medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020.

The copyright holder for this preprint

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 1

Suggestibility in functional neurological disorder: A meta-analysis

Lillian Wieder, BSc1, Richard J. Brown, PhD, ClinPsyD2,3, Trevor Thompson, PhD4, & Devin B. Terhune,

PhD1

1
Department of Psychology, Goldsmiths, University of London, London, UK
2
School of Health Sciences, University of Manchester, Manchester, UK
3
Complex Trauma and Resilience Research Unit, Greater Manchester Mental Health NHS Foundation
Trust, Manchester Academic Health Sciences Centre, Manchester, UK
4
School of Human Sciences, University of Greenwich, London, UK

Corresponding author:
Devin B. Terhune, PhD
Department of Psychology
Goldsmiths, University of London
8 Lewisham Way
New Cross, London, UK SE14 6NW
d.terhune@gold.ac.uk

Word count: 2,965 words

NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 2

Abstract

Objective Responsiveness to direct verbal suggestions (suggestibility) has long been hypothesized to

represent a predisposing factor for functional neurological disorder (FND) but previous research has yielded

conflicting results. The aim of this study was to quantitatively evaluate whether FND patients display

elevated suggestibility relative to controls via meta-analysis.

Methods Four electronic databases were searched in November 2019, with the search updated in April

2020, for original studies assessing suggestibility using standardized behavioural scales or suggestive

symptom induction protocols in FND patients and controls. The meta-analysis followed Cochrane,

PRISMA, and MOOSE guidelines. Data extraction and study quality coding were performed by two

independent reviewers. Standardized suggestibility scores and responsiveness to symptom induction

protocols were used to calculate standardized mean differences (SMDs) between groups.

Results Of 26,643 search results, 19 articles presenting 11 standardized suggestibility datasets (FND:

n=316; control: n=360) and 11 symptom suggestibility datasets (FND: n=1285; control: n=1409) were

included in random-effects meta-analyses. Meta-analyses revealed that FND patients displayed greater

suggestibility than controls on standardized behavioural scales (SMD, 0.48 [95% CI, 0.15, 0.81]) and greater

responsiveness to suggestive symptom induction (SMD, 1.39 [95% CI, 0.92, 1.86]). Moderation analyses

presented mixed evidence regarding the extent to which effect sizes covaried with methodological

differences across studies. No evidence of publication bias was found.

Conclusions These results corroborate the hypothesis that FND is characterized by heightened

responsiveness to verbal suggestion. Atypical suggestibility may confer risk for FND and be a cognitive

marker that can inform diagnosis and treatment of this condition.

medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 3

Introduction

Functional neurological disorder (FND) is characterized by impaired motor or cognitive functioning that

resembles neurological pathology but is not adequately explained by it.1 FND has a prevalence of 4-12 per

100,0002, 3 and represents ~16% of clinical referrals and visits to epilepsy clinics.4 It is associated with

considerable diagnostic delays and frequent misdiagnosis,5 which add to the already significant

psychological, social, and economic impact of the condition.6

FND has long been hypothesized to be characterized by elevated responsiveness to direct verbal

suggestions (suggestibility).7 Suggestibility is theorized to confer vulnerability for FND8 through aberrant

meta-awareness of intentions,9-12 the capacity for suggestions to trigger automatized behavioural routines or

mental representations,8 and/or a tendency to form precise (symptom) priors that override motor and

perceptual systems.13, 14 The use of suggestion to provoke FND symptoms is widely used to aid diagnosis15

and functional symptoms are responsive to suggestion-based treatments, such as hypnosis and placebo.16 In

addition, suggestibility has been shown to predict prognosis17 and response to treatment18 in FND patients.

Conditions with germane symptom profiles, such as dissociative disorders, are also characterized by

elevated hypnotic suggestibility.19

Despite these various strands of evidence, the empirical association between suggestibility and FND

seems to be highly variable.20, 21 In order to quantify the evidence for elevated suggestibility in FND

patients, we conducted a random-effects meta-analysis of controlled studies of suggestibility on standardized

behavioural scales and in response to suggestive symptom induction protocols. Secondary analyses

investigated moderating influences on patient-control differences, such as experimenter blindness,22

methodological quality, the inclusion of a hypnotic induction,23 and symptom provocation method.15

Method

Eligibility criteria

Inclusion criteria included (1) English language; (2) full paper in a peer-reviewed journal; (3) patient sample

characterized by FND/symptoms, encompassing conversion disorder (DSM), dissociative neurological

disorder (ICD), specific functional neurological syndromes (e.g., non-epileptic seizures), and conditions

where functional neurological symptoms are a diagnostic feature (i.e. DSM-IV somatization disorder;
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 4

Briquet’s syndrome); (4) inclusion of a control group; and either (5) use of a standardized behavioural

measure of direct verbal suggestibility23, or (6) assessment of symptom induction through suggestion (e.g.,

suggestive seizure induction15).

