Transcranial Magnetic Stimulation And Its
Effectiveness In Affective Disorders
Thomas AM Kramer, MD
Transcranial Magnetic Stimulation in Humans
Transcranial magnetic stimulation (TMS) was patented in 1903 for the treatment of
depression. Research investigating this treatment has exploded recently, as evidenced by the
fact that only 1 paper on TMS was published in 1985 whereas 192 were published in 1997.
This technique involves running electrical current through a coil, which creates a magnetic
field. This magnetic field can penetrate the skull and reach cortical tissue to a depth of
approximately 2 cm. It is speculated that this magnetic field causes a biologic current in the
neurons that it makes contact with. There is an enormous variability from patient to patient
as to how powerful this magnetic field should be. Placing the coil over the motor strip in
area 4 of the brain and increasing the power until the thumb on the contralateral side moves
determines the calibration. This is referred to as the motor-evoked potential or motor
threshold, and the treatment dose is usually given at a range of 80% to 100% of this.
Exercise usually increases this response, and depression and chronic fatigue syndrome have
been noted to decrease it in a state-dependent fashion.
Many variables have yet to be determined in treatment with TMS.
One of these is the frequency with which the treatment is applied,
measured in cycles per second. There has been some evidence that
high-frequency treatment (20 cycles per second) can create an
excitable state in the patients and that low-frequency treatments (1
cycle per second) can be "quenching," causing long-term potentiation
of depression. There is also some controversy about which
hemisphere of the brain the treatment should be applied to. While the
treatment is almost always applied to the prefrontal cortex, there has
been some evidence that treatment of the left side decreases happy
mood and treatment of the right side decreases sad mood. In a
study of 1-cycle-per-second treatment of both the left and right side
in 19 normal subjects, the authors found absolutely no difference in
any mood, cognitive, or physiologic variable at this low-frequency
treatment, except a statistically significant difference in diastolic blood
pressure (subjects treated on the left side had a decrease of 9 mm
Hg, and no blood pressure effect was seen with right-sided
treatment). The authors doubt whether this finding is of any
Ebmeier and colleagues did a study randomizing 15 patients with
major depression into 1 of 3 groups: 5 cycles per second, 10 cycles
per second, or 20 cycles per second. Patients were treated twice per
day with 500 stimulations for a total of 1000 stimulations per day for 5
days. They achieved a reduction of 44% on Hamilton depression
scales, with no significant differences in treatment effects across the
3 groups. They then looked at the motor-evoked potentials both
before and after treatment and found no significant changes.
Finally, this group of authors did some neuroimaging studies using
single photon emission computed tomography (SPECT) both before
and after TMS treatments and found the primary change to be an
increase in activity in the medial prefrontal cortex.
ECT and TMS
The similarities and differences between TMS and electroconvulsive therapy (ECT) offer
some interesting perepectives on treatment. Clinical similarities between the 2 treatments
include their apparent effectiveness in the treatment of depression, mania, and Parkinson's
disease. Unlike TMS, ECT requires administration of anesthesia and induction of a seizure
and usually causes some memory impairment, but its efficacy is extremely well documented
in the literature. The first controlled trial of TMS was done in 1986, and there have been
fewer than 10 known controlled trials since then, so its efficacy has not been as well
Biologic similarities between the 2 treatments include an
anticonvulsant effect for both, increased apomorphine stereotypy in
animal models, and downregulation of beta-adrenergic receptors.
Although both treatments precipitate early gene expression in the
brain, they do so in different areas. With ECT, early gene expression
occurs most often in the hippocampus, whereas with TMS, early
gene expression occurs primarily in the paraventricular nucleus of the
Recent studies have shown new growth of neurons in the brain
throughout life. These authors have demonstrated that ECT causes
new neuronal growth in the hippocampus, and that this new growth
increases with the number of ECT treatments. This growth is also
sustained; it is not subject to apoptosis (programmed cell death).
These findings are consistent with current research findings that
show that chronically depressed patients have decreased brain
volume, particularly in the hippocampus. Finally, these authors
speculate that TMS may be effective by resetting circadian dysthymia
by the growth of new sympathetic neurons from the paraventricular
nucleus to the pineal gland.
