Comparative incidence rates of mild adverse effects to transcranial magnetic stimulation

Abstract

Objectives:
Past research has largely neglected to investigate mild adverse effects (MAEs) to transcranial magnetic stimulation (TMS), including headache and nausea. Here we explored the relationship between MAEs, participant characteristics (age and gender) and protocol parameters, including mode of application, coil geometry, stimulated brain region, TMS frequency, TMS intensity, and active vs. sham stimulation.

Methods:
Data from 1270 standard post-monitoring forms was obtained from 113 healthy participants. Analyses aimed to identify the risk factors associated with MAE reports and specific symptoms.

Results:
The overall rate of MAEs across TMS sessions was ∼5%, with ∼78% of symptoms occurring post-session. Initial TMS sessions were followed by a higher MAE incidence rate relative to later testing sessions. No associations between participant characteristics, TMS frequency, or intensity were observed.

Conclusions:
TMS-related MAEs are relatively common and may be exacerbated by initial expectations or anxieties of participants. A significant proportion of MAEs may reflect reporting of coincidental phenomena that are unrelated to TMS. Recommendations for future safety studies are proposed and monitoring documentation is provided.

Significance:
Our findings illustrate the importance of standardized monitoring of MAEs. Such research aids our understanding of how MAEs arise and may lead to interventions for reducing their incidence.

Full text

Comparative incidence rates of mild adverse effects to transcranial
magnetic stimulation
Leah Maizey
a,
⇑
, Christopher P.G. Allen
a
, Martynas Dervinis
a
, Frederick Verbruggen
b
, Alice Varnava
a,c
,
Michail Kozlov
a
, Rachel C. Adams
a
, Mark Stokes
d
, Jane Klemen
a
, Andreas Bungert
a
, Charles A. Hounsell
a
,
Christopher D. Chambers
a
a
Cardiff University Brain Research Imaging Centre, School of Psychology, Cardiff University, Cardiff CF10 3AT, UK
b
Psychology, College of Life and Environmental Sciences, University of Exeter, Exeter EX4 4QG, UK
c
Department of Psychology, Swansea University, Swansea SA2 8PP, UK
d
Oxford Centre for Human Brain Activity, University of Oxford, Oxford OX3 7JX, UK
article info
Article history:
Accepted 25 July 2012
Available online xxxx
Keywords:
Transcranial magnetic stimulation
Safety
Mild adverse effects
Post-monitoring
Risk factors
highlights
We report the rate of mild adverse effects (MAEs) to transcranial magnetic stimulation (TMS) over a
four-year period at Cardiff University, across 1270 experimental sessions in healthy participants.
Subsequent to both sham and active TMS sessions, we found an overall MAE rate of 5%; with the onset
of 78% occurring after participants had left the laboratory.
Additional analyses indicated that 37% of MAEs reported may be associated with expectations or anx-
ieties regarding TMS in naïve participants; routine monitoring of MAEs is recommended and screening
documentation is provided.
abstract
Objectives: Past research has largely neglected to investigate mild adverse effects (MAEs) to transcranial
magnetic stimulation (TMS), including headache and nausea. Here we explored the relationship between
MAEs, participant characteristics (age and gender) and protocol parameters, including mode of applica-
tion, coil geometry, stimulated brain region, TMS frequency, TMS intensity, and active vs. sham stimula-
tion.
Methods: Data from 1270 standard post-monitoring forms was obtained from 113 healthy participants.
Analyses aimed to identify the risk factors associated with MAE reports and specific symptoms.
Results: The overall rate of MAEs across TMS sessions was 5%, with 78% of symptoms occurring post-
session. Initial TMS sessions were followed by a higher MAE incidence rate relative to later testing ses-
sions. No associations between participant characteristics, TMS frequency, or intensity were observed.
Conclusions: TMS-related MAEs are relatively common and may be exacerbated by initial expectations or
anxieties of participants. A significant proportion of MAEs may reflect reporting of coincidental phenom-
ena that are unrelated to TMS. Recommendations for future safety studies are proposed and monitoring
documentation is provided.
Significance: Our findings illustrate the importance of standardized monitoring of MAEs. Such research
aids our understanding of how MAEs arise and may lead to interventions for reducing their incidence.
Ó2012 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights
reserved.
1. Introduction
Through modulation of cortical activity, transcranial magnetic
stimulation (TMS) has become an invaluable tool in experimental
and clinical neuroscience (Rossi et al., 2009). Due to the relatively
non-invasive nature of TMS, it has proven a favourable brain
stimulation technique over many of its predecessors (e.g. electro-
convulsive therapy: Pascual-Leone et al., 1993; Loo et al., 2008;
Janicak et al., 2008). However, TMS is not without medical risks.
Guidelines aimed at reducing the incident rates of the most severe
known risk, TMS-induced seizure, have received careful attention
(see Wassermann, 1998 and Rossi et al., 2009, for overviews).
1388-2457/$36.00 Ó2012 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.clinph.2012.07.024
⇑
Corresponding author. Tel.: +44 2920874000x77499.
E-mail address: leahmaizey@gmail.com (L. Maizey).
Clinical Neurophysiology xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