Exclusion criteria included (1) studies in which suggestion was used to aid diagnosis; (2) case

studies/series or non-empirical papers; (3) overlapping/insufficient data; (4) use of interrogative

suggestibility scales, which capture a different form of suggestibility characterized by high compliance24;

and (5) studies of patients with functional somatic syndromes not specifically characterized by functional

neurological symptoms (e.g., fibromyalgia).

Search strategy

PubMed, PsycINFO, Web of Science and Academic Search Complete databases were searched in November

2019 for eligible studies using terms relating to suggestibility and FND (see online supplemental content

1) and then integrated into a single database. The search was repeated in April 2020 but yielded no new

studies. The reference lists of all eligible studies (and relevant review papers) were manually searched to

identify additional studies. Authors were contacted when data were unavailable or to clarify ambiguities in

methodology.

Study selection

Two raters (LW and a second rater) independently screened and assessed all studies for their eligibility using

a two-stage procedure. First, all titles and abstracts were screened and articles not meeting eligibility criteria

were rejected. Second, all remaining papers were reviewed to establish a final list of articles. Discrepancies

at either stage were resolved in consultation with a third reviewer (DBT) and sometimes a fourth reviewer

(RJB). Authors of eligible studies were contacted to address any questions regarding insufficient data. Of 5

author groups contacted, 4 (80%) provided data sufficient to permit study inclusion or answered queries that

justified exclusion.

Data extraction

The two outcome types (standardized suggestibility and symptom suggestibility) were measured using

continuous and categorical measures, respectively. After exclusion of two studies with overlapping data,
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 5

data from eligible studies were extracted and coded independently by LW and the second rater using a data

extraction form including: (i) study details (title, year, geographical location), (ii) diagnosis, (iii), diagnosis

method, (iv) demographics (sample size and gender and age distributions), (v) study design details

(suggestibility type [standardized or symptom], inclusion of a hypnotic induction, scale, administration

method [live or recorded presentation], and scoring method [self or experimenter]), provocation method [see

Table 2], and experimenter blindness [blind, unblind, or unreported]), and (vi) descriptive statistics (Ms and

SDs [standardized suggestibility] or response counts [symptom suggestibility]). Symptom suggestibility

response counts include only responses identified as typical for the respective patient (typicality reported:

k=8; unreported: k=3). When studies reported results for more than one provocation method (k=2), the

rounded mean was used. Two studies included data for both standardized and symptom suggestibility. There

was 91% agreement between the two raters and discrepancies were resolved with a third reviewer (DBT).

Study quality

A 13-item scale was developed to assess study quality (see online supplemental content 1). Items were

adapted from an earlier meta-analysis that included items based on Cochrane criteria and PRISMA

recommendations25 and a range of other methodological criteria such as experimenter blindness. LW and the

second rater independently rated each item categorically (0=criterion not met, 1=criterion met) and a

summed total was computed for each study. Agreement between raters was 90% (kappa=.73) and

discrepancies were resolved with DBT.

Meta-analysis and meta-regression

Individual study effect sizes included between-group differences in suggestibility that were computed with

standardised mean differences (SMDs; Hedges’s gs) using Review Manager (RevMan v. 5.3, 2014; The

Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen). Symptom suggestibility data consisted

of binary outcomes that were used to compute odds ratios, which were subsequently transformed to SMDs26

in MATLAB (v. 2017b, MathWorks, Natick, MA). SMDs were coded such that positive values reflected

greater suggestibility in FND patients than controls.

Publication bias was assessed by examining funnel plots of effect sizes against standard errors for

asymmetry, as might occur due to a small number of studies with small or non-significant effect sizes. We
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 6

also tested for asymmetry using Egger’s bias test27 where p<.05 is indicative of asymmetry. We report

revised effect sizes correcting for asymmetry using the trim-and-fill method28 and funnel plots in

Figure 3.

Random-effects meta-analyses were performed in JASP (v. 0.8.6, 2019; JASP Team, Netherlands).

Outlier detection was made on the basis of studentized residuals (>|3.3|).29 There were no outliers in the

standardized suggestibility data (range: -1.63 to 1.33) or the symptom suggestibility data (range: -2.01 to

1.26). Moderating effects were assessed using meta-regression analyses whenever data were available for at

least 2 studies at each level of a categorical moderator and at least 10 studies for continuous moderators.30

Moderators included five categorical measures (experimenter blindness, hypnotic induction, symptom

provocation method [suggestion vs. nocebo; symptom studies], symptom typicality [not reported vs. report;

symptom studies]; and control type [non-clinical vs. clinical; standardized studies]) and one continuous

measure (methodological quality).

