as a result of relaxation of local smooth muscle or elevated blood pressure. It is for this reason that experts in the measurement of peripheral vascular activity recommend that these other cardiovascular variables should also be measured 1241. The approach taken here was to use a photoelectric transducer and to sample the amplified output rapidly with a laboratory computer. This would enable indirect
139
Mechanisms of chronic headaches measurement
of the maximum,
minimum
and average
distension
of the artery
to be
stored for later analysis. Rather than calculating difference scores between the maximum and minimum values which would approximate the usual pulse amplitude measure, the measures would be analysed separately to avoid confounding. This study also included continuous measures of blood pressure and heart rate so that the mechanisms of vascular changes could be explored. A final variation in this study was in the experimental conditions employed.
Instead of comparing headache sufferers and controls before, during and after stress, this study aimed to manipulate head pain while monitoring psychophysiological processes that might med;ate the pain. METHOD Subjects
Headache sub_jects were recruited via newspaper articles. Inclusion criteria for the study were: (a) headache frequency of at least one a week; (b) headache chronicity of at least 12 months; (c) diagnosed by family doctor as suffering from tension headache, migraine or combined headache; and average
(d) aged 19-59 yr. Control subjects were recruited by asking headache subjects to bring along a friend of same sex and similar age (? 3 yr) who does not experience regular headaches (an average frequency of less than one per month). This procedure was followed because it results in controls who are matched for age and sex, but also tends to result in matching on other variables such as socioeconomic status and education. Headaches were diagnosed as in our previous studies [26, 271. The approach used was a slightly modified version of one developed by Philips 1281 and is based on three migrainous features similar to those used by many neurologists [291: (a) unilateral onset of headache; (b) accompanied by nausea and/or vomiting; and (c) sensory prodromata. Subjects responded to questions about each of these features using a five-point scale (never, rarely, sometimes, usually, always). The diagnosis ‘migraine’ was applied if patients responded ‘always’ or ‘usually’ to two or three of the three questions (or if they responded ‘always’ or ‘usually’ to ‘accompanied by vomiting’); ‘tension’ was applied if they responded ‘never’ or ‘rarely’ to two or three of the three questions; and ‘mixed’ was applied in all other cases. A total of 5.5 headache and I4 control subjects satisfied the inclusion criteria. Subject characteristics are presented in Table I. TABLE
N
Group
Migraine Tension Mixed Control *Standard
19 23 I3 I4 deviations
I.--SUBJECT CHARACTERISTICS
Sex ratio (M : F) 2 : I7 9 : 14 4:9 4 : 10 included
Mean age
(yr) 39.3 37.6 39.1 33.1
(10.6)* (IO. I) (9.5) (9.5)
Mean headache chronicity (yr) 16.6 (11.7) 16.6 (11.1) 20.9 (12.1) -
in parentheses.
Apparutus Physiological activity was recorded with a Grass model 7D polygraph and a Cromemco Z-2D computer. Electromyographic (EMG) activity was monitored with three electrodes placed over the frontalis muscle two and a half centimetres above the eyebrows. Raw EMG up to a frequency of 500 HZ was sampled every millisecond, enabling average EMG levels to be calculated. Electrocardiographic (ECG) activity was also monitored every millisecond, electrodes being placed at the top of the sternum. at the fifth left intercostal space, and at the left elbow (ground). The radial artery pulse was continuously monitored by an electromechanical surface transducer [301 whose output was also sampled at 1 KHZ, enabling calculation of pulse transit time (radial R-wave to pulse interval, RPI). Cerebral arterial activity
was monitored with a photoelectric pulse sensor (Grass model PPS) placed over the zygomatico-facial branch of the superficial temporal artery. Absolute SBP and DBP measurements were taken during the session with a Copal model U-215 electronic sphygmomanometer. All sessions
were carried
out in a temperature-controlled
(2O”C), sound-attenuated,
psychophysiology
140
P. R. MARTIN et al
laboratory. The laboratory was divided door and a one-way mirror.