However, relatively little research has considered risk factors for
more mild adverse effects (MAEs), including headache and nausea.
This is particularly true for those MAEs that occur in the hours fol-
lowing the application of TMS, after the participant has left the lab-
oratory. Investigating the origins of such risk factors may open
avenues for reducing their incidence.
TMS exploits the principle of electromagnetic induction to pro-
duce small electrical currents in the cortex beneath the scalp site of
stimulation (Wagner et al., 2009). The application of TMS not only
generates electrical currents in brain tissue, but also in the inter-
vening muscle and nerve fibres in the scalp (Rossi et al., 2009). This
ancillary activation is the likely cause of tension headaches, local
pain and peripheral muscle twitches (Pascual-Leone et al., 1993;
Wassermann, 1998; Rossi et al., 2009). Local pain and headache,
in particular, are the most common MAEs previously reported,
affecting between 23% (Machii et al., 2006) and 40% of participants
who receive repetitive (r)TMS (Rossi et al., 2009; Oberman et al.,
2011). In addition, some reports suggest that symptoms may
diminish with successive sessions (Machii et al., 2006; O’Reardon
et al., 2007; Janicak et al., 2008). Here we further explore risk fac-
tors that have been implicated in previous research as being caus-
ally or theoretically related to MAEs, including the mode of TMS
application, intensity and frequency of stimulation, site of stimula-
tion, coil geometry, auditory and tactile artefacts, and active vs.
sham stimulation.
1.1. Mode of TMS application
Different neuronal effects can be achieved by administering
TMS in various modes (Jahanshahi et al., 1997; Rossi et al., 2009).
One approach involves the delivery of a single TMS pulse at times
relative to stimulus onset (Amassian et al., 1989). Research into
MAEs associated with single-pulse stimulation is limited, yet its
application is considered to be relatively harmless (Jahanshahi
et al., 1997). Indeed, the most recent international safety guide-
lines for TMS (Rossi et al., 2009) merely note that neck pain, tooth-
ache and paresthesia (tingling sensation or numbness of the skin)
are possible with single-pulse stimulation.
In contrast, greater attention has been paid to those protocols
associated with a potentially elevated seizure risk, including rTMS
and patterned theta-burst stimulation (TBS) (Machii et al., 2006).
Both rTMS and TBS involve multiple TMS pulses in rapid succession
to increase the effectiveness and duration of changes in cortical
excitability (Jahanshahi et al., 1997; Machii et al., 2006). Low dis-
continuation rates (primarily owing to MAEs) following rTMS are
reported, with approximately 4.5% of all participants excluded
from further participation (O’Reardon et al., 2007; Janicak et al.,
2008). Overall crude risk of MAEs after TBS has been estimated
at 5% (4.8% for healthy participants) with a crude risk per session
of 1.1% (Oberman et al., 2011).
1.2. Intensity and frequency
Incidents of headache and local pain have been positively corre-
lated with both the intensity of stimulator output and frequency of
stimulation (Wassermann, 1998; Loo et al., 2008; Rossi et al.,
2009
1
). Headaches have been reported after single-pulse TMS when
suprathreshold intensities were administered (i.e. >100% motor
threshold (MT): Rossi et al., 2009). Machii et al. (2006) speculate that
the incidence of MAEs may be related to whether intensity of stim-
ulation is set according to motor or phosphene thresholds (PT),
noting that a greater stimulator output is often required to elicit
phosphenes relative to motor responses.
1.3. Site of TMS application
Dense coverage of muscle nerve endings towards the front of
the scalp may explain the greater incidence of headaches and local
pain following frontal TMS in comparison to more medial or pos-
terior sites (Wassermann, 1998; Machii et al., 2006; Loo et al.,
2008). Frontal stimulation has also been associated with dental
pain, due to aggravation of the trigeminal nerve (Ropohl et al.,
2004). However, MAEs may also be induced by stimulation of pos-
terior cortical sites. Satow et al. (2002) found incidents of nausea
subsequent to right cerebellar stimulation in two of eight partici-
pants who were administered low frequency rTMS, possibly owing
to inadvertent stimulation of the posterior fossa. Neck pain has
also been reported after TMS application to such posterior sites
where the neck muscles can inadvertently be stimulated (Satow
et al., 2002); however such effects could also be due to participants
sustaining a constant head and neck position throughout an exper-
imental session (Machii et al., 2006).
1.4. Coil geometry
The shape and size of the coil used to administer TMS pulses has
a direct influence on the spread of the induced electric field and the
depth to which stimulation is possible, both in the cortex and on
the scalp surface (Wagner et al., 2009). Theoretically, one might
anticipate coil geometry to have a direct impact on MAEs, yet Rossi
et al. (2009) argue that this is unlikely. However, the authors also
recognise that no studies have tested this claim.
1.5. Auditory and tactile artefacts
The auditory artefacts of TMS increase with the intensity of
stimulator output and can exceed 140 dB (Counter et al., 1990;
Rossi et al., 2009). Increases in auditory threshold or feelings of
‘fullness in the ears’ have been reported in instances where ear
protection has not been used (e.g. Pascual-Leone et al., 1993). At
the same time, the tactile artefact associated with peripheral nerve
stimulation during TMS has been associated with paresthesia. As
expected, this appears to be more frequent with rTMS as opposed
to single-pulse protocols (Rossi et al., 2009).
1.6. Active vs. sham stimulation
Sham stimulation is often employed as a control condition in
TMS studies. During sham, an active coil is oriented on the scalp
in a way that produces little to no tactile effects but replicates the
auditory artefact (Loo et al., 2000). Few reports overtly describe
the details of the sham methods utilised. According to Lisanby
et al. (2001), coil orientation for sham stimulation is critical in
determining the density of the magnetic flux reaching the cortex.
This complication would therefore also apply to muscle and nerve
stimulation within the scalp tissue. MAEs have been reported after
sham stimulation but are thought to be directly related to the ori-
entation of the coil to the scalp surface (Loo et al., 2000). In a review
of the efficacy of rTMS in depression, Loo et al. (2008) concluded
that 16% of patients reported headache after sham stimulation
and 15% reported pain and/or discomfort (compared to 28% and
39% for active stimulation, respectively). However, the authors note
that there was a decrease in these incidences associated with great-
er angular displacement from the scalp surface. Indeed, Lisanby
et al. (2001) demonstrated that an active figure-8 coil oriented
90°to the scalp surface, with one wing touching the scalp, can
1
Although see Machii et al. (2006) who report a greater incidence of symptoms in
studies where TMS was applied at 61 Hz in comparison to >1 Hz to non-motor areas.
The authors attribute this to longer train durations in the use of lower frequencies.
2L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