Results

Study inclusion

A PRISMA diagram showing study selection is presented in Figure 1. Principal features of these studies,

including diagnosis criteria and procedures, are presented in Tables 1 and 2.

** Figure 1 about here **

** Table 1 about here **

** Table 2 about here **

Study and participant characteristics

The 19 papers included 11 standardized suggestibility studies (FND: n=316, control: n=360) and 11

symptom suggestibility studies (FND: n=1285, control: n=1409). The standardized studies were published

between 1984 and 2009 and were conducted in the US (k=5), the Netherlands (k=4), and the UK (k=2). The

symptom studies were published between 1994 and 2018, and were conducted in the US (k=7), Puerto Rico

(k=1), France (k=1), India (k=1) and the UK (k=1). The mean percentage of females for standardized studies

was 86.13 for FND (k=8) and 45.71 for controls (k=7), whereas for symptom studies (k=6) it was 77.17 for

FND and 61.33 for controls. In the standardized studies, the mean age for FND was 31.22 (SD=8.01) (k=6),
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 7

and for controls it was 36.96 (SD=5.09) (k=7), whereas for the symptom studies (k=6), mean age for FND

was 33.41 (SD=8.79), and 37.20 (SD=14.38) for controls.

Details of the types of standardized scales and provocation methods as well as the use of a hypnotic

induction are provided in Table 1 and Table 2, respectively. A hypnotic induction was used in 9

standardized studies and 3 symptom studies. All symptom studies included verbal suggestion but varied in

their use of various nocebo procedures.

Methodological quality criteria

Ratings for each study on each of the 13 methodological quality criteria items are shown in online

supplemental content 1, Tables 1 and 2. Although some of the criteria were met by the majority of studies,

multiple criteria were not reliably met. Only 5 of 22 studies (23%) reported that the experimenter was blind

to group, 18 (82%) described the inclusion/exclusion criteria, 16 (73%) described the diagnosis procedure

and criteria in adequate detail, 10 (45%) described participant characteristics and 3 (14%) exhibited

demographic comparability between patients and controls. In the standardized studies, 7 of 11 (64%)

described the scale and procedure in adequate detail, and only 1 (9%) included a measure to correct for

compliance. In the symptom studies, 8 of 11 (73%) clearly described the provocation method.

Meta-analysis of standardized suggestibility

Meta-analysis of 11 standardized behavioural suggestibility studies found that patients with an FND

exhibited greater suggestibility than controls, SMD=0.48 [0.15, 0.81], Z=2.84, p=.004 (see Figure 2).

Positive results were observed in 8 studies with a high inconsistency of effect sizes across studies, I2=73%.

Jackknife analysis in which each study effect was sequentially omitted and the analysis re-performed

indicated that the group difference was reliably significant (SMD range: 0.40-0.56).

** Figure 2 about here **

Meta-analysis of symptom suggestibility

Meta-analysis of 11 symptom suggestibility studies found that patients with FND displayed greater

responsiveness than controls, SMD=1.39 [0.92, 1.86], z=5.77, p<.001 (see Figure 2). Overall positive

responses were observed in 49% of patients with FND and 6% of controls (sensitivity=49%,

specificity=94%). Ten of the 11 studies exhibited positive SMDs although there was substantial
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 8

inconsistency in effect sizes across studies, I2=99%. Jackknife analysis revealed that the group difference

remained significant after omitting each study in a sequential manner (SMD range: 1.25 to 1.54).

Meta-analysis of standardized vs. symptom suggestibility

The weighted effect size for symptom studies was significantly greater than that for standardized studies,

z=2.60, p=.009. This difference remained stable, z=2.29, p=.022, after removing the two studies included in

both data sets. When the two data sets were aggregated, the cumulative standardised effect size was slightly

less than 1, SMD=0.96 [0.62, 1.29].

Publication bias

Egger’s test did not suggest asymmetry in the distribution of effect sizes in the standardized studies, z=0.69,

p=.49, or in the symptom studies, z=1.50, p=.13. A trim and fill estimate produced only a slight reduction in

effect sizes for standardized studies, ΔSMD=-.05, and symptom studies, ΔSMD=-.10 (see Figure 3).

** Figure 3 about here **

Meta-regression of standardized and symptom suggestibility

Given the observed inconsistency in the magnitude of effects, a set of meta-regression analyses considered

whether group differences were moderated by binary and continuous predictors pertaining to study

methodologies (see Table 3). Effect sizes were significantly larger when a hypnotic induction was included

in standardized studies, but not in symptom studies. Effect sizes were also larger in studies that reported

whether suggested symptoms were typical for the patient relative to those that did not report this

information; this implies that effect sizes are not inflated by the inclusion of atypical symptoms in response

rates for suggestive symptom induction. By contrast, effect sizes were not significantly related to the type of

control (clinical vs. non-clinical; standardized studies) or suggestive induction protocol (suggestion vs.

nocebo [suggestion and sham]; symptom studies). Experimenter blindness did not significantly moderate

group differences with numerical differences in opposing directions for standardized and symptom studies.

