Cravings, currents and cadavers: What is the magnitude of tDCS effects on food craving outcomes?

Full text

Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ynns20
Nutritional Neuroscience
An International Journal on Nutrition, Diet and Nervous System
ISSN: 1028-415X (Print) 1476-8305 (Online) Journal homepage: http://www.tandfonline.com/loi/ynns20
Cravings, currents and cadavers: What is the
magnitude of tDCS effects on food craving
outcomes?
Peter A. Hall & Cassandra J. Lowe
To cite this article: Peter A. Hall & Cassandra J. Lowe (2018): Cravings, currents and cadavers:
What is the magnitude of tDCS effects on food craving outcomes?, Nutritional Neuroscience, DOI:
10.1080/1028415X.2018.1513678
To link to this article: https://doi.org/10.1080/1028415X.2018.1513678
Published online: 28 Aug 2018.
Submit your article to this journal
Article views: 15
View Crossmark data

LETTER TO THE EDITOR
Cravings, currents and cadavers: What is the
magnitude of tDCS effects on food craving
outcomes?
Peter A. Hall
1
, Cassandra J. Lowe
2
1
School of Public Health & Health Systems, University of Waterloo, Canada,
2
Brain and Mind Institute, Western
University, Canada
Dear Editor,
We read with interest the recent meta-analytic
review by Mostafavi et al.
1
on the efficacy of transcra-
nial direct current stimulation (tDCS) for modulation
of food cravings and consumption, updating our meta-
analysis published one year prior.
2
The findings of
Mostafavi et al.
1
suggested highly reliable effects for
tDCS on both outcomes, with a magnitude in the
medium-to-large size across the studies included in
the review. The reported magnitude of tDCS effects
was more than double that reported in our earlier
meta-analysis, despite the modest time lag between
the two analyses. Such findings are somewhat surpris-
ing in light of the widely reported null effects of tDCS
on cognitive test performance reported in the existing
research literature.
3
In this commentary, we explore some of the poten-
tial methodological origins of Mostafavi et al.’s
1
find-
ings, and open a dialogue regarding some important
avenues for further refinement of tDCS and future
evaluations of its efficacy.
A tale of two methods
There are some important methodological differences
between the Mostafavi et al.
1
meta-analysis and our
original meta-analysis. Our original meta-analysis
reviewed all published and unpublished studies on
non-invasive imaging methods –including both
rTMS and tDCS –with food cravings and consump-
tion as outcomes. We conducted a combined analysis
with all neuromodulation methods, and separate sub-
analyses for tDCS and rTMS, in order to compare
the efficacy of these quite different approaches to
neuromodulation. Separate systematic reviews were
conducted for single-session experimental studies and
multi-session therapeutic trials, for reasons that we
felt were quite important: single doses of stimulation
in the experimental context quantify acute effects,
while multi-session trial outcome effects represent
any acute effects at the time of outcome assessment,
plus any cumulative effects across sessions. Indeed it
is the cumulative component of the effect that is
thought to be critical for any therapeutic application
of neuromodulation for eating-related clinical con-
ditions. Our meta-analysis focused on the single-
session effects, given that, at the time, there were not
sufficient numbers of therapeutic trials to conduct a
meta-analysis for treatment studies (of tDCS or
rTMS).
Estimating separate effect sizes for single-session
and multi-session tDCS is important for guiding
sample size determinations of experimental studies
(which normally employ single-session stimulation)
versus clinical treatment studies (which normally
employ multi-session stimulation). If single-session
effects are weaker than multi-session effects, using an
overall effect size estimate derived from a mix of
single and multi-session tDCS findings could lead to
insufficient a-priori sample size estimates for exper-
imental studies involving tDCS, and possibly unrealis-
tic expectations from clinical samples of the effects of a
single stimulation session.
In our primary single-session analyses, we observed
reliable effects of rTMS but less consistent support for
tDCS.
2
Recognizing the explosive growth in studies
involving both methodologies occurring in the year
since the original meta-analysis was submitted for
publication, we subsequently updated the meta-analy-
sis.