reduce the energy reaching the cortex by 67–73% in comparison to
an active coil placed tangentially on the scalp.
1.7. Limitations of previous research
As noted by Machii et al. (2006) and Oberman et al. (2011),
there is a general lack of overt reports of MAEs within the TMS lit-
erature. Incident rates of MAEs may be underestimated as atten-
tion is understandably focused on more serious adverse effects,
including TMS-induced seizure. Studies that have explored MAEs
tend to focus on one parameter at a time (e.g. mode of TMS
application, see Oberman et al., 2011), rather than exploring the
potential interactions between different parameters. Furthermore,
incidents of adverse effects may have been missed as reports
generally include only those symptoms that have occurred during
a TMS session, ignoring side effects that may have a later onset. To
enable precise causal inferences regarding the origin of all adverse
effects, standardized and rigorous monitoring of side effects is
required (Machii et al., 2006; Oberman et al., 2011).
An additional concern is that the majority of literature reviews
focus on clinical populations, either alone or in conjunction with
reports from non-clinical populations (e.g. Loo et al., 2008; Ober-
man et al., 2011). Causation of adverse symptoms may therefore
be clouded by potential neurological deficits or the effects of med-
ication as opposed to the effects of TMS per se. By focusing solely
on non-clinical populations the baseline risk of MAEs may be
uncovered.
1.8. The current study
Here we present data collected via standard post-monitoring
questionnaires, from TMS studies undertaken over a four-year per-
iod at the Cardiff University Brain Research Imaging Centre (CUB-
RIC). In contrast to previous studies, we adopted a comparative
approach to explore the incident rates across a range of participant
characteristics and protocol parameters. The aim of this systematic
analysis was to uncover the TMS-related factors that most clearly
predict MAEs. Furthermore, we not only studied symptoms re-
ported during a session, but those that occurred in the 24-h period
following TMS.
2. Methods
2.1. Data inclusion and TMS parameters
Incidences of MAEs were documented for all TMS studies at
CUBRIC between 2008 and 2012, inclusive. All experiments were
undertaken for investigative as opposed to clinical purposes; the
exploration of TMS-related MAEs was undertaken post hoc and
was not the primary objective of these studies. All experiments
were approved by the local research ethics committee at the
School of Psychology, Cardiff University. In total, post-monitoring
forms for 1270 TMS sessions were included, across twenty-two
studies.
Table 1
The protocol parameters associated with each study. Two forms of TBS were administered: continuous and intermittent (600 pulses in 40 s and 600 pulses over 3 min,
respectively as described in Huang et al., 2005). For the purposes of assessing the association of TBS with MAEs, these protocols were collapsed into a single ‘TBS’ condition.
*
Study Mode Site Coil Duration Abs. Intensity %MT/%PT
**
Freq. nSessions Sham Spacer CT
1 Double r-aIPS, r-pIPS Fig-8 1 h+ 36–83% 120%MT 10 23 45 No N/A No
2 Single V1 (calcarine) Circ 1 h+ 47–70% 120%PT N/A 5 42 Same Yes No
3 Sing + doub V1 (calcarine) Circ 1 h+ 70% 120%PT 25 1 12 Same No No
4 Sing + doub Occ (pole), Circ 1 h+ 25–70% 95%PT 25 22 85 Same No No
(+single alone) V1 (calcarine)
5 cTBS r-IFG, r-IFJ Fig-8 40s 23–46% 59–80%MT 50 11 37 Sep Yes Yes
6 rTMS r-AG Fig-8 1 h+ 37–86% 110–160%MT 25 31 68 Sep No Yes
7 cTBS r-DLPFC, r-IFG Fig-8 40s 25–45% 65–80%MT 50 9 27 Sep Yes Yes
8 iTBS l-DLPFC Circ 3mins 21–49% 49–82%MT 50 32 79 Sep Yes Yes
9 cTBS AG Fig-8 40s 17–58% 80%MT 50 25 75 Sep Yes No
(left and right)
10 cTBS r-aIPS, r-pIPS, r-FEF Fig-8 40s 19–50% 70–80%MT 50 23 135 Sep Yes Yes
11 Single r-FEF, r-IFG, r-PPC Fig-8 1 h+ 36–70% 90–120%MT N/A 19 49 Same No Yes
12 cTBS/rTMS l-IFG Fig-8 40s/ 1 h+ 19–89% 80–140%MT 5–50 19 49 Sep No Yes
13 Double Occ (pole) Fig-8 1 h+ 50–70% 93–120%MT 20 13 24 Same No Yes
14 c/iTBS V1 (calcarine) Circ 40s/3mins 32–42% 80%MT 50 7 9 Sep Yes No
15 c/iTBS Occ (pole), Circ 40s/3mins 26–50% 80%MT 50 26 101 Sep Yes No
V1 (calcarine)
16 cTBS r-IFG, r-IFJ, r-SMA Fig-8 40s 20–39% 51–80%MT 50 19 65 Sep Yes Yes
17 cTBS + Sing Occ (pole), V1 (calcarine) Circ 1 h+ 52–87% 80%MT 50 12 24 Same No No
18 Single r-IFG, r-IFJ Fig-8 1 h+ 34–45% 100–105%MT N/A 3 7 Sep No Yes
19 rTMS r-aIPS, r-pIPS, r-SMG Fig-8 1 h+ 45–65% 71–122%MT 10 8 19 Same Yes Yes
20 rTMS r-Parietal, r-SMA Fig-8 1 h+ 56–87% 142.5%MT 1 7 17 Sep Yes Yes
21 rTMS Occ (pole) Fig-8 1 h+ 22–75% 100–115%MT 10 13 26 N/A N/A No
22 Single M1/Occ Fig-8 1 h+ N/A N/A N/A 104 275
***
N/A N/A N/A
(left and right)
*
Nb. Mode = mode of TMS application; Double = double pulse; Single = single pulse; Sing + doub = single and double pulse; cTBS = continuous theta-burst stimulation;
iTBS = intermittent theta-burst stimulation; rTMS = repetitive transcranial magnetic stimulation; cTBS + Sing = continuous transcranial magnetic stimulation and single
pulse; c/iTBS = continuous or intermittent theta-burst stimulation; Site = site of TMS application; l- = left hemisphere; r- = right hemisphere; IFG = inferior frontal gyrus;
IFJ = inferior frontal junction; AG = angular gyrus; aIPS = anterior intraparietal sulcus; pIPS = posterior intraparietal sulcus; FEF = frontal eye field; PPC = posterior parietal
cortex; Occ = occipital; M1 = primary motor cortex; SMA = pre-supplementary motor area; SMG = supramarginal gyrus; Coil = coil geometry; Circ = circular coil; Fig-
8 = figure-8 coil; Duration = duration of time that TMS application spread over; 40s = 40 s; 3mins = 3minutes; 1 h+ = 1 h or longer; Abs. Intensity = absolute Intensity of TMS
expressed as a % of stimulator output; %MT/%PT = stimulator output, expressed as a percentage of motor or phosphene threshold;%MT = percent of motor thresh-
old;%PT = percent of phosphene threshold; Freq = frequency of stimulation; n = number of participants; Sessions = number of sessions; Sham = whether or not sham was
applied in a separate session to active stimulation; Same = sham applied in same session as active session; Sep = sham applied as a separate session to the active sessions;
Spacer = whether an acrylic plastic spacer was used between the coil and scalp surface during sham stimulation; CT = whether a comfort threshold was used to test the
comfort of the protocol prior to the testing sessions.
**
Values vary due to consideration of comfort thresholds (see Section 2.3c).
***
Number of sessions is higher than number of participants due to compilation of induction sessions with left and right hemisphere MTs.
L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 3
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