Similarly, methodological quality related to effect sizes in opposing directions: greater quality was

significantly associated with lower effect sizes in standardized studies, but with larger (albeit non-

significantly) effect sizes in symptom studies. These effects were primarily driven by procedure description.

Standardized effect sizes were smaller in studies that met criteria for clarity of inclusion/exclusion criteria:
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 9

z=-2.54, p=.011; diagnostic procedure: z=-2.82, p=.005; and scale administration procedure: z=-2.28, p=.022

(all other ps>.09). In contrast, symptom effect sizes were larger in studies that met criteria for clarity of

diagnostic procedure, z=3.23, p=.001 (all other ps>.09).

Discussion

The results of this meta-analysis suggest that FND patients display elevated suggestibility on standardized

behavioural scales and in response to suggestive symptom induction protocols, consistent with theoretical

predictions to this effect.7, 20, 21 There was no evidence for publication bias although there was considerable

heterogeneity in both data sets, which was partly explained by methodological variability.

These findings are consistent with models proposing responsiveness to suggestion as a vulnerability

factor for FND and those attributing functional symptoms to precise symptom priors.8, 13 Recent theoretical

work conceptualizes functional symptoms as arising from an interaction between multiple factors including

an autonomic arousal response, inhibitory deficits that increase the likelihood of body misrepresentation,

and symptom priors that collectively trigger the automatic activation of rogue symptom representations or

conditioned actions.8 Suggestibility may confer heightened sensitivity to symptom-specific cues or

dissociative responses to stressors. Moreover, suggestibility has been proposed to reflect a generalized

tendency to form precise priors that override various motor and perceptual systems,31 which is thought to be

a key process in FND13 and symptom reporting more generally.14 This aligns with research showing that

hypnotic suggestibility predicts symptom severity in FND patients.32 Observed links between hypnotic

suggestibility and emotional responsiveness to social cues,23 highlight the potential role of suggestibility in

symptom modeling or triggering through social observation.8 The perception that functional symptoms are

extra-volitional may be further augmented by aberrant meta-awareness of intentions, as observed in both

FND11, 12 and high hypnotic suggestibility.9, 10

Although the two forms of suggestibility moderately covary,33 FND patients were more responsive to

symptom-specific (SMD=1.39 [0.92, 1.86]) than to standardized suggestions (SMD=0.48 [0.15, 0.81]),

implying selectively greater suggestibility for functional symptoms. Indeed, the sole non-significant

provocation study34 administered suggestions for generic symptoms that did not necessarily mirror patients’

symptom profiles. Previous research similarly found that highly dissociative individuals and dissociative and

acute stress disorder patients are more responsive to suggestions for dissociative experiences (e.g.,
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 10

amnesia).19 Symptom suggestibility in FND may thus partly reflect a kindling process whereby symptoms

become more responsive to verbal suggestion over time.8

Atypical suggestibility in FND patients complements research showing greater suggestibility in germane

conditions, such as dissociative and stress disorders19, 20 and somatoform disorders.35 Many of these

conditions are characterized by detachment and/or compartmentalization symptoms.36 Variability in such

symptoms may account for heterogeneity in the suggestibility profiles of FND patients, with heightened

responsiveness to suggestions being specific to, or more pronounced among, patients with marked

dissociative symptomatology.19, 37 Moderation analyses indicated that the administration of a hypnotic

induction was associated with greater standardized suggestibility among FND patients. This is consistent

with the proposal that individuals with compartmentalization symptoms benefit more from a hypnotic

induction19 although the mechanistic basis of this effect remains unclear.23

These effects attest to the efficacy of suggestion in the diagnosis of FND.15 Suggestive symptom

induction displayed high specificity (94%) although sensitivity is poor (49%), indicating that this technique

is insufficient as a standalone diagnostic procedure. The inclusion of sham methods or a hypnotic induction

were not associated with greater discrimination of FND patients and controls relative to verbal suggestion

alone but warrant further attention. Suggestive symptom induction is likely to be especially valuable in

suggestible populations such as adolescent and elderly samples23 or patients with comorbid dissociative or

stress disorder diagnoses.19, 20 It may also inform diagnosis of comorbid non-epileptic seizures in epilepsy

patients38 and medication prescription in FND patients.39 Insofar as suggestibility is a positive predictor of

outcome with suggestion-based treatments,18 these results also support greater incorporation of suggestion

techniques in treatment protocols, which show promising results in randomized-controlled trials.16 However,

they also highlight the need to control for suggestion, or consider its role, in diagnostic and treatment

procedures, particularly those that evoke strong response expectancies.