into subject
and experimenter
areas by a wall incorporating
a
A total of IO variables were recorded: eight physiological (six continuous variables using the polygraph and two periodic measures); and two subjective, PM~~, P~/il; crnd P,~Y. P,,x corresponds to the peak of the temporal arterial pulse signal and ih an index of the maximum distension of the artery; PU~N corresponds to the pre-pulse minimum level of the pulse signal and is an index of the minimum distension of the artery; PAv is an average of the samples taken of the pulse signal. Radial RPI. This is the interval between the occurrence of the ECG R-wave and the arrival of the associated pressure pulse at a point in the radial artery. Radial RPI is inversely correlated with systolic blood pressure [311, and to a lesser extent with diastolic blood pressure. Interbeat interval. Interbeat interval (IBI) is the time in millisec between successive ECG R-waves and is therefore an inverse of heart rate. Individual IBIS were averaged over 30.set intervals. Forehead EMG. Averages were computed based on samples of raw EMG taken throughout the session. Each of the above variables was averaged over 30.set intervals for the purpose of statistical analysis. S~stofic and diastolic blood pressure. SBP and DBP measurements via the inflating cuff were taken at the beginning and end of each phase. Headache intensify and srress. Subjects rated their headache intensity and perceived stress on scales from 0 to IO0 at the end of every second trial. Procedure Subjects were told that the experiment was a study of the development of pain problems and ways of dealing with them without drugs but rather by using the individual’s own resources. They were asked to sit in a comfortable armchair and electrodes and transducers were attached at the beginning of the session. The session was divided into a 6-min adaptation period followed by four IO-min phases. The session was further subdivided into trial and intertrial intervals each 30 set in duration. The beginning and end of trials were indicated by red and green lights, respectively. Subjects were asked to remain as still as possible during trial periods and to restrict their movements during the intertrial intervals. Phase I. This phase consisted of a baseline condition during which subjects were told to relax with their eyes open. Phase If. This phase consisted of the pain induction procedure for which a paradigm developed by Schweiger and Parducci [32J was employed. Subjects were told that an electric current stimulator would deliver a low voltage current through their foreheads. They were informed that the current would be too mild to be felt on the skin, but it had produced mild headaches in past studies, headaches which disappeared as soon as the current was turned OK. They were told that the procedure would be terminated if at any time they indicated that it was too uncomfortable. Subjects were also requested to carry out a perceptual task during this phase consisting of matching the brightness of two outer strips to a middle strip. The experimenter adjusted the middle strip to diRerent levels of brightness and subjects had to match these levels by adjusting a dial which controlled the brightness of the outer strips. Subjects were told that this task might either prevent headaches, reduce their intensity, or have no effect at all. At the beginning of this phase, subject5 were told that the experimenter would switch on the electric current stimulator. Leads were connected between the junction box into which the forehead EMG electrodes were attached and the current atimulatol-. Authenticity was added by operating switches which resulted in an illumination of a red light and a tone being emitted from the apparatus, but no current was passed through the electrodes. At the end of the phase, the switches on the current stimulator were turned off and the leads connecting the subject to the stimulator were disconnected. Following Schweiger and Parducci [321, this paradigm was administered in a ‘stressful’ context. Hence, the inevitably stressful appearance of a psychophysiology laboratory was further accentuated by placing on view anxiety-inducing equipment such as dentistry drills. Schweiger and Parducci reported that more than two thirds of a sample of 34 college students reported mild headaches in response to this paradigm. They argued against the reports simply reflecting compliance with perceived demand characteristics both on the grounds that their instructions did not suggest that the experimenter wanted them to report headaches, and the claims of subjects at the post-experimental interview that their headaches were ‘real’. Phase III. This phase was a recovery period. As in phase I. subjects were told to relax with their eyes open.
Mechanisms
of chronic
headaches
141
Phase IV. This phase was designed to result in pain reduction and used the elation condition of a mood induction procedure reviewed by Clark [331, Subjects were played a 7-min section of music (a jolly Overture of ‘Coppelia’ by Delibes) accompanied by instructions to try and get into the happy mood suggested by the music. They were told that the music would not automatically get them into a happy mood. and that they would have to try hard to get into the mood. They were instructed to use any method they wished providing it didn’t involve movement. This procedure was used because in a previous study of ours on the relationship between headaches and mood it resulted in a significant reduction in headache intensity [261.
RESULTS
P&n manipulation Mean headache intensity scores were calculated for each of the four phases and entered into a 4 x 4 repeated measures ANOVA. The analysis revealed that pain had been manipulated successfully as the main effect for Phases was significant (F = 4.04, df = 3,186, p