In total, six different modes of TMS were applied to seventeen
cortical sites. Either a circular coil (90 mm) or a figure-8 coil (50 or
70 mm) was used to administer TMS via a Magstim Super Rapid or
a Magstim Rapid
2
biphasic stimulator. All sessions were conducted
with single coils and all parameters were within the international
safety guidelines of Wassermann (1998) and Rossi et al. (2009)
(see Table 1 for all protocol parameters). All experimenters were
trained to a standard level to ensure consistency in both the admin-
istration of TMS procedures and adherence to safety guidelines.
For ‘active’ TMS sessions the coil was oriented tangential to the
scalp surface. All sham TMS sessions were conducted with the coil
oriented 90°to the scalp surface. For figure-8 coils, one of the coil
wings touched the scalp surface. To minimise direct cortical stimu-
lation while maintaining any contact artefact, a 10 mm acrylic plas-
tic spacer was positioned between the coil and scalp for more
powerful protocols (see Table 1)(Lisanby et al., 2001). Ear plugs
were provided for all sessions in line with previous recommenda-
tions (Pascual-Leone et al., 1993; Wassermann, 1998; Rossi et al.,
2009) and consecutive TMS sessions were separated by at least 24 h.
2.2. Participants
A total of 113 unique participants (69 female and 44 male) aged
between 18 and 41 were recruited (M= 25.32, SD = 4.82). All par-
ticipants were neurologically healthy and were screened for med-
ical contraindications to TMS. Specifically, no participant was
currently taking any neuroactive medication or had a history of
frequent or severe headaches or migraines, drug abuse, brain in-
jury or any other brain-related conditions (e.g. stroke or disease),
or had a family history of seizure and/or epilepsy, or had sustained
any head injury that had resulted in concussion or unconscious-
ness (see Appendix A in the Supplementary material for the spe-
cific screening questionnaire employed). Participants completed
the screening questionnaire during an induction interview with
the experimenter and prior to participating in any new TMS stud-
ies. All participants had normal or corrected-to-normal vision. Par-
ticipants took part in multiple studies (mean number of sessions
participated in = 11.24, SD = 11.46) and informed consent was re-
ceived prior to participation in each study. Immediately prior to
receiving TMS, participants were further screened for state-depen-
dent contraindications, including recent alcohol or recreational
drug use, fatigue, or excessive consumption of caffeine (see Appen-
dix B in the Supplementary material).
2.3. Recruitment protocol and intensity setting
Prior to each study, all participants took part in a standard
recruitment protocol involving:
(a) An initial induction session: potential risks were explained,
participants were screened (see above), and TMS was
administered (10 pulses in the vicinity of M1 at intensities
30–50% of stimulator output).
(b) An intensity-calibration session: either via a distance-adjusted
MT or PT (Stokes et al., 2005, 2007; Varnava et al., 2011). MT
was estimated using the observation of movement method
(Kozel et al., 2000; McConnell et al., 2001; Varnava et al.,
2011). TMS pulses were applied to the M1 region of the scalp
to produce overt contractions in the contralateral hand. MT
was defined as the intensity of stimulator output required
to produce five observable contractions for every 10 TMS
pulses, at the site where the most pronounced contraction
was observed (Stokes et al., 2005). PT was determined sim-
ilarly, with TMS pulses applied to the occipital cortex in
order to induce the perception of a phosphene (Franca
et al., 2006).
(c) Comfort threshold session: for more powerful protocols or
those involving TMS to frontal sites, a comfort threshold
was obtained a priori (see Table 1) to ensure that the desired
frequency and intensity of TMS was not uncomfortable for
the participant. Intensity was adjusted on a site-specific
basis and then matched between sites according to the low-
est comfortable intensity (e.g. see Verbruggen et al., 2010).
During comfort threshold acquisition, test pulses were
applied to the desired site and frequency for that protocol.
A staircase method was employed in which the stimulator
output was varied until the comfort threshold was reached;
this was either the target intensity required for that protocol
or the point where discomfort was reported. Critically, if dis-
comfort was reported before the target intensity was
reached, where possible the target intensity was lowered
(see Table 1), otherwise participants were excluded from
further participation in that particular study.
2.4. Post-monitoring
A standard post-monitoring (PM) form was provided to partic-
ipants after each TMS session to assess whether they had experi-
enced any adverse effects during the session or in the following
24-h. Participants were specifically asked whether they had expe-
rienced any incidence of seizure, fainting or collapse, dizziness,
nausea or vomiting, headache, muscular aches, muscle spasm or
twitches, insomnia, sensory problems, difficulties speaking or
understanding speech, lack of coordination, or slowness or impair-
ment of thought (see Appendix C in the Supplementary material).
Participants were also encouraged to document any MAEs they
had experienced other than those listed. Further information con-
cerning the nature of adverse symptoms (e.g. longevity or severity)
was documented where possible.
2.5. Statistical analyses
Analyses focused on whether specific participant characteristics
(gender and age) or protocol parameters (mode of TMS application,
site of stimulation, coil geometry, frequency, intensity, and active
vs. sham stimulation) were relevant risk factors for MAEs. Age of
participant at time of involvement was separated into one of two
categories: either above or below the mean. The categorisation of
mode of application, site of stimulation and coil geometry can be
found in Table 1. Duration of TMS was categorised as either short
duration (i.e. 3 min or less) or long duration (i.e. 1 h or more; see
Table 1). Intensity of stimulation was explored in respect to the
absolute percentage of stimulator output, or indexed as a percent-
age of MT or PT. If protocol intensities were based on MT or PT,
then intensities were categorised as either sub-threshold (i.e.
<100%MT or PT) or supra-threshold (i.e. P100%MT or PT). Where
active and sham conditions were present in the same session the
corresponding PM form was treated as ‘active’. Sham data was
not included in any ‘stimulation site’ analyses as coil positioning
may not be exact (either placed on an arbitrary basis or based on
an ‘average’ location where more than one cortical site was stimu-
lated in a protocol). A minimum alpha level of .05 was applied to all
statistical tests. Due to the categorical nature of the data, a mixture
of binomial and multinomial logistic regression analyses were em-
ployed, followed by non-parametric inferential statistics (chi-
square; and Fisher’s Exact tests where expected frequencies were
less than 5). Analyses were conducted at two levels:
(1) Analysis at the level of reports: data from all PM forms were
analysed to explore the risk factors (participant and protocol
parameters) across 1270 MAE reports. This level of analysis
focused on whether one or more MAEs were reported or
not, as opposed to the specific type of MAE.
4L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

(2) Analysis of specific MAEs: data from PM forms where adverse
symptoms were reported were analysed. This level of analy-
sis sought to explore whether specific parameters were asso-
ciated with specific side effects. To reduce statistical
limitations caused by limited spread through key regions
of the data set, adverse symptoms that were likely to co-
occur were categorically merged. These categories were:
headache;nausea (including dizziness, nausea and vomit-
ing); muscular problems (including muscular aches, spasms
and twitches); and other (including lack of co-ordination,
sensory problems, slowness or impairment of thought,
insomnia and any other symptoms not pre-categorised).
3. Results
3.1. Analysis at the level of reports
Table 2 presents the frequency and category of all reported
MAEs and the corresponding protocol details. No incidents of sei-
zure or auditory side-effects were observed. Of the 1270 PM forms
collated, 546 sessions were undertaken by male participants, and
724 by female participants. Table 3 shows the percentage use of
each parameter across all PM forms. MAEs were reported on
62 PM forms (4.88%; with 78 symptoms reported in total, see Sec-
tion 3.2). In total, 44 of the 113 participants who took part in these
experiments reported at least one MAE (39%). Even though the pro-
portion of participants that reported MAEs was substantial, exclu-
sion rates were modest. Only five participants were excluded from
further participation after an initial session, and 10 from later
experimental sessions. These exclusion rates indicate that the
majority of MAEs were very mild (e.g. a slight headache as opposed
to more severe sensory or muscular problems, see Table 2) and
participants expressed a clear intent to continue. MAE data relating
to all participants were retained in the analyses unless otherwise
specified.
3.1.1. Active vs. sham TMS
Although more MAEs were associated with active reports
(5.38%) than with sham reports (2.09%), preliminary analysis re-
vealed this difference to be of only marginal significance (
v
2
(1,
N= 1270) = 3.76, p= .052). Where active and sham sessions were
completed separately, all protocol parameters were matched, with
the obvious exception that there was no direct cortical stimulation
associated with sham sessions (see Table 1). By analysing active
and sham PM forms separately, insights into MAEs that are associ-
ated with TMS testing procedures rather than TMS per se may be
uncovered.
3.1.2. Active reports
Data relating to all participant (gender and age) and protocol
variables (mode of application, site of application, coil geometry
and duration of stimulation) for active TMS sessions were entered
into a backward stepwise binary logistic regression analysis to
determine whether any of these variables could predict whether
an MAE was reported. A significant model (
v
2
(5, N=1079) =
18.24, p= .003) was revealed, in which mode of TMS application
was the only significant contributor (Wald
v
2
(5, N=1079) =
16.77, p= .005; Nagelkerke’s R
2
=.049). This effect was driven solely
by TBS (b=3.81; Wald
v
2
(1, N=1079) = 13.44, p<.001). As shown
in Table 3, TBS accounted for a similar number of PM forms to sin-
gle-pulse applications, but the percentage of MAEs found within
each mode of application was lower for TBS than for other modes.
Therefore TBS appears to be an important predictor for reduced
occurrence of MAEs. Subsequent chi-square analyses confirmed a
significant association between all modes of TMS application and
MAEs (p= .003, Fisher), with a significantly greater percentage of
MAEs associated with single-pulse sessions relative to TBS sessions
Table 2
The frequency of the occurrence of each adverse symptom reported with relevant protocol details. For definitions see
*
below and
*
Nb. for Table 1.
Mode Site Coil Active Freq Intensity Duration Incidents per category Additional information for ‘other’ symptoms
Head Nausea Musc Other
Single M1 Fig-8 Active N/A N/A 1 h+ 11 5 5 8 Sensation in wrist/hand (4); insomnia (1); bruising
sensation/ pain at site of TMS(2); lack of
coordination (1)
Occ Fig-8 Active N/A N/A 1 h+ 4 2 2 Tingling body sensation(1); nosebleed (1)
Occ Circ Active N/A N/A 1 h+ 2 2 2 Sensory problems (1); slowness of thought (1)
FEF Fig-8 Active N/A 120%MT 1 h+ 1
IFG Fig-8 Active N/A 95–100%MT 1 h+ 2
PPC Fig-8 Active N/A 95%MT 1 h+ 1 2
Double aIPS Fig-8 Active 10 Hz 120%MT 1 h+ 1
pIPS Fig-8 Active 10 Hz 120%MT 1 h+ 1 Jawache (1)
cTBS + Sing Occ Circ Active 50 Hz 80%MT 1 h+ 1
rTMS AG Fig-8 Active 25 Hz 160%MT 1 h+ 3
aIPS Fig-8 Active 10 Hz 119%MT 1 h+ 1 Lack of co-ordination (1)
Occ Fig-8 Active 10 Hz 100–115%MT 1 h+ 3
Sing + doub V1 Circ Active 25 Hz 95%PT 1 h+ 3 1
c/iTBS aIPS Fig-8 Active 50 Hz 80%MT 3 min max 1
FEF Fig-8 Active 50 Hz 80%MT 3 min max 1
IFG Fig-8 Active 50 Hz 80%MT 3 min max 1
IFJ Fig-8 Active 50 Hz 62% MT 3 min max 1 Tingling sensations on left side of body (1)
V1 Circ Active 50 Hz 80%MT 3 min max 1 1 Temporary speeded thoughts/actions (1)
Occ Circ Active 50 Hz 80%MT 3 min max 3
DLPFC Circ Active 50 Hz 78–80%MT 3 min max 1 1 Tiredness post-session (1)
Sham Circ Sham 50 Hz 80%MT 3 min max 2
Sham Fig-8 Sham 50 Hz 25–80%MT 3 min max 1 1 Insomnia (1)
Total number of each adverse symptom 33 19 8 18
*
Nb. Head = headache; Musc = muscular problems; 3 min max = maximum of 3 minutes of stimulation.
L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 5
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

(9% and 3%, respectively;
v
2
(1, N= 777) = 15.27, p< .001). No other
significant association between adverse reports and mode of
application was found.
The reduced occurrence of MAEs associated with TBS may be
due to the typically low absolute stimulator intensities at which
these protocols are applied (see Table 1). Intensity and frequency
data were substituted into the regression analysis in place of mode
of application (gender, age, site of application, coil geometry and
duration of stimulation remained). To circumvent potential issues
of multiple collinearity between these variables (r=.62,
p< .001), separate regression analyses were carried out including
frequency and then intensity. No significant regression models
were observed. Due to initial induction and intensity-calibration
sessions, many of the single-pulse protocols yielded little frequency
and intensity information (see Section 2.3 and study 22, Table 1).
Further exploration of the data suggested that the higher incidence
of MAEs associated with single-pulse sessions was the result of all
participants completing an initial single-pulse session prior to fur-
ther experimental participation (i.e. an induction session, or ses-
sions involving the acquisition of MT or PTs
2
). Indeed, significantly
more MAE reports were associated with initial single-pulse sessions
(40%) as opposed to later active TMS sessions (12%;
v
2
(1,
N=1079) = 33.97, p< .001). MAEs reported after later experimental
protocols may be confounded by the omission of five participants
who had experienced an MAE of sufficient severity to warrant com-
plete exclusion; this effectively prevented these participants from
contributing further to incidence rates for MAEs. However, removal
of these participants from the data had no demonstrable effect on
the regression model (model
v
2
(5, N= 1072) = 48.35, p= .001, Nage-
lkerke’s R
2
= .14; mode of application: Wald
v
2
(5, N= 1072) = 13.69,
p= .02; TBS: b=3.63; Wald
v
2
(1, N= 1072) = 10.97, p= .001) or sub-
sequent association analyses (mode of application: p=.02, Fisher;
single-pulse vs. TBS:
v
2
(1, N= 770) = 11.48, p= .001), with a greater
number of MAEs still associated with single-pulse sessions (8%)
relative to TBS sessions (3%).
This greater proportion of MAEs within initial sessions may
reflect a particular sensitivity to MAEs in naïve participants. To test
this possibility, the regression analyses were repeated with all ini-
tial sessions excluded. The entry of all of the participant or protocol
variables (gender, age, mode of application, site of application, coil
geometry and duration of stimulation) led to a non-significant
regression model and no significant associations between mode
of application and MAEs (p= .52, Fisher).
Differences in the overall duration of TMS between single-pulse
and TBS application, or the preclusion of potential MAEs by exclud-
ing participants who do not pass a comfort threshold, do not appear
to be adequate alternative explanations for these findings. Although
the duration of stimulation for TBS protocols was shorter than that
for single-pulse protocols (see Table 1), duration of stimulation
was not shown to be a significant predictor of MAEs in any of the
regression analyses. In addition, the effects of comfort threshold
did not account for differences in MAEs between single-pulse TMS
and TBS. Although excluding those participants that did not pass a
comfort threshold may, in effect, prevent them from contributing
to MAEs in that study, no difference in the associations were found
between MAEs and whether or not a comfort threshold was con-
ducted (
v
2
(1,N= 804) = .78, p= .38
3
). These results further indicate
that incidence of MAEs was greater following initial single-pulse ses-
sions compared with later experimental sessions, which may indeed
reflect an initial sensitivity to MAEs in naïve participants.
The application of TMS at supra-threshold intensities (i.e.
P100%MT or PT) may account for why there is such a high inci-
dence of MAEs associated with initial single-pulse sessions.
Supra-threshold stimulator intensities are likely to have been
applied during intensity-calibration sessions (see Section 2.3b).
Although the specific intensities used in these intensity-calibration
sessions was not recorded, analyses revealed no significant associ-
ation between MAE incidence and whether the intensities used in
later testing sessions were set above or below MT (p= .83, Fisher)
or PT (p= .25, Fisher), or matched to MT or PT (p= .79, Fisher). These
results indicate that variance in stimulator output relative to the
excitability threshold is not reliably predictive of MAEs, but that
MAE reports are, once again, more likely to be driven by the initial
sensitivity to TMS in naïve participants.
This premise is further illustrated through analyses regarding
MAEs and cortical site of stimulation. Although site of stimulation
was not a significant predictor of MAEs in any of the regression anal-
yses, further exploration revealed a significantly higher rate of MAEs
associated with the threshold sites of M1 and occipital cortex com-
bined (10%) vs. all other sites (3%;
v
2
(1, N= 1079) = 21.60, p< .001),
for M1 alone (10%) vs. all other sites (4%;
v
2
(1,N= 1079) = 13.48,
p< .001), and for occipital stimulation alone (9%) vs. all other sites
(5%;
v
2
(1,N= 1079) = 4.25, p= .04). Yet no significant difference
was found between M1 and occipital sites and their association with
MAEs (10% and 9%, respectively;
v
2
(1,N= 378) = .08, p= .77).
3.1.3. Sham reports
All analyses completed with active reports were repeated with
sham reports only. There was no significant association between
incidence of MAEs and any of the participant variables (gender or
age) or protocol variables (mode of application, site of application,
coil geometry and duration of stimulation, frequency or absolute
intensity; p> .37, Fisher, for all analyses). Of the 191 PM forms
Table 3
The percentage use and the corresponding percentage of MAEs within each mode of TMS application, each site of stimulation and each coil geometry. Values are shown separately
for both active and sham stimulation. For definitions see
*
below and
*
Nb. for Table 1.
Mode Site
**
Coil
Single TBS rTMS Sing + doub Double cTBS + Sing Post Frontal M1 Occ Sham Fig-8 Circ
% Use
Active 36.24 35.77 12.14 7.23 6.4 2.22 35.19 19.69 18.66 11.42 15.04 73.68 26.32
Sham 0.52 90.58 8.9 N/A N/A N/A N/A N/A N/A N/A N/A 64.4 35.6
% MAEs
Active 9.21 2.59 4.58 3.85 2.9 4.17 2.91 3.2 10.13 8.97 2.09 5.79 4.23
Sham 0 2.31 0 N/A N/A N/A N/A N/A N/A N/A N/A 1.63 2.94
*
Nb. Post = cortical sites located posterior to M1; Frontal = cortical sites located anterior to M1.
**
For efficiency of presentation, cortical sites are divided into Frontal areas (located anterior to the motor cortex); and Posterior areas (located posterior to the motor cortex).
M1 and occipital stimulation are shown separately as these sites were stimulated most frequently in initial single-pulse sessions. Due to the arbitrary nature of the
positioning of the coil for sham stimulation, these sessions are shown as a separate cortical ‘site’.
2
For all analyses initial sessions are inclusive of induction session and participant’s
initial MT or PT.
3
All inductions, MTs and PTs were excluded from these analyses, as CT sessions
would be based on the intensities calibrated as a result of these sessions.
6L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

completed subsequent to sham stimulation, MAEs were only
reported on 4, and all corresponded to TBS administration (see
Table 2). Similar rates of MAEs were associated with active TBS ses-
sions (2.59%) and sham TBS sessions (2.09%). There was no signif-
icant difference in the association of active vs. sham TBS and MAEs
(p> .99, Fisher), demonstrating that the incidence of these MAEs
may have been coincidental or due to factors unrelated to direct
cortical stimulation.
3.2. Analysis of specific MAEs
When considered at the level of reports (via PM forms), the inci-
dent rates of MAEs are likely to be underestimated. This is due to
the exclusion of participants who report MAEs early in the se-
quence of experimental sessions and the repeated sampling of
those participants who never report MAEs. Therefore we tested
whether specific symptoms were associated with particular partic-
ipant characteristics (gender or age) or protocol parameters (mode
of application, site of application, coil geometry and duration of
stimulation, frequency or absolute intensity; see Section 2.5 for
classification of adverse symptoms).
This level of the analysis included 44 participants. Across all re-
ports, participants completed between 1 and 41 sessions each
(M=11.32, SD =11.07). Analyses were based on 78 symptoms re-
ported in total, owing to multiple symptoms reported on 12 of the
62 adverse PM forms. Of the 44 participants who had experienced
MAEs, 42 had experienced these following active TMS rather than
sham. Eleven participants reported further MAEs after subsequent
protocols, accounting for 25 of 62 sessions that provoked MAEs
(mean number of sessions with MAEs = 3.18, SD =1.40), implying
a predisposition for MAEs in some participants. Indeed, MAEs were
reported after 20% of all sessions for those experiencing multiple
MAEs, compared to 10% for participants who reported MAEs after
a single session. This difference was reliable (
v
2
(1,N= 498) = 8.87,
p= .003) even though there was no significant difference in the
number of sessions completed by each group (t(42) = 1.46, p= .68).
Removal of the initial sessions from the data still indicated that
there was a general predisposition in some participants to experi-
ence MAEs (
v
2
(1,N= 449) = 12.82, p< .001). For those participants
who reported symptoms after more than 1 session, 18% of
completed sessions now included MAEs, compared with 6% for
participants who reported MAEs after a single session. For those par-
ticipants that experienced MAEs, more reports were associated with
single-pulse (82%) sessions as opposed to experimental sessions
(9%;
v
2
(1,N= 498) = 40.12, p< .001).
There was no difference in the distribution of MAEs across cat-
egories reported during initial sessions vs. experimental protocols
(p= .79, Fisher).
3.2.1. Active vs. sham
Analysis of active vs. sham data did not reveal any significant
difference in associations with each category of MAE (p= .61, Fish-
er). Symptoms reported subsequent to sham stimulation were all
reported by participants who had experienced multiple symptoms.
To reduce statistical limitations caused by limited spread through
key regions of the data set (as sham data was based on only 4 po-
sitive PM forms), analyses were carried out both exclusive and
inclusive of sham data.
3.2.2. Analysis of specific MAEs excluding sham TMS
Fig. 1 shows the overall percentages of each category of adverse
symptoms reported. Due to the infrequent use of each protocol
parameter across studies at this level, only gender and age could
be included in a multinomial regression analysis, neither of which
reliably predicted the reported MAE category (Model
v
2
(6,N=
74) = 7.23, p=.301; Nagelkerke’s R
2
= .10). No significant
associations between type of MAE reported and protocol variables
was found (mode of TMS application, p= .34, Fisher; or coil geome-
try, p= .89, Fisher). There was no difference in the category of MAE
associated with initial or later sessions (p= .66, Fisher).
Since these studies included stimulation of a wide range of cor-
tical regions, we also explored the potential association between
specific symptoms and cortical sites based on previous literature
(e.g. Wassermann, 1998; Satow et al., 2002; Machii et al., 2006;
Loo et al., 2008). In contrast to previous expectations, incidence
of headache was not significantly associated with frontal stimula-
tion
4
compared with all other symptoms and sites (
v
2
(1,N= 74) =
.51, p= .48) However, in accordance with Satow et al. (2002) nausea
was more likely to be associated with occipital stimulation com-
pared with all other symptoms and sites (
v
2
(1,N= 74) = 4.54,
p= .03); nausea was reported in 43% of MAEs associated with occip-
ital stimulation, compared with 19% for other sites. Circular coils
were more frequently used in protocols targeting occipital regions
but coil geometry was not associated with nausea symptoms in com-
parison to other symptoms (p=.36,Fisher).
3.2.3. Analysis of specific MAEs including sham TMS
All analyses completed exclusive of sham data were repeated
with sham data included. No significant predictors of, or associa-
tions with, any specific MAE category were observed (p> .59,
Fisher, for all analyses).
3.3. Onset of specific MAEs
Of all 78 symptoms reported (over both active and sham ses-
sions), 14 (18%) were reported to be present during the session
only, and 61 (78%) reported only in the 24 h after TMS. The remain-
ing symptoms of nausea, dizziness and headache were reported
subsequent to a single TMS session and occurred both during the
session and in the following 24 h.
4. Discussion
The aim of this study was to provide comparative incidence
rates for a range of MAEs associated with TMS. The overall inci-
dence rate of MAEs across sessions is comparable with previous re-
views of TBS protocols, converging on 5% (e.g. Oberman et al.,
2011). Although this figure may represent an underestimation of
TMS-induced MAEs (see Results 3.2), the incidence rates across
participants (39%) are comparable to previous reports (e.g. Machii
et al. (2006) report that 40% + of participants experience MAEs
Fig. 1. Percentages of each category of MAE reported across all post-monitoring
forms (including both active and sham data).
4
Here, frontal sites were classified as those cortical sites anterior to and inclusive
of the motor cortex.
L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 7
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

after rTMS). In line with previous findings, headaches were found
to be the most common MAE (e.g. Machii et al., 2006; Loo et al.,
2008; Rossi et al., 2009; Oberman et al., 2011). In contrast to pre-
vious safety studies, we undertook a comparative post hoc ap-
proach in which the effects of participant and protocol
parameters were considered. Previous studies and literature re-
views have either reported only the quantity and type of incidents,
or have explored risk factors associated with one mode of TMS
application (e.g. Machii et al., 2006; Oberman et al., 2011). Addi-
tionally, this study was completed based on data collected from
neurologically healthy participants, in contrast to previous studies
that also included data from clinical populations (e.g. Machii et al.,
2006; Oberman et al., 2011).
Our results differ somewhat from previous observations that
MAEs were more likely to occur after rTMS or TBS than after sin-
gle-pulse TMS (e.g. see Rossi et al., 2009). The majority of MAEs ob-
served here were associated with single-pulse stimulation, as
opposed to alternative protocols previously reported as carrying
a higher risk (Machii et al., 2006). Our findings suggest that the
greater frequency of MAEs associated with single-pulse sessions
may be due to increased sensitivity in naïve participants to MAEs
prior to initial TMS sessions (i.e. induction, MT or PT sessions) as
opposed to subsequent experimental protocols. Data including ini-
tial single-pulse sessions indicated that TMS of M1 and occipital
cortex (MT and PT sites), considered either separately and in con-
junction, were associated with a greater incidence of MAEs com-
pared with all other sites. However, no such difference was
found between M1 and other cortical sites during later experimen-
tal sessions.
These findings cannot be explained by duration of stimulation
during a session, or comfort threshold exclusions (i.e. the exclusion
of participants from a study if they found the desired frequency
and intensity of TMS uncomfortable; see Section 2.3c), as these fac-
tors were assessed in our analyses. Instead it seems likely that
naïve participants may have expectations or anxieties about the
sensations and subsequent side effects of TMS, leading to relatively
higher reporting of coincidental phenomena. Past research has
indicated that the incidence of MAEs may be reduced with succes-
sive testing sessions (Machii et al., 2006; O’Reardon et al., 2007;
Janicak et al., 2008). This may well be due to the reduction in anx-
ieties and expectations regarding TMS application over time.
It is notable that some MAEs were reported subsequent to sham
stimulation. The magnetic flux reaching the scalp and cortex for
sham conditions presented here is negligible due to the coil orien-
tation implemented and the use of acrylic plastic spacers for more
powerful protocols (see Lisanby et al., 2001). It appears that inci-
dence of sham-related MAEs for TBS sessions may be coincidental
or related to TMS-evoked anxiety or other non-specific stressors
present in experimental situations. To unearth whether MAEs are
due to physiological factors or the reporting of coincidental phe-
nomena, future research could explore MAEs between different
experimental settings (e.g. involving EEG or MRI). If TMS-evoked
anxiety is the origin of some MAEs, then future research would ben-
efit from personality profiling through anxiety-related measures
prior to participation. Although some clinical safety studies report
the use of psychometric questionnaires (e.g. Loo et al., 2008), there
has been no direct exploration of whether MAEs correlate with such
measures.
Although the risk of MAEs is higher in clinical populations
(Oberman et al., 2011), it is unclear whether this is due to the
expression of anxiety-related traits and differential expectations
of TMS in clinical groups/settings, or the underlying pathophysiol-
ogy of the disease-state and concurrent use of medication (Machii
et al., 2006; Loo et al., 2008). The inclusion of personality measures,
and comparative research based on clinical and non-clinical
populations, could help untangle the contribution of these factors.
Aside from the effects of initial session sensitivity, analyses
indicated that MAE incidence rates found in subsequent experi-
mental protocols were more likely to be associated with occipital
stimulation compared with other cortical sites. This observation
is consistent with previous findings by Satow et al. (2002) and sug-
gests that occipital stimulation could lead to activation of the pos-
terior fossa, resulting in nausea symptoms. Importantly, comfort
thresholds were not used in any of our experimental protocols tar-
geting occipital cortex. The screening of participants using comfort
thresholds prior to such applications would be expected to
enhance participant comfort and may provide a sufficient interven-
tion to reduce or eliminate such MAEs.
Our study presents a post-monitoring methodology to TMS
safety research. While this comparative approach assesses the rate
of MAEs across a range of participant and protocol parameters, it is
important to note that our findings emerged from post hoc analyses
within studies where uncovering the basis of MAEs was not the
main objective. By grouping different protocols, the analyses are
compiled across varying stimulation approaches. The conclusions
we make regarding the influence of specific parameters (e.g. fre-
quency, intensity, etc.) on TMS-related MAEs are therefore, to an ex-
tent, tentative; particularly when considering the application of
rTMS and TBS. Before concrete inferences can be made, it is impor-
tant that variations in these parameters are explored directly. For a
more complete understanding of the complex nature of TMS-re-
lated MAEs, future studies could be specifically designed for that
purpose using factorial a priori designs (e.g. Satow et al., 2002).
In conclusion, it is reasonable to expect the focus of research to
be centred on TMS protocols associated with the most severe ad-
verse effects, including seizure. However, exploration of protocols
associated with more common MAEs should not be undervalued
when ensuring participant comfort and safety. This study high-
lights the importance of monitoring MAEs to TMS. Moreover, not
all MAEs occurred during a TMS session; instead, the onset of most
symptoms was reported post-session (78% of MAE reports). The
advantages of standard post-monitoring are manifold and we rec-
ommend that post-screening be adopted broadly in TMS studies.
Participant responses to follow-up questionnaires may be more
informative when probing whether any of a range of specific symp-
toms was experienced after the previous TMS session as opposed
to only during the session. Standardized post-monitoring across
all sessions will thus enhance the inferences made in future com-
parative studies and meta-analyses. In accordance with the recom-
mendations by Machii et al. (2006) and Oberman et al. (2011),
documentation of specific participant characteristics and protocol
parameters will enrich understanding of the origin of MAEs and
may lead to interventions for reducing their likelihood.
Acknowledgements
These studies were supported by the Biotechnology and Biolog-
ical Sciences Research Council (BB/E020291/1, CDC & JK; David
Phillips Fellowship, CDC), the Wales Institute of Cognitive Neuro-
science (CPGA, JK, AV, CDC), Research Foundation-Flanders (FV),
and the Welsh Assembly Government (Academia for Business
grant HE 07 COL 3012, CDC & AB). We are grateful to Jacky Boivin
and the School of Psychology Ethics Committee for inspiring this
research.
Appendix A. Supplementary data
Supplementary information associated with this article, includ-
ing TMS screening and post-monitoring forms, can be found in the
online version, at http://dx.doi.org/10.1016/j.clinph.2012.07.024.
8L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024

References
Amassian VE, Cracco RQ, Maccabee PJ, Cracco JB, Rudell A, Eberle L. Suppression of
visual perception by magnetic coil stimulation of human occipital cortex.
Electroencephalogr Clin Neurophysiol 1989;74:458–62.
Counter AS, Borg E, Lofqvist L, Engr B, Brismar T. Hearing loss from the acoustic
artefact of the coil used in extracranial magnetic stimulation. Neurology
1990;40:1159–62.
Franca M, Koch G, Mochizuki H, Huang Y, Rothwell JC. Effects of theta burst
stimulation protocols on phosphene threshold. Clin Neurophysiol
2006;117:1808–13.
Huang Y-Z, Edwards MJ, Rounis E, Bhatia KP, Rothwell JC. Theta burst stimulation
report of the human motor cortex. Neuron 2005;45:201–6.
Jahanshahi M, Ridding MC, Limousin P, Profice P, Fogel W, Dressler D, et al. Rapid
rate transcranial magnetic stimulation – a safety study. Electroencephalogr Clin
Neurophysiol 1997;105:422–9.
Janicak PG, O’Reardon JP, Sampson SM, Husain MM, Lisanby SH, Rado JT, et al.
Transcranial magnetic stimulation in the treatment of major depressive
disorder: a comprehensive summary of safety experience from acute
exposure, extended exposure, and during reinforcement treatment. J Clin
Psychiatry 2008;69:222–32.
Kozel FA, Nahas Z, DeBrux C, Molloy M, Lorberbaum JP, Bohning DE, et al. How coil-
cortex distance relates to age, motor threshold, and antidepressant response to
repetitive transcranial magnetic stimulation. J Neuropsychiatry Clin Neurosci
2000;12:376–84.
Lisanby SH, Gutman D, Luber B, Shroeder C, Sackeim HA, Sham TMS. Intracerebral
measurement of the induced electrical field and the induction of motor-evoked
potentials. Biol Psychiatry 2001;49:460–3.
Loo CK, McFarquhar TF, Mitchell PB. A review of the safety of repetitive transcranial
magnetic stimulation as a clinical treatment for depression. Int J
Neuropsychopharmacol 2008;11:131–47.
Loo CK, Taylor JL, Gandevia SC, McDarmont BN, Mitchell PB, Sachdev PS.
Transcranial magnetic stimulation (TMS) in controlled treatment studies: are
some ‘‘sham’’ forms active? Biol Psychiatry 2000;47:325–31.
Machii K, Cohen D, Ramos-Estebanez C, Pascual-Leone A. Safety of rTMS to non-
motor cortical areas in healthy participants and patients. Clin Neurophysiol
2006;117:455–71.
McConnell KA, Nahas Z, Shastri A, Lorberbaum JP, Kozel FA, Bohning DE, et al. The
transcranial magnetic stimulation motor threshold depends on the distance
from coil to underlying cortex: a replication in healthy adults comparing two
methods of assessing distance to cortex. Biol Psychiatry 2001;49:454–9.
Oberman L, Edwards D, Eldaief M, Pascual-Leone A. Safety of theta burst
transcranial magnetic stimulation: a systematic review of the literature. Clin
Neurophysiol 2011;28:67–74.
O’Reardon JP, Solvason HB, Janicak PG, Sampson S, Isenberg KE, Nahas Z, et al.
Efficacy and safety of transcranial magnetic stimulation in the acute treatment
of major depression: a multisite randomized controlled trial. Biol Psychiatry
2007;62:1208–16.
Pascual-Leone A, Houser CM, Reese K, Shotland LI, Grafman J, Sato S, et al. Safety of
rapid-rate transcranial magnetic stimulation in normal volunteers.
Electroencephalogr Clin Neurophysiol 1993;89:120–30.
Ropohl A, Hiller M, Elstner S, Sperling W, Kornhuber J, Bickel A. Dental pain during
repetitive transcranial magnetic stimulation. Eur Psychiatry 2004;19:457–8.
Rossi S, Hallett M, Rossini PM, Pascual-Leone A. Safety, ethical considerations, and
application guidelines for the use of transcranial magnetic stimulation in
clinical practice and research. Clin Neurophysiol 2009;120:2008–39.
Satow T, Mima T, Hara H, Oga T, Ikeda A, Hashimoto N, et al. Nausea as a
complication of low-frequency repetitive transcranial magnetic stimulation of
the posterior fossa. Clin Neurophysiol 2002;113:1441–3.
Stokes MG, Chambers CD, Gould IC, Henderson TR, Janko NE, Allen NB, et al. Simple
metric for scaling motor threshold based on scalp-cortex distance. application
using transcranial magnetic stimulation. J Neurophysiol 2005;94:4520–7.
Stokes MG, Chambers CD, Gould IC, English T, McNaught E, McDonald O, et al.
Distance-adjusted motor threshold for transcranial magnetic stimulation. Clin
Neurophysiol 2007;118:1617–25.
Varnava A, Stokes MG, Chambers CD. Reliability of the ‘observation of movement’
method for determining motor threshold using transcranial magnetic
stimulation. J Clin Neurosci Methods 2011;201:327–32.
Verbruggen F, Aron AR, Stevens M, Chambers CD. Theta burst stimulation
dissociates attention and action updating in human inferior frontal cortex.
PNAS 2010;107:13966–71.
Wagner T, Rushmore J, Eden U, Valero-Cabre A. Biophysical foundations underlying
TMS: setting the stage for an effective use of neurostimulation in the cognitive
neurosciences. Cortex 2009;45:1025–34.
Wassermann EM. Risk and safety of repetitive transcranial magnetic stimulation:
report and suggested guidelines from the international workshop on the safety
of repetitive transcranial magnetic stimulation, June 5–7, 1996.
Electroencephalogr Clin Neurophysiol 1998;108:1–16.
L. Maizey et al. / Clinical Neurophysiology xxx (2012) xxx–xxx 9
Please cite this article in press as: Maizey L et al. . Clin Neurophysiol (2012), http://dx.doi.org/10.1016/j.clinph.2012.07.024