Limitations

The principal limitations of these data concern methodological variability across studies. Methodological

quality was significantly or descriptively related to effect sizes in both data sets albeit in opposite directions.

Among standardized studies, older studies that did not precisely specify inclusion/exclusion criteria and/or

diagnostic and scale administration procedures tended to exhibit larger effect sizes; in contrast, precise
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 11

specification of diagnostic procedures was associated with larger effect sizes in symptom studies. In most

studies, the operator administering the assessment was not blind to patient group, which may inflate effect

sizes22, although there was no evidence for an experimenter effect in the standardized studies. The sole blind

symptom study had a lower effect size than the remainder of the studies but was still large in magnitude

(SMD=0.93 [0.61, 1.26]). Most studies included clinical controls (e.g., epilepsy patients), which raises issues

regarding generalizability although effect sizes did not significantly relate to control type. Symptom

suggestibility estimates are also confounded by baseline symptom frequency, which is not incorporated into

these assessments.15 This potentially renders patients with high symptom frequency at an increased risk of

false positive responses, although there is evidence that this is not the case.40 The studies also varied in

whether successful responses to symptom induction protocols were contingent upon the typicality of the

response, which accounted for variability in effect sizes. Standardized studies were limited insofar as only

one controlled for compliance. In addition, standardized suggestibility scales include a disproportionate

number of suggestions for dissociative and functional symptoms (e.g., paralysis),20, 23, 32 raising the question

of whether elevated suggestibility in FND generalizes beyond these symptoms. Collectively, these findings

underscore the need for optimization and standardization of suggestive symptom induction protocols,15

compliance-correction,23 and more diverse suggestion batteries.

Conclusions

This meta-analysis corroborates the long-held view that FND is characterized by elevated suggestibility.

Increased suggestibility has direct implications for the risk factors underlying this condition, the use of

suggestion to aid diagnosis, the utility of suggestion-based treatments for functional symptoms, and

heterogeneity within this population.

Acknowledgements

The authors thank Che Ofuasia, Goldsmiths, University of London, for contributing to the data coding and

evaluation.
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 12

Contributors

All authors conceived the project. LW carried out the database searches and data coding with assistance

from RJB, TT, and DBT. LW and DBT performed the meta-analysis with assistance from RJB and TT. LW

and DBT drafted the initial manuscript. All authors reviewed and approved the final version of the

manuscript.

Funding

This study was supported by Bial Foundation bursary 70/16 (DBT) and a Gyllenbergs Foundation

fellowship (DBT).

Competing interests

No, there are no competing interests for any author

Patient consent for publication

Not required.

Provenance and peer review

Not commissioned; externally peer reviewed.

medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 13

References

1. Espay AJ, Aybek S, Carson A, et al. Current concepts in diagnosis and treatment of functional

neurological disorders. JAMA Neurol 2018;75:1132-1141.

2. Carson AJ, Brown R, David AS, et al. Functional (conversion) neurological symptoms: Research

since the millennium. J Neurol Neurosurg Psychiatry 2012;83:842-850.

3. Benbadis SR, Allen Hauser W. An estimate of the prevalence of psychogenic non-epileptic seizures.

Seizure 2000;9:280-281.

4. Stone J, Carson A, Duncan R, et al. Who is referred to neurology clinics? The diagnoses made in

3781 new patients. Clin Neurol Neurosurg 2010;112:747-751.

5. Reuber M. The etiology of psychogenic non-epileptic seizures: Toward a biopsychosocial model.

Neurol Clin 2009;27:909-924.

6. Gelauff J, Stone J, Edwards M, Carson A. The prognosis of functional (psychogenic) motor

symptoms: a systematic review. J Neurol Neurosurg Psychiatry 2014;85:220-226.

7. Janet P. The major symptoms of hysteria. London, UK: Macmillian & Co., 1907.

8. Brown RJ, Reuber M. Towards an integrative theory of psychogenic non-epileptic seizures (PNES).

Clin Psychol Rev 2016;47:55-70.

9. Lush P, Naish P, Dienes Z. Metacognition of intentions in mindfulness and hypnosis. Neuroscience

of Consciousness 2016:1-10.

10. Terhune DB, Hedman LRA. Metacognition of agency is reduced in high hypnotic suggestibility.

Cognition 2017;168:176-181.

11. Baek K, Donamayor N, Morris LS, et al. Impaired awareness of motor intention in functional

neurological disorder: Implications for voluntary and functional movement. Psychol Med 2017;47:1624-

1636.

12. Jungilligens J, Wellmer J, Schlegel U, Kessler H, Axmacher N, Popkirov S. Impaired emotional and

behavioural awareness and control in patients with dissociative seizures. Psychol Med 2019:1-9.

13. Edwards MJ, Adams RA, Brown H, Parees I, Friston KJ. A Bayesian account of 'hysteria'. Brain

2012;135:3495-3512.