TMS -- Treatment or Research Tool
Recent efforts by George to combine functional MRI scans with TMS have presented new
and interesting challenges. Since both use magnetism at fairly high power, this presented an
engineering challenge. One important difference is that the MRI scanner is essentially
always on, but the TMS coil is only activated for brief pulses of energy. These authors were
able to develop the equipment so that the image and the pulse magnetism would alternate,
allowing them to image what is going on in the brain while TMS is taking place. One of the
first things they looked at is the motor threshold calibration phenomenon. It is the distance
between the cortex and the coil that determines the electricity required to move the thumb.
They determined this energy to be approximately 60% of what is needed, with the patient's
physiology providing the other 40%. They could find no difference between volitional and
TMS thumb movement on a functional scan. They found that TMS-induced movements
involve a very focused response of the neurons involved, and not a "blob of tissue" being
activated by the coil. They also speculated that contralateral movement of the other hand
would dampen the effect of the TMS-induced movements, since this is a physiologic
phenomenon of attenuation of electrical activity in the contralateral side of the brain, where
movement is taking place. To their surprise, they found no such effect. They were also
curious as to whether repeated pulses of TMS would have a plateau effect on motor
threshold; the 2 studies they conducted to resolve this issue yielded contradictory results.
Finally, they found some increases in blood flow in a dose-dependent fashion at 120% of
the motor threshold.
It is important to review what is and is not known about TMS. There is
a newfound acceptance of the idea that changes in mood can occur
with subseizure stimulation. With only 2 cm or less of direct effect, we
are unable to access deep brain structures, but there may be a
"prefrontal window" to allow us to get reasonably close.
In the future, further combinations of imaging technology and TMS
may allow us real-time focusing on specific brain structures during
treatment. These combinations may also allow us to look at specific
neurotransmitters during treatment. A study by George on efficacy
of TMS showed a difference of only 5 points on the Hamilton
depression scale. This result is of interest from a research standpoint
but is less than clinically relevant. Looking at all of the controlled
studies of TMS together, he draws the following conclusions:
1. The speed of onset is a minimum of 2 weeks.
2. The location seems to be the prefrontal areas of the skull. There is some
contradiction in the laterality data, but there is some evidence for better effect on
the right side.
3. The treatments seem to have more of a direct antidepressant effect than a stimulant
4. Although there are little data indicating that TMS is effective in bipolar depression,
data have shown that it is not particularly effective in older populations, psychotic
patients, or those patients whose anatomy makes for a distance of more than 16
mm between their cortex and the coil during treatments (ie, a thick skull).
We have no definitive data on the best frequency, ideal hemisphere, or optimum power for
TMS. We also have no definitive safety data, although it is important to note that the
seizure risk with TMS is lower than that associated with any known antidepressant
medication. In conclusion, what we need are better clinical trials of longer duration to fully
assess the viability of this exciting new treatment.
Antidepressant Properties of ECT and TMS
It is interesting to note that there are antidepressant properties of both ECT and TMS. These
2 treatments clearly have different mechanisms of action and may have different clinical
utility. Although the effectiveness of ECT is extremely well established, TMS is, at best,
still being developed, but holds great promise for having significantly fewer side effects.
One of the most clinically important unanswered questions about this comparison is which
types of patients will benefit from which therapy.
A new approach to combining the properties of both of these
mechanisms is seen in the recent development of convulsive
magnetic stimulation treatment (MST) and an analysis of why such
treatment may be of clinical utility. Although the fact that TMS does
not involve seizure induction is seen as a great advantage, it is
believed that the seizure in ECT is what imparts a great deal of its
efficacy. There are some serious problems with convulsive therapy
initiated with electricity. There is remarkable variability of seizure
threshold in ECT, with recent data showing a 100-fold difference
between patients (Sackeim). The human skull is not particularly
even with regard to electrical conduction, and it has multiple "current
sinks." ECT has a dose-response relationship for both efficacy and
side effects, and regional changes in the brain are responsible for
both the effects of the treatment and the adverse events. Finally, it is
very difficult to stop the spread of electrical activity to vulnerable
areas of the brain.