4
The conclusions of this updated analysis were
much the same, but we did find in this updated analy-
sis that tDCS was indeed effective for modulating crav-
ings, but with substantially less efficacy than rTMS. In
Correspondence to: Peter Hall, Prevention Neuroscience Laboratory,
School of Public Health & Health Systems, LHN 2731, University of
Waterloo, 200 University Avenue West, Waterloo, Ontario N2L 3G1,
Canada. Email: pahall@uwaterloo.ca
© 2018 Informa UK Limited, trading as Taylor & Franc is Group
DOI 10.1080/1028415X.2018.1513678 Nutritional Neuroscience 2018 1

both publications, we highlighted the fact that tDCS
was a methodology currently undergoing significant
refinement, and that further refinement of methods
may render its efficacy stronger. A good example of
this is the examination of the dosage issue in
Mostafavi et al.,
1
which clarifies the possible origin
of earlier null effects involving tDCS throughout the
literature, as well as the potential for tDCS to affect
cognitive mediators in a more consistent way with
2 mA applications.
A second important methodological difference
between our original meta-analysis and the
Mostafavi et al.
1
meta-analysis that needs to be
addressed is the inclusion of duplicate study samples.
In our meta-analysis, for both the main analyses and
method sub-analyses, the assumption of one effect
size per participant sample was persevered. The
inclusion of one effect size per participant sample is
an important methodological caveat in meta-analyses
(i.e. the assumption of independence of data points),
as the inclusion of data from duplicate participant
samples inflates the homogeneity of findings (see
Borenstein et al.,
6
). For all studies that included
more than one active stimulation condition within
the same set of participants (e.g. different stimulation
sites) we created a composite score by averaging the
effect sizes across conditions. However, in the
Mostafavi et al.
1
meta-analysis the authors treated
duplicate participant samples as independent data
points, which violates the assumption of independence
and raises the possibility that the observed effects were
less precise than would otherwise be the case.
Currents and cadavers
A recently published study provides some independent
support for the proposition that current strength
matters with tDCS. Using a very clever paradigm
involving living mice and human cadavers,
Vöröslakos et al.
7
showed that concerns of current
attenuation via conduction through the skull at the
1 mA current are well founded, both in the living
rodent models and also when testing current intensity
in brain tissues of the unfixed human cadavers using
EEG. The current attenuation by the skull and scalp
was estimated at 80% for the former, and 60% for
the latter. The significance of this study was in the
demonstration that the hypothesized attenuation of
current via bone and other tissues is a valid one and
that 2 mA stimulation intensities would likely be the
minimum intensity required to affect brain networks
in a meaningful way in living humans.
Additional developments have taken place in
relation to the relative focality of tDCS, using tech-
niques such as high definition tDCS, which will
likely be combined with the higher intensity to
produce more reliable experimental and clinical
effects in the future. Several early examples of such
studies have provided promising results (c.f.,
8
).
Overall, there is good reason to believe that 1 mA
stimulation is an insufficient intensity, and that more
focal application might be an important direction
forward for tDCS technologies.
More methods
The current intensity issue aside for the moment, we
took the step of applying our methodological
approach to the tDCS data available up to 10
August 2018; three new studies
9–11
since the publi-
cation of the update to our meta-analysis (see
4
). We
present these findings for the sake of comparability
to our original meta-analysis. Likewise, consistent
with our meta-analysis and subsequent update, all
effect sizes were transformed to Hedge’sg; this trans-
formation corrects for effect size inflation due to
small sample sizes. Effect sizes are reported as
Cohen’sdin the Mostafavi et al.
1
meta-analysis.
Our findings for single-session stimulation indicate
a marginally significant effect of tDCS on food crav-
ings (g=−.233, Z=−1.948, P=.051; k=12), and
a significant effect of tDCS on the consumption of
appetitive calorie-dense foods (g=−.430,
Z=−2.140, P=.032; k=10; Figures 1 and 2).
Current intensity was not a significant moderator of
the craving effects (Q
B
=2.967, P=.085). Of note,
only two studies used 1 mA current intensities
(g=−.063), and the effect size for studies using
2 mA current intensities is comparable to the observed
overall effect (g=−.289).