14. Van den Bergh O, Witthoft M, Petersen S, Brown RJ. Symptoms and the body: Taking the

inferential leap. Neurosci Biobehav Rev 2017;74:185-203.

medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 14

15. Popkirov S, Gronheit W, Wellmer J. A systematic review of suggestive seizure induction for the

diagnosis of psychogenic nonepileptic seizures. Seizure 2015;31:124-132.

16. Deeley Q. Hypnosis as therapy for functional neurologic disorders. Handb Clin Neurol

2016;139:585-595.

17. Gambini O, Demartini B, Chiesa V, Turner K, Barbieri V, Canevini MP. Long-term outcome of

psychogenic nonepileptic seizures: The role of induction by suggestion. Epilepsy Behav 2014;41:140-143.

18. Moene FC, Spinhoven P, Hoogduin KA, van Dyck R. A randomized controlled clinical trial of a

hypnosis-based treatment for patients with conversion disorder, motor type. The International journal of

clinical and experimental hypnosis 2003;51:29-50.

19. Terhune DB, Cardeña E. Dissociative subtypes in posttraumatic stress disorders and hypnosis:

Neurocognitive parallels and clinical implications. Current Directions in Psychological Science

2015;24:452-457.

20. Bell V, Oakley DA, Halligan PW, Deeley Q. Dissociation in hysteria and hypnosis: Evidence from

cognitive neuroscience. Journal of Neurology, Neurosurgery, and Psychiatry 2011;82:332-339.

21. Brown RJ, Reuber M. Psychological and psychiatric aspects of psychogenic non-epileptic seizures

(PNES): A systematic review. Clin Psychol Rev 2016;45:157-182.

22. Holman L, Head ML, Lanfear R, Jennions MD. Evidence of experimental bias in the life sciences:

Why we need blind data recording. PLoS Biol 2015;13:e1002190.

23. Terhune DB, Cleeremans A, Raz A, Lynn SJ. Hypnosis and top-down regulation of consciousness.

Neurosci Biobehav Rev 2017;81:59-74.

24. Gudjonsson GH. Interrogative suggestibility and compliance. In: Ridley AM, Gabbert F, La Rooy

DJ, eds. Suggestibility in legal contexts: Psychological research and forensic implications. Chichester, UK:

Wiley-Blackwell, 2013: 45-61.

25. Thompson T, Terhune DB, Oram C, et al. The effectiveness of hypnosis for pain relief: A

systematic review and meta-analysis of 85 controlled experimental trials. Neurosci Biobehav Rev

2019;99:298-310.

26. Hasselblad V, Hedges LV. Meta-analysis of screening and diagnostic tests. Psychol Bull

1995;117:167-178.
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 15

27. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple,

graphical test. BMJ 1997;315:629-634.

28. Duval S, Tweedie R. Trim and fill: A simple funnel-plot-based method of testing and adjusting for

publication bias in meta-analysis. Biometrics 2000;56:455-463.

29. Viechtbauer W, Cheung MW. Outlier and influence diagnostics for meta-analysis. Res Synth

Methods 2010;1:112-125.

30. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions. West Sussex:

Wiley, 2008.

31. Martin JR, Pacherie E. Alterations of agency in hypnosis: A new predictive coding model. Psychol

Rev 2019;126:133-152.

32. Roelofs K, Hoogduin KA, Keijsers GP, Naring GW, Moene FC, Sandijck P. Hypnotic susceptibility

in patients with conversion disorder. J Abnorm Psychol 2002;111:390-395.

33. Khan AY, Baade L, Ablah E, McNerney V, Golewale MH, Liow K. Can hypnosis differentiate

epileptic from nonepileptic events in the video/EEG monitoring unit? Data from a pilot study. Epilepsy

Behav 2009;15:314-317.

34. Chabrol H, Peresson G, Clanet M. Lack of specificity of the traditional criteria for conversion

disorders. Eur Psychiatry 1995;10:317-319.

35. Wickramasekera IE. Somatization: Concepts, data, and predictions from the high risk model of

threat perception. J Nerv Ment Dis 1995;183:15-23.

36. Holmes EA, Brown RJ, Mansell W, et al. Are there two qualitatively distinct forms of dissociation?

A review and some clinical implications. Clin Psychol Rev 2005;25:1-23.

37. Roelofs K, Keijsers GP, Hoogduin KA, Naring GW, Moene FC. Childhood abuse in patients with

conversion disorder. Am J Psychiatry 2002;159:1908-1913.

38. Kutlubaev MA, Xu Y, Hackett ML, Stone J. Dual diagnosis of epilepsy and psychogenic

nonepileptic seizures: Systematic review and meta-analysis of frequency, correlates, and outcomes. Epilepsy

Behav 2018;89:70-78.

39. Gogia B, Rai PK, Matthys SA, et al. Use of noninvasive induction techniques in the diagnosis of

PNES. Epilepsy Behav 2019;99:106491.

medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 16

40. Popkirov S, Jungilligens J, Gronheit W, Wellmer J. Diagnosing psychogenic nonepileptic seizures:

Video-EEG monitoring, suggestive seizure induction and diagnostic certainty. Epilepsy Behav 2017;73:54-

58.
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 17

Figure 1. PRISMA flowchart of study selection process

Figure 2. Forest plots of Standardised Mean Differences (SMDs) (with 95% confidence intervals) from

(left) 11 standardized suggestibility studies and (right) 11 symptom suggestibility studies. Marker sizes

reflect study weights with smaller markers denoting smaller weights.

Figure 3. Funnel plots of standardized mean differences (SMDs) from (left) 11 standardised suggestibility

studies and (right) 11 symptom suggestibility studies. Filled circles denote individual study effect sizes and

empty circles denote estimated missing individual effect sizes due to publication bias imputed using the trim

and fill method. Summary SMDs [95% CI] using the trim-and-fill method were SMD=0.43 [0.10, 0.76]

(standardized suggestibility studies) and SMD=1.29 [0.84, 1.75] (symptom suggestibility studies).
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 18

Table 1. Characteristics of included studies measuring standardized suggestibility

Source Diagnostic criteria FND Controls Scale Hypnotic
induction
Formal Procedure Diagnosis n Age Diagnosis n Age
(% [M (SD)] (% [M (SD)]
female) female)
Khan et al., 2009 - MI NES 24 (-) - (-) ES 16 (-) - (-) HIP Y

Brown et al., 2008 DSM- CR SD 19 (95) 40.9 (14.1) DYS 17 (59) 45.4 (11.1) BSS N
IV
Roelofs et al., 2002 DSM- MI CD 50 (84) 37.2 (11.9) mAD 50 (82) 36.4 (11.1) SHSS:C Y
IV
Litwin and Cardeña, - MI NES 10 (100) 30.5 (9.9) ES 31 (45) 35.2 (8.9) SHCS Y
2001
Moene et al., 2001 / DSM- CR CD 96 (-) - (-) NCC 82 (57) 40.1 (14.3) SHCS Y
Spinhoven et al. III-R
1991*
Barry et al., 2000 - MI, CR NES/ES 36 (-) - (-) ES 22 (-) - (-) HIP Y

Goldstein et al., - MI, CR NES 20 (80) 34.4 (12.4) NCC 20 (45) 36 (8.5) CIS N
2000
Kuyk et al., 1999 - MI NES 20 (80) 25 (-) ES 17 (18) 37 (-) SHCS Y

Kuyk et al., 1995 - MI, CR NES 6 (50) 19.3 ES 7 (14) 28.6 (-) SHCS Y

Bliss, 1984a - CR CD 18 (100) - (-) NCC 49 (-) - (-) SHSS:C Y

Bliss, 1984b - CR BS 17 (100) - (-) NCC 49 (-) - (-) SHSS:C Y

Notes. DSM=Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association); MI=Medical
investigations; CR=case note review for ruling out organic causes; PDF=positive diagnostic features; NES=non-epileptic seizures;
SD=somatization disorder; CD=conversion disorder; ES=epileptic seizures; BS=Briquet’s syndrome; DYS=dystonia; mAD=mixed
affective disorders; NCC=non-clinical controls; HIP = Hypnotic Induction Profile; BSS = Barber Suggestibility Scale; SHSS:C =
Stanford Hypnotic Susceptibility Scale: Form C; SHCS = Stanford Hypnotic Clinical Scale; CIS = Creative Imagination Scale; - =
Not reported; * = Control data were drawn from Spinhoven et al. (1991).
 medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 19

Table 2. Characteristics of included studies measuring symptom suggestibility

Source Diagnostic criteria FND Controls Method Responsiveness: n (%)

Formal Procedure Diagnosis n (% Age Diagnosis n (% Age Suggestion/noceb Hypnotic FND Control
female) [M (SD)] female) [M (SD)] o induction
Kandler et al., - MI, CR NES 416 (-) - (-) ES 568 (-) - (-) VS, HV, PS N 120 40 (7%)
2018 (29%)
Chen-Block et al., - MI, CR NES/ES 466 - (-) ES 581 - (-) VS, HV, PS, AP, N 240 15 (3%)
2016 (70) (54) ST (52%)

Goyal et al., 2014 - MI NES 140 21.38 ES 50 (34) 21.18 VS, PS, VTF, N 84 (60%) 0 (0%)
(71) (10.83) (9.98) TC, MS, SI

Arain et al., 2009 - MI NES 24 (67) - (-) ES 54 (65) - (-) VS, HV N 8 (33%) 6 (11%)

Khan et al., 2009 - MI NES 24 (-) - (-) ES 16 (-) - (-) VS Y 11 (46%) 2 (13%)

Kenney et al., FW PDF PT 12 (-) 42.5 (11) ET 33 (-) 56.8 (17) VS, HV, VTF N 6 (50%) 5 (15%)
2007
Martinez-Taboas, - CR, PDF NES 8 (88) 32 (9.2) ES 5 (100) 30 (6.2) VS, ARS Y 8 (100%) 0 (0%)
2002
Barry et al., 2000 - MI, CR NES/ES 47 (-) - (-) ES 22 (-) - (-) VS Y 36 (77%) 1 (5%)

Chabrol et al, DSM- CR, PDF CD 15 (-) 44 (14) mND 40 (-) 53 (17) VS N 6 (40%) 17 (43%)
1995 IV
Lancman et al., - MI NES 93 (77) 26.7 (-) ES 20 (70) 28.2 (-) VS, AP N 72 (77%) 0 (0%)
1994
Walczak et al., MI NES 40 (90) 33.9 (11.6) ES 20 (45) 34 (11.7) VS, SI N 33 (83%) 5 (25%)
1994
Notes. DSM=Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association); MI=Medical
investigations; CR=case note review for ruling out organic causes; PDF=positive diagnostic features; Fahn and Williams diagnostic
criteria; NES=non-epileptic seizures; ES=epileptic seizures; PT=psychogenic tremor; CD=conversion disorder; ES=essential tremor;
mND=mixed neurological disorders; VS = verbal suggestion; HV=hyperventilation; PS=photic stimulation; AP=alcohol patch;
ST=subjective trigger; VTF=vibrating tuning fork; TC=temple compression; MS=moist swab; SI=saline injection; ARS=age
regression suggestion; - = Not reported.
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Suggestibility in functional neurological disorder 20

Table 3. Meta-regression analyses for standardized and symptom suggestibility studies (SMD [95% CIs] (k))
Moderator
Suggestibility type Analysis groups Z p I2

Hypnotic induction No Yes

Standardized -0.37 [-0.82, 0.09] (2) 0.66 [0.34, 0.97] (9) 2.74 .006 61%
Symptom 1.30 [0.76, 1.85] (8) 1.66 [0.67, 2.64]) (3) 0.62 .54 100%
Experimenter blindness Unblind Blind
Standardized 0.46 [0.03, 0.89] (7) 0.54 [-0.09, 1.16] (4) 0.20 .84 76%
Symptom 1.44 [0.94, 1.93] (10) 0.93 [0.61, 1.26] (1)
Suggestive symptom Suggestion Nocebo
induction method
Symptom 1.20 [0.11, 2.28] (4) 1.50 [0.93, 2.08] (7) 0.69 .49 100%
Suggestive symptom Not reported Typical
typicality
Symptom 0.60 [-0.14, 1.34] (3) 1.70 [1.14, 2.26] (8) 2.07 .038 100%
Control sample type Non-clinical Clinical
Standardized 0.58 [-0.08, 1.24] (4) 0.42 [0.01, 0.82] (7) -0.40 .69 76%
Methodological quality
Standardized 12-item scale -2.21 .027 66%
Symptom 10-item scale 1.76 .079 99%
medRxiv preprint doi: https://doi.org/10.1101/2020.05.30.20117705; this version posted June 3, 2020. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license .

Records identified through Additional records identified

database searching through other sources
(n = 26,642) (n = 1)

Records after duplicate removal

(n = 20,194)

Records screened Records excluded (n = 19,575)

(n = 20,194)

Abstracts and/or full-text Full-text articles excluded (n = 600)

articles assessed for Case studies: 212
eligibility (n = 619) Non-clinical sample: 171
(n = ) Non-eligible clinical sample: 74
No control group: 126
Insufficient data: 5
Duplicate data: 2
Studies included in
Suggestion used for diagnosis: 10
quantitative synthesis
(meta-analysis) (n = 22,
from 19 papers)
Source SMD [95% CI] Source SMD [95% CI]

Litwin & Cardeña 2001 Chabrol et al. 1995

Brown et al. 2008 Arain et al. 2009

Goldstein et al. 2000 Kandler et al. 2018

Moene et al. 2001 Kenney et al. 2007

Barry et al. 2000 Khan et al. 2009

Roelofs et al. 2002 Walczak et al. 1994
Khan et al. 2009 Martinez-Taboas 2002
Kuyk et al. 1999 Chen-Block et al. 2016
Bliss 1984b Barry et al. 2000

Bliss 1984a Lancman et al. 1994

Kuyk et al. 1995 Goyal et al. 2014

Summary SMD Summary SMD

Favours controls Favours FND Favours controls Favours FND

 SMD SMD
(standardized suggestibility) (symptom suggestibility)