The current controversy between unilateral vs bilateral ECT adds to
these problems. Unilateral ECT is thought of as having fewer side
effects but also less efficacy. Right unilateral ECT at high dose may
be as effective as bilateral ECT. Seizures, however, may lack
antidepressant efficacy but create enough of an insult to the brain for
adverse effects to occur. Current paths, and not intensity of seizure,
determine efficacy. The key to ECT efficacy is enhancing inhibition in
the prefrontal cortex, particularly on the right side. Bilateral ECT
tends to concentrate current in the anterior frontal cortex, while
unipolar ECT tends to be even over the anterior two thirds of the
cortex it is applied to. As the dose increases, unipolar ECT's current
distribution more closely resembles bilateral ECT. ECT responders
have been shown to have a state-dependent 80% increase in seizure
threshold, a decrease in cerebral blood flow in the anterior cortex, an
increase in delta waves in the frontal cortex, and an increase in theta
waves in the medial temporal cortex.
The comparison between ECT and MST may show MST solving
some of these problems. The pulse of energy used to give ECT is
between 1 and 2 milliseconds, which is supraphysiologic and
certainly inefficient. MST (and TMS) has pulses at 0.15 milliseconds
given at variable frequencies. MST may give us the opportunity to do
focal inductions of seizures with active inhibition of vulnerable areas.
Both left and right unilateral ECT works in many patients. Right
unilateral ECT may be better for depression, and left unilateral ECT
may be more effective for mania. Laterality may be more critical for
TMS, as may be frequency or the site of the treatment. ECT may
resemble low-frequency TMS more than high-frequency TMS in its
Repetitive Transcranial Magnetic Stimulation (rTMS) in the
Elderly depressed patients are often treatment resistant or intolerant of medication. Since
TMS may be effective for depression and has no known cognitive side effects, it may be a
good treatment for this population. Unfortunately, there are data showing a decreased
response rate to TMS in the elderly.
Mosimann and colleagues conducted a controlled double-blind
study of depressed patients over the age of 40 in which 9 patients
underwent sham TMS and 14 received real TMS. The patients in the
treatment group were given 9 days of treatment at 100% of motor
threshold at 20 cycles per second for 16,000 pulses per day. Sham
TMS was done by tilting the coil at 90 degrees away from the patient
during treatment. Patients were assessed for mood and cognitive
status on days 1 and 9, and for mood on day 5. The results of this
study showed a response rate to sham TMS of 17% and a response
rate to real TMS of 20.4%, with no significant difference. It did,
however, show a small improvement in cognitive function. While the
study was disappointing, it does indicate an interesting potential for
this treatment in the elderly, given its cognitive results.
1. Steele JD, Glabus MF, Shajahan PM, Ebmeier KP. Increased cortical inhibition in
depression: a prolonged silent period with transcranial magnetic stimulation.
Psychol Med. 2000;30:565-570.
2. Ebmeier KB, Shajahan PM, Glabus MF. Neuroimaging of TMS effects in humans.
Int J Neuropsychopharmacol. 2000;3:S26. Abstract: S.16.1.
3. Bolwig T. ECT and TMS - Similarities and dissimilarities in their development as
antidepressive treatment modality. Int J Neuropsychopharmacol. 2000;3(suppl
1):S27. Abstract: S.16.2.
4. George MS, Nahas Z, Avery D. TMS-treatment or research tool - where are we 5
years after the beginning? Int J Neuropsychopharmacol. 2000;3(suppl 1):S27.
5. George MS, Lisanby SH, Sackeim HA. Transcranial magnetic stimulation:
applications in psychiatry. Arch Gen Psychiatry. 1999;56:300-311.
6. Sackeim HA. Antidepressant properties of ECT and TMS: modulation of prefrontal
cortex function. Int J Neuropsychopharmacol. 2000;3(suppl 1):S27. Abstract
7. Mosimann UP, Schlaepfer TE. rTMS in the elderly - cognitive enhancer,
antidepressant or research tool? Int J Neuropsychopharmacol. 2000;3(suppl 1):S28.