256 Headaches
the patient in early sessions of biofeedback but it has been
suggested that the therapist•s presence, particularly if overly
active or intrusive, can become a distraction and interfere with
the training (Borgeat, Hade, Larouche, & Bedwani, 1980).
Hence, biofeedback training is designed to be therapist-
guided in the initial phases, with an effort to move in the
direction of increased self-regulation on the part of the patient
as training proceeds.
For both types of biofeedback training described next, 8 to
16 sessions of training are usually provided, typically between
20 and 40 minutes in duration (or long enough for training
to be effective but brief enough to minimize the likelihood of
fatigue). Instead of a universal prescriptive for the length of
treatment, the number of sessions is more usefully determined
by the individual patient•s response to treatment. Training
may be discontinued when maximum bene“t has been
achieved, as in a signi“cant reduction in headache activity or
when the reduction in headache activity plateaus or stabilizes.
In some cases, a reduction of headache activity may not have
occurred. In these cases, it may be useful to determine whether
the patient has achieved suf“cient skill at physiological self-
regulation of the target response. If the patient has achieved
suf“cient skill and is able to apply these skills in real-life
settings but has not experienced a reduction in headache
activity, other treatment options may be indicated.
EMG biofeedbackis relatively straightforward and can be
performed both in the clinic and at home with portable de-
vices. The aim of EMG biofeedback training is to decrease
muscle tension (as evidenced by electrical activity) of the
frontal muscles of the forehead (e.g., Budzynski et al., 1973),
although other muscles may be targeted in a similar fashion if
these muscles appear to play an important role in the individ-
ual•s headache activity. To achieve these aims, patients are
encouraged to experiment with a variety of methods of phys-
iological self-control (such as relaxation exercises, imagery
exercises, or breathing exercises) while receiving feedback
about their performance via an EMG device. Often, the train-
ing portions of the biofeedback sessions proceed in brief in-
tervals of 1 to 5 minutes in length, interspersed with brief
pauses that provide an opportunity for rest periods and dis-
cussion with the therapist. Across sessions, patients are
encouraged to further increase and re“ne their self-regulatory
skills in this manner.
Thermal biofeedbackalso generally aims to increase phys-
iological self-regulation. Speci“cally, the aim is to increase
peripheral body temperature or a hand-warming response. To
achieve these aims, patients are encouraged to experiment
with a variety of methods of physiological self-control (such
as relaxation exercises, imagery exercises, or breathing exer-
cises) while receiving feedback about their “nger tempera-
ture. Relaxation may be induced by recall prior to start of
biofeedback session. Often, autogenic phrases or imagery are
used during thermal biofeedback training sessions as a means
of raising peripheral body temperature. An adaptation phase
and baseline period are often used to note baseline tempera-
ture, followed by training phases that proceed in short inter-
vals characterized by voluntary efforts to warm the hands.
Some have suggested that it may be bene“cial for patients to
achieve a certain criterion level during training (e.g., be able
to increase “nger temperature to a certain temperature value
within a speci“ed period or for a speci“ed length; Fahrion,
Norris, Green, Green, & Snarr, 1986). Although this makes
sense from a clinical perspective, there is minimal data to
support this notion.
The mechanisms of action for these therapies are not fully
clear, as the data suggest that the direction of change in EMG
level and “nger temperature and extent of physiological
control achieved are not predictive of outcome. Similarly,
comparisons of relaxation therapies and biofeedback inter-
ventions often “nd equivalence, suggesting that the ef fects
are not speci“c to the type of therapy employed but rather
due to nonspeci“c ef fects that may have an underlying
relaxation mechanism (Cohen, McArthur, & Rickles, 1980;
Primavera & Kaiser, 1992). It is possible that a generalized
relaxation response or physiological self-control is the com-
mon denominator and active ingredient in these therapies,
rather than the directional change in a speci“c physiological
process. Alternative explanations of the mechanism of action
of these therapies have included alteration of cognitive
and behavioral responses to stress and improved coping
(Andrasik & Holroyd, 1980b) and cognitive changes such as
an increased sense of perceived control and mastery (Cohen
et al., 1980). Cognitive changes that may underlie the effec-
tiveness of biofeedback may be mediated by performance
feedback that suggests •successŽ (Holroyd, Penzien, Hursey,
et al., 1984), allowing for increased perceptions of control
and mastery. In short, research into the psychophysiological
mechanisms of biofeedback has led to the suggestion that
cognitive factors may play an important role in the ef“cacy of
behavioral and physiological self-regulation interventions;
however, our understanding of these mechanisms remains
•rudimentaryŽ (Gauthier, Ivers, & Carrier, 1996).Cognitive Behavioral InterventionsThis type of therapy has been labeled variously as cognitive
behavior therapy, cognitive stress coping therapy, cognitive
therapy, stress management, or other terms. In addition
to the evidence from biofeedback studies that suggests that
cognitive factors play a role in the treatment of recurrent