- Email: [email protected]
Functional neurological symptoms modulate processing of emotionally salient stimuli
Functional neurological symptoms modulate processing of emotionally salient stimuli Johanna Fiess M.Sc, Brigitte Rockstroh Ph.D., Roger Schmidt M.D., Christian Wienbruch Ph.D., Astrid Steffen Ph.D. PII: DOI: Reference:
S0022-3999(16)30434-2 doi:10.1016/j.jpsychores.2016.10.007 PSR 9229
To appear in:
Journal of Psychosomatic Research
Received date: Revised date: Accepted date:
19 February 2016 20 September 2016 20 October 2016
Please cite this article as: Fiess Johanna, Rockstroh Brigitte, Schmidt Roger, Wienbruch Christian, Steffen Astrid, Functional neurological symptoms modulate processing of emotionally salient stimuli, Journal of Psychosomatic Research (2016), doi:10.1016/j.jpsychores.2016.10.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Functional neurological symptoms modulate processing of emotionally salient stimuli
IP
T
Running title: Automatic emotion processing and FNS
Johanna Fiess, M.Sc., Department of Psychology, University of Konstanz, Germany,
SC R
[email protected]
Brigitte Rockstroh, Ph.D., Department of Psychology, University of Konstanz, Germany,
NU
[email protected]
Roger Schmidt, M.D., Neurological Rehabilitation Center Kliniken Schmieder, Konstanz,
MA
Germany, [email protected]
Christian Wienbruch, Ph.D., Department of Psychology, University of Konstanz, Germany, [email protected]
CE P
Corresponding author:
TE
[email protected]
D
Astrid Steffen, Ph.D., Department of Psychology, University of Konstanz, Germany,
Johanna Fiess, Department of Psychology, University of Konstanz,
AC
P.O. Box 905, D-78457 Konstanz, Germany Phone: +49-7531-884604, Fax: +49-7531-884601 E-mail: [email protected]
Keywords: functional neurological symptoms (FNS); affective pictures; rapid serial visual presentation (RSVP); magnetoencephalography (MEG); conversion; dissociative movement and sensibility disorders.
Number of tables: 2; number of figures: 3; supplementary materials: 2
1
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Abstract
Objective: Dysfunctional emotion processing has been discussed as a contributing factor to
T
functional neurological symptoms (FNS) in the context of conversion disorder, and refers to
IP
blunted recognition and the expression of one‟s own feelings. However, the emotion
SC R
processing components characteristic for FNS and/or relevant for conversion remain to be specified. With this goal, the present study targeted the initial, automatic discrimination of
NU
emotionally salient stimuli.
Methods: The magnetoencephalogram (MEG) was monitored in 21 patients with functional
MA
weakness and/or sensory disturbance subtypes of FNS and 21 healthy comparison participants (HC) while they passively watched 600 emotionally arousing, pleasant, unpleasant or neutral
D
stimuli in a rapid serial visual presentation (RSVP) design. Neuromagnetic activity was
CE P
central regions of interest.
TE
analyzed 110−330 ms following picture onset in source space for prior defined posterior and
Results: As early as 110 ms and across presentation interval, posterior neural activity modulation by picture category was similar in both groups, despite smaller initial (110−150
AC
ms) overall and posterior power in patients with FNS. The initial activity modulation by picture category was also evident in the left sensorimotor area in patients with FNS, but not significant in HC. Conclusions: Similar activity modulation by emotional picture category in patients with FNS and HC suggests that the fast, automatic detection of emotional salience is unchanged in patients with FNS, but involves an emotion-processing network spanning posterior and sensorimotor areas.
2
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Introduction
Altered emotion processing has been discussed as a factor that potentially contributes to
T
the generation of functional neurological symptoms (FNS) ever since FNS have been
IP
conceptualized in the context of conversion or dissociative disorders [cf. 1, 2-7]: Breuer and
SC R
Freud linked FNS to a conversion of intrapsychic distress into physical symptoms [2], and Janet linked psychoform and somatoform symptoms (like FNS) to the dissociation of psychobiological systems (thoughts, sensations, and behavior) consequent upon psychological
NU
trauma and individual predispositions [1, 3-5, 8]. Pavlov connected emotion processing and
MA
FNS when proposing that excessive cortical inhibitory control over (subcortically generated) emotional activities (centered in subcortical areas) flooded other neurological pathways,
D
thereby causing FNS−like paralysis or anesthesia [cf. 7, 9, 10].