In conclusion, we find that there is indeed evidence
for the efficacy of single-session tDCS now that a criti-
cal mass of studies has been reached for inclusion in
meta-analytic reviews. However, the magnitude of
the overall effects is more similar to our original esti-
mate, and notably lower than that provided by the
recent analysis by Mostafavi et al.
1
Future avenues
Some important considerations in future tDCS
research will be a systematic examination of multiple
current intensity effects within the same study (i.e.
involving random assignment to 1 and 2 mA current
conditions within the same study, using the same
equipment and montage). Testing a priori hypotheses
of current strength effects on outcomes will be impor-
tant in order to rule out conflation of current intensity
with montage arrangement, equipment type, or
sample characteristics. Likewise, demonstration of
dose–response effects in such studies will help to deal
with expectancy effects that might drive comparisons
between inactive sham and experimental conditions
in some earlier studies; this is particularly important
for clinical trials involving multiple stimulation
Letter to the Editor
Nutritional Neuroscience 2018
2

sessions, which involve considerable opportunity for
participants to guess their assigned treatment
condition.
Finally, beyond efficacy, safety and tolerability of
higher current intensities will need to be established,
given that much of the existing research of this kind
involves 1 mA or 1 and 2 mA studies together. If inten-
sities higher than 2 mA are attempted, careful report-
ing of safety and tolerability parameters will need to
accompany measurement of the primary outcome.
Relatedly, a general issue in all neuromodulation
research is historically inconsistent reporting of toler-
ability and adverse event data, and a standardized
reporting avenue would assist in this respect, for
both tDCS and rTMS researchers, particularly
outside of the context of experimental trials in settings
where reporting of such events are not mandatory.
Continued attention to the measurement of
outcome (both craving and, even more so, consump-
tion) will be crucial for the validity of the findings to
be assured. Subtle expectancy effects can be conveyed
by cues in the eating environment, and the relative
palatability of the food options and the cleanliness of
the space may introduce unwanted floor effects on
consumption. Finally, affective dimensions of experi-
ence are crucial, as any contexts that may introduce
disgust reactions or shame could introduce more
noise into the assessment of brain system effects on
Figure 1 Forest plot for food cravings as a function of single-session tDCS.
Figure 2 Forest plot for food consumption as a function of single-session tDCS.
Letter to the Editor
Nutritional Neuroscience 2018 3

actual consumptive behavior in the laboratory setting,
or even in clinical trials where objective measurement
of eating takes place.
Conclusions
In conclusion, we suggest that the best single-session
effect size estimate for 2 mA tDCS on food cravings
is g=−.289 when targeting the dlPFC. This estimate
is similar in magnitude to the earlier estimate in our
original meta-analysis
2
and subsequent update.
4
This
estimate would be appropriate for calculating sample
size determinations for experiments involving tDCS
with food craving outcomes in human participants.
Multi-session tDCS effect sizes may prove to be
larger in magnitude, however further trial data –
including those findings extracted from the unpub-
lished literature –will need to be compiled in order
to calculate a reliable estimate of 2 mA application
of tDCS for reducing food cravings in the therapeutic
context. Currently, three studies have been published
with cravings outcomes
12–14
and four with consumption
outcomes.
10,12,13,15
However, only two of the above trials
employ 2 mA current stimulation intensities
10,14
and
only 3 among those measuring food consumption.
10,14,15
Attention to methodological considerations noted
above within the tDCS research literature will ensure
that continued technological innovation takes place
and is directed appropriately. We appreciate Mostafavi
et al.’s effort to bring organization to this rapidly evol-
ving field of inquiry and look forward to additional
findings involving tDCS from this and other research
groups. We encourage pre-registration of a systematic
review protocol in PRISMA or other platforms for
careful consideration of tDCS multi-session treatment
effects as new studies emerge in the published and
unpublished research literature. Mostafavi and col-
leagues would be well-positioned to lead this effort.
Funding
This work was supported by Social Sciences and
Humanities Research Council of Canada [435-2017-
0027].
ORCID
Cassandra J. Lowe http:// orcid.org/0000-0003-
3830-5283
References
1 Mostafavi SA, Khaleghi A, Mohammadi MR, Akhondzadeh S.
Is transcranial direct current stimulation an effective modality in
reducing food craving? A systematic review and meta-analysis.
Nutr Neurosci 2018.doi:10.1080/1028415X.2018.1470371
2 Lowe CJ, Vincent C, Hall PA. Effects of noninvasive brain
stimulation on food cravings and consumption: a meta-analytic
review. Psychosom Med 2017;79(1):2–13. doi:10.1097/PSY.
0000000000000368
3 Horvath JC, Forte JD, Carter O. Quantitative review finds no evi-
dence of cognitive effects in healthy populations from single-
session transcranial direct current stimulation (tDCS). Brain
Stimulation 2015;8(3):535–50. doi:10.1016/j.brs.2015.01.400
4 Hall PA, Lowe C, Vincent C. Brain stimulation effects on food
cravings and consumption: an update on Lowe et al. (2017)
and a response to Generoso et al. (2017). Psychosom Med
2017;79(7). doi:10.1097/PSY.0000000000000504
5 Borenstein M, Hedges LV, Higgins JPT, Rothstein, HR.
Introduction to meta-Analysis. Hoboken NJ: John Wiley &
Sons; 2011.
6 Lipsey MW, Wilson DB. Practical meta-analysis. Thousand
Oaks, CA: Sage Publications; 2001.
7 Vöröslakos M, Takeuchi Y, Brinyiczki K, Zombori T, Oliva A,
Fernández-Ruiz A, et al. Direct effects of transcranial electric
stimulation on brain circuits in rats and humans. Nat Commun
2018;9(1), 483. doi:10.1038/s41467-018-02928-3
8 Nikolin S, Loo CK, Bai S, Socrates D, Donel MM. Focalised
stimulation using high definition transcranial direct current
stimulation (HD-tDCS) to investigate declarative verbal learning
and memory functioning. Neuroimage 2015;117(15):11–19. doi:
10.1016/j.neuroimage.2015.05.019
9 Grundeis F, Brand C, Kumar S, Rullmann M, Mehnert J, Pleger
B. Non-invasive prefrontal/frontal brain stimulation is not effec-
tive in modulating food reappraisal abilities or calorie consump-
tion in obese females. Front Neurosci 2017;11:334. doi:10.3389/
fnins.2017.00334
10 Heinitz S, Reinhardt M, Piaggi P, Weise CM, Diaz E, Stinson EJ,
et al. Neuromodulation directed at the prefrontal cortex of sub-
jects with obesity reduces snack food intake and hunger in a ran-
domized trial. Am J Clin Nutr 2017;106(6):1347–57. doi:10.
3945/ajcn.117.158089
11 Ray MK, Sylvester MD, Osborn L, Helms J, Turan B, Burgess
EE, et al. The critical role of cognitive-based trait differences
in transcranial direct current stimulation (tDCS) suppression of
food craving and eating in frank obesity. Appetite 2017;116:
568–74. doi:10.1016/J.APPET.2017.05.046
12 Jauch-Chara K, Kistenmacher A, Herzog N, Schwarz M,
Schweiger U, Oltmanns KM. Repetitive electric brain stimu-
lation reduces food intake in humans. Am J Clin Nutr
2014;100(4):1003–09. doi:10.3945/ajcn.113.075481
13 Leong SL, De Ridder D, Vanneste S, Sutherland W, Ross S,
Manning, P. High definition transcranial pink noise stimulation
of anterior cingulate cortex on food craving: an explorative study.
Appetite 2018;120:673–78. doi:10.1016/J.APPET.2017.10.034
14 Ljubisavljevic M, Maxood K, Bjekic J, Oommen J, Nagelkerke
N. Long-term effects of repeated prefrontal cortex transcranial
direct current stimulation (tDCS) on food craving in normal
and overweight young adults. Brain Stimul 2016;9(6):826–33.
doi:10.1016/j.brs.2016.07.002
15 Gluck ME, Alonso-Alonso M, Piaggi P, Weise CM, Jumpertz-
Von Schwartzenberg R, Reinhardt M, et al. Neuromodulation
targeted to the prefrontal cortex induces changes in energy
intake and weight loss in obesity. Obesity 2015;23(11):2149–56.
doi:10.1002/oby.21313
Letter to the Editor
Nutritional Neuroscience 2018
4