TE
Within the framework of psychological emotion processing theories, emotion processing is defined by distinct processes such as the initial, automatic detection of emotional and salient
CE P
“ecologically important” [11, p. 773] stimuli and/or controlled processing, the conscious processing of one‟s own emotional responses or their regulation. Evaluating dysfunctional
AC
emotion processing in individuals with FNS indicated a different initial detection of emotional stimulus category: in an incidental affective task, patients with FNS responded with a greater amygdala activity to unpleasant stimuli (measured by functional magnetic resonance imaging, fMRI) and also showed less activity habituation across trials compared to healthy control participants [12]. Moreover, patients with FNS exhibited less modulation of amygdala activation by stimulus valence (unpleasant – pleasant) than HC [13]. Aybek et al. [12] further showed that emotionally salient stimuli prompted greater activation in the supplementary motor area (SMA) in patients with FNS than in controls, and a greater functional connectedness between amygdala and SMA [13, 14], thus, between brain areas associated with emotion processing and movement preparation.
3
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Referring to emotion regulation, we [15, this journal] found less frontocortical but more sensorimotor neuromagnetic activity in patients with FNS than in healthy control participants, when subjects activated trained strategies of cognitive reappraisal upon presentation of
IP
T
unpleasant pictures [16]. When subjects passively watched the unpleasant and neutral pictures
SC R
in a second condition within the same design, patients and controls displayed similar activity in posterior areas. While the latter activity was associated with the processing of the emotional salience of visual stimuli [11, 17, 18], sensorimotor activity was discussed as an
NU
indication of the close association of emotion and sensorimotor processing in patients with
MA
(dominant sensory and motor) FNS.
Complementing these findings, the present study scrutinized the initial, automatic discrimination of emotionally salient stimuli using a rapid serial visual presentation (RSVP)
TE
D
design to enforce automatic detection and high-resolution magnetoencephalography (MEG) to depict the fast cortical processing. The RSVP design has been shown to prompt greater
CE P
posterior activity to arousing pleasant and unpleasant compared to neutral stimuli as early as 150 ms after stimulus onset [17, 19-21] and has been used to study dysfunctional emotion
AC
processing in patients with posttraumatic stress disorder (PTSD), major depression disorder (MDD) or stress-related syndromes [20-22]. In the present study, neuromagnetic activity modulation by picture category was compared between patients with FNS and healthy comparison participants (HC) with the hypotheses that: (1) patients with FNS show enhanced processing of emotionally salient relative to neutral stimuli (referring to evidence from fMRI studies [12, 13]) in posterior areas associated with a visual emotional processing network [e.g. 17]); and that (2) patients with FNS display greater activity in sensorimotor areas than HC (referring to evidence of simultaneous activity in emotion- and movement-related areas in emotion recognition [12, 13] and emotion regulation [15] tasks). 4
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Methods Participants
T
The study involved 21 patients with FNS and 21 healthy comparison participants (HC).
IP
Samples overlapped with those reported in Fiess et al. [15; this journal]. Patients with an ICD
SC R
diagnosis of conversion disorder (ICD-10 codes F44.4, F44.6, F44.7) were recruited from the local rehabilitation center (Kliniken Schmieder) where they were following comprehensive treatment protocols involving individual and group psychotherapy, physiotherapy and
NU
occupational therapy. Diagnoses were given by two or more experienced psychiatrists and
MA
neurologists following standardized ICD-10 guidelines, with at least one core negative somatoform dissociative symptom (e.g., motor disorders, hypesthesia) required for a
D
diagnosis of conversion disorder. Symptom duration varied between 7 and 41.5 months
TE
(interquartile range) around the median of 12 months. Dominant symptoms were motor
AC
CE P
weakness and/or sensory disturbances on the left and/or right side of the body (see Table 1A).
5
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Table 1A: Dominant symptoms in patients with FNS (n = 21) Right-sided
Both sides
None
Motor weakness (n)
6
1
12
2
Sensory disturbances (n)
5
3
9
IP
T
Left-sided
4
Patients with
SC R
Table 1B: Sociodemographic and clinical information of study samples HC
vs. HC
21
21
Gender (f/m)
13/8
11/10
n. s.
Age (M±SD)
42±13.4
48±14.3
n. s.
Years schooling (M±SD)
11±3.2
11±1.6
n. s.
SDQ-20 (median (range))
33 (28–40)
21 (20–22.5)
U = 26, z = -4.9; p < 0.001, r = 0.76
1.4 (0.3–1.9)
0.1 (0–0.3)
U = 82.5, z = -3.5; p < 0.001, r = 0.54
8 (0.5–24)
0 (0–4)
U = 108, z = -2.9; p < 0.01, r = 0.45
213 (15–585)
34 (8–75.5)
U = 78, z = -2.6; p < 0.01, r = 0.39
18
20
Left-handed
1
1
Ambidextrous
2
0
D
CE P
PSSI (median (range)) ETI (median (range)) Handedness (n)
AC
Right-handed
TE
SCL-90-R (median (range)) Depression
MA
n
NU
FNS
Patients with FNS
n. s.
Note. FNS = patients with functional neurological symptoms; HC = healthy control participants; SDQ20 = Somatoform Dissociation Questionnaire (scores range from 20−100); SCL-90-R = 90-item Symptom Checklist (scores range from 0−4); PSSI = Posttraumatic Stress Scale-Interview severity score (scores range from 0−86); ETI = Early Trauma Inventory.
6
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
HC were recruited from the local community through flyers and verbal advertising in order to be demographically comparable to the patient sample. HC were screened using the Mini International Neuropsychiatric Interview [MINI; 23] to exclude any psychiatric disorders.
IP
T
From n = 24 individuals screened with the MINI, n = 3 were excluded because of a
SC R
hypomanic episode in the past, a current major depressive episode, and alcohol abuse, respectively. Recruitment was continued until the control group consisted of n = 21 participants. As evident from Table 1B, groups did not differ with respect to mean age, gender
NU
distribution or years of school education. Patients with FNS and HC with central nervous lesions (e.g., degenerative disorders, tumors) were not included. For patients with FNS, this
MA
screening was accomplished upon admission following a standard protocol that included screening for neuropathology, clinical structural MRI and electromyography as clinically
D
indicated. All participants had normal or corrected-to-normal vision. The Fisher-Freeman-
TE
Halton test (an extension of the Fisher exact test for r x c contingency tables) confirmed that
CE P
groups did not differ significantly in handedness, as assessed using the Edinburgh Handedness Inventory [24; see Table 1B].
AC
Setting
The study protocol was approved by the ethics committee of the University of Konstanz. Participants were informed about the design and procedures and provided their written informed consent. Thereafter, demographic and self-report data (see below) were assessed. A separate session comprised the rapid serial visual presentation (RSVP) protocol [17] while the MEG was being monitored. Stimulus material For the RSVP, highly arousing pleasant (n = 100), unpleasant (n = 100) and neutral pictures (n = 100) were chosen from the International Affective Picture System [IAPS; 25] on the basis of their normative ratings. Brightness, contrast and color spectra of the stimuli were 7
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
matched across picture categories. Each stimulus was presented once within each of two series of 300 pictures (total 600 stimuli) without a perceivable gap for 333 ms each (3 Hz, 60 Hz refresh rate) in a pseudorandom sequence. The two picture series were presented without a
IP
T
break, the presentation time was approximately 4 minutes. Participants were instructed to
SC R
maintain their focus on a small, centrally located fixation cross overlaying each picture and to watch the sequence of pictures without any specific task. Data assessment and analysis
NU
Clinical data: The severity of somatoform dissociation symptoms was determined using
MA
the Somatoform Dissociation Questionnaire [SDQ-20; 26, German version by 27]. Participants were further characterized by comorbid psychopathology using the Posttraumatic
D
Stress Scale-Interview [PSSI; 28, 29] for screening trauma-related symptoms, and the German
TE
version of the 90-item Symptom Checklist [SCL-90-R; 30, 31, 32]. Moreover, adverse childhood experiences, which have been related to psychopathology, were screened using the
CE P
German version of the Early Trauma Inventory [ETI; 33, German version by 34, see 35]. Given that self-report data of HC were not normally distributed, group differences were
AC
evaluated with non-parametric Mann-Whitney U-tests reporting exact significance values. Effect sizes were calculated as suggested by Fritz et al. [36, p. 12] by the formula r = z / √N. MEG data acquisition and analysis: Magnetic fields were continuously recorded using a 148-channel whole-head magnetometer (MAGNES 2500 WH, 4D Neuroimaging, San Diego, USA) with a sampling rate of 678.16 Hz, and filtered with an analog bandpass filter from 0.1 to 200 Hz. The participant‟s head position was monitored using five coils (nasion, inion, Cz, left and right ear canal) and headshape, digitized with a Polhemus 3Space® Fasttrack. Signals recorded by eleven MEG reference sensors were used to remove external, non-biological noise.
8
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Preprocessing of the MEG data was conducted with FieldTrip, a MATLAB-based toolbox [37]. Trials containing movement artifacts and superconducting quantum interference device (SQUID) jumps were rejected based on visual inspection. Five dysfunctional sensors were
IP
T
excluded from further analysis. Independent component analysis (ICA) was used to visually
SC R
detect eye or cardiac artifacts, which were subsequently removed from the data. Frequencies above 40 Hz were filtered with a FIR low pass filter using a Kaiser window. The number of artifact-free trials retained did not differ between groups or picture category (means ±
NU
standard deviation, FNS: neutral – 196.7±2.78, unpleasant – 197±3.3, pleasant – 197.1±2.19; HC: neutral – 196±2.3, unpleasant – 196.3±2.72, pleasant – 197.1±3). Average event-related
MA
magnetic fields were calculated for each picture category and subject. Source analysis. The L2 minimum-norm was calculated based on a one-shell spherical
D
head model with evenly distributed 2 (azimuthal and polar direction) × 350 dipoles as a
TE
source model using the MATLAB-based software EMEGS© [38]. The L2 minimum-norm
CE P
estimate enables an enhanced resolution of brain activations generating the magnetic field without a priori assumptions regarding the location and number of current sources [39]. This
AC
calculation was based on information on the center of a sphere fitted to the digitized head shape and the positions of the MEG sensors relative to the head. A source shell radius of 87% of the spherical volume conductor head radius was chosen, which roughly corresponded with the grey matter volume. Across all participants and conditions, the Tikhonov regularization parameter λ was set to 0.2. Source data were used for hypothesis testing: In a first step, the global field power of the L2 minimum-norm across time was plotted separately for each group and each stimulus category. In a second step, time windows and regions distinguishing groups and stimulus categories were determined based on (a) previous evidence of dominant activity modulation by stimulus valence in posterior cortical regions [17, 19, 21], (b) on previous reports of
9
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
sensorimotor activity distinguishing patients with FNS and controls in an emotion regulation task [15], and (c) on visual inspection (see also Figure 2A, 3B, and Supplementary Figure). Because previous research indicated a right-hemispheric dominance of activity modulation by
IP
T
picture category in healthy controls [e.g. 17, 21], hemisphere (left vs. right) was taken into
SC R
account as additional factor.
The modulation of neural activity by group and by stimulus category was evaluated in factorial repeated-measures ANOVAs with the between-subject factor Group (patients with
NU
FNS vs. HC) and the within-subject factor Emotion (pleasant vs. neutral vs. unpleasant
MA
stimuli) separately for the three time windows of interest (110–150 ms, 150–220 ms, and 220–300 ms). In three separate ANOVAs dependent variables of neural activity were overall activity (global field power), and activity in the two regions of interest (ROIs), the latter
D
including the within-subject factor Hemisphere (left vs. right). For interactions, degrees of
TE
freedom were corrected using Greenhouse-Geisser‟s estimates of sphericity where
CE P
appropriate; adjusted p values are reported. Confounds by multiple comparisons were controlled by Bonferroni-corrected post-hoc analyses. A potential confound by comorbid
AC
symptoms of depression, trauma-related symptoms, amount of adverse childhood experiences and the modulation by stimulus salience was evaluated by Spearman's rank-order correlations including the individual questionnaire scores (depression score of the SCL-90-R, PSSI symptom score, amount of adverse childhood experiences, ETI) and the overall neural activity across all emotional picture categories (pleasant, neutral and unpleasant). In addition (in patients with FNS), the questionnaire scores were considered as covariates of no interest in the repeated measures ANOVA.
10
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Results
Figure 1 illustrates stimulus-evoked brain activation as global field power across time
T
separately for the two groups and the three picture categories. Activity differences between
IP
groups and stimulus conditions are evident in the three time windows (110–150 ms, 150–220
SC R
ms and 220–330 ms, see Table 2A). Table 2 summarizes statistical effects for overall activity
AC
CE P
TE
D
MA
NU
(Table 2A), the posterior, and the central ROI (Table 2B, 2C).
11
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
D
Figure 1: Global field power averaged separately for patients with FNS (functional
TE
neurological symptoms; black) and HC (healthy comparison participants; grey) as well as for the three picture categories pleasant (dotted), neutral (dash-dotted) and unpleasant (solid).
CE P
Grey bars mark the three time windows selected for analyses (110−150 ms, 150−220 ms, and
AC
220 −330 ms after stimulus onset).
12
ACCEPTED MANUSCRIPT Automatic emotion processing and FNS
CR
IP
T
Table 2: Statistical effects on neural activity in the three latency windows, separately for the global power, the posterior and central regions of interest. 110–150 150–220 220–330 F(1,40) = 5.9; p = 0.02 n. s. n. s.
B: Posterior areas (FNS x HC) Group Emotion Hemisphere Group x Emotion Group x Hemisphere Emotion x Hemisphere Group x Emotion x Hemisphere
F(1,40) = 4.28; p = 0.04 F(1.57,62.95) = 3.11; p = 0.05 n. s. n. s. n. s. n. s. n. s.
C: Central areas (FNS x HC) Group Emotion Hemisphere Group x Emotion Group x Hemisphere Emotion x Hemisphere Group x Emotion x Hemisphere
F(1,40) = 4.72; p = 0.04 F(2,80) = 3.68; p = 0.03 F(1,40) = 10.05; p < 0.01 n. s. n. s. F(2,80) = 4.18; p = 0.02 n. s.
n. s. F(2,80) = 6.18; p < 0.001 n. s.
n. s. F(1.54,61.72) = 5.94; p < 0.01 F(1,40) = 7.02; p = 0.01 n. s. n. s. n. s. n. s.
n. s. F(2,80) = 6.97; p < 0.01 n. s. n. s. n. s. F(2,80) = 3.71; p = 0.03 n. s.
n. s. n. s. F(1,40) = 20.99; p < 0.001 n. s. n. s. n. s. n. s.
n. s. n. s. F(1,40) = 15.68; p < 0.001 n. s. n. s. n. s. n. s.
MA N
TE D
CE P
AC
n. s. F(2,80) = 2.63; p = 0.08 n. s.
US
A: Global power (FNS x HC) Group Emotion Group x Emotion
Note. FNS = functional neurological symptoms; HC = healthy comparison participants.
13
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
As evident in Figure 1, patients with FNS showed smaller overall activity than HC mainly in the initial time window. This was confirmed by a main effect Group in the initial (110-150 ms) interval, while groups did not differ in the two subsequent time windows (see Table 2A).
IP
T
Table 2 further indicates that modulation of activity by stimulus category is area-specific (in
SC R
the posterior ROI) and only marginally evident in global activity. In line with previous evidence [17, 21], main effects Emotion are verified for the posterior ROI, but evident in global activity only in the 220−330 ms time window.
NU
Falsifying hypothesis (1), the modulation of dipole activity by emotional picture content in the posterior ROI was similar in both groups, that is, the Group x Emotion interaction did not
MA
reach significance (p > .05; main effects Emotion in all three time windows p ≤ .05, see Table 2B and Figure 2A). In line with previous results [17, 21], a right-hemispheric dominance of
D
posterior dipole activity was verified for the 150−220 ms interval independent of group and
TE
stimulus content (main effect Hemisphere, p ≤ .05; see Table 2B), and a greater modulation
CE P
by stimulus content in the right compared to the left hemisphere in the 220−300 ms window (Emotion x Hemisphere interaction; see Table 2B). Planned group-specific analyses verified
AC
the modulation of dipole activity by emotional picture content (main effect Emotion) for patients with FNS (F(2,40) = 5.28; p = 0.01) and HC (F(2,40) = 4.45; p = 0.02) in the late time window (220−330 ms), and additionally for HC in the intermediate time window (150−220 ms, F(2,40) = 4.15; p = 0.02).
14
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
Figure 2A: Upper rows: Statistical parametric topographical maps of the main effect Emotion (pleasant, neutral, and unpleasant picture category) separately averaged over all participants for the three time windows (110–150 ms, 150–220 ms and 220–330 ms following stimulus onset). Lower row: Schematic layout of the dipole distribution in the posterior areas used for statistical analyses per time window. B Estimated neural strength (ordinate: dipole activity expressed as minimum norm estimates, mean ± standard error in nA/m) in posterior areas plotted separately for groups (grey: patients with functional neurological symptoms, FNS; black: healthy comparison participants, HC) and picture category (pleasant, neutral, unpleasant) within each of the three time windows (abscissa) as well as for the left and right hemisphere. 15
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Referring to hypothesis (2), Figure 3 illustrates the modulation of neural activity by emotional picture category in the central ROI. The group difference reported for initial overall
T
activity (110−150 ms) was indicated for the central ROI (main effect Group, Table 2C and
IP
Figure 3). However, initial activity modulation was evident only in the left-hemispheric
SC R
central ROI (main effect Emotion, Emotion x Hemisphere interaction, see Table 2C and Figure 3B). Although the Group x Emotion interaction did not reach significance (F(2,80) = 1.26; p = .29), planned comparisons verified the prominent left-hemispheric neural activity
NU
modulation for patients with FNS (F(2,40) = 6.47; p < 0.01), but not for HC (F(2,40) = 1.65;
MA
p = .24).
A potential impact of comorbid psychopathology on the reported neural activity was
D
evaluated by correlation analyses and repeated measures ANOVA. Patients with FNS
TE
experienced more somatoform dissociation, more comorbid psychopathology including trauma-related symptoms, and a greater childhood stress load than HC (see Table 1B and
CE P
Supplementary Table). Yet, neither depression, trauma-related symptoms scores nor childhood stress load revealed significant correlations with the global power (p > .1) or ROI
AC
specific activity (p > .1). Similarly, controlling for comorbid symptoms and childhood stress load as factor in the respective ANOVAs in patients with FNS had no impact on the Emotion main effect (F(2,43) = 4.71; p = 0.02). These results do not indicate a major influence of comorbid psychopathology on the reported neural activity modulation.
16
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
Figure 3: A Estimated neural strength (ordinate: dipole activity expressed as minimum norm estimates, mean ± standard error in nA/m) in the central region of interest (upper right corners: schematic layout of the dipole distribution per hemisphere used for statistical analyses) plotted separately for groups (grey: patients with functional neurological symptoms, FNS; black: healthy comparison participants, HC) and picture category (pleasant, neutral, unpleasant) within each of the three time windows (abscissa: 110–150 ms, 150–220 ms and 220–330 ms following picture onset) as well as for the left and right hemisphere. B Topographies representing statistical parametric maps of the contrasts between picture categories (pleasant-neutral and unpleasant-neutral) per group (patients with FNS and HC) in the time window 110–150 ms following picture onset. 17
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Discussion
Dysfunctional emotion processing has been discussed as a contributing factor to functional neurological symptoms [FNS; 12-14, 40, 41, 42], but the specific process or
IP
T
processes that are dysfunctional in or relevant for FNS remain unclear. The present study
SC R
addressed the capability of patients with FNS to automatically distinguish emotionally salient and neutral stimuli in the rapid flow of visual stimuli with the hypothesis that cortical signs of this initial recognition of emotionally salient information differ from the pattern usually
NU
observed in healthy individuals, suggesting abnormal emotion processing at a very early
MA
processing stage.
Falsifying the hypothesis (1) of enhanced neuromagnetic activity in response to
D
emotionally salient (relative to neutral) pictures, patients with FNS and HC exhibited similar
TE
activity modulation by emotional picture category in posterior areas within 110−330 ms following stimulus onset. Whereas fMRI studies reported augmented subcortical (amygdala)
CE P
responses to emotionally salient stimuli [12, 13], the present MEG-study added evidence that cortical responses to visual emotional input within the first 110−330 ms distinguish
AC
emotionally salient from neutral stimuli in patients with FNS and healthy controls alike. Whether and in which way such early and fast cortical responding was related to subcortical (amygdala) activation could not be determined from the MEG-measurement with sufficient precision. A relationship might be assumed, as Vuilleumier et al. [43] showed that visual cortex activation by different facial expressions (sad, happy) can be modified by amygdala activity, potentially via fast feedback loops [see also 11]. The present results suggest that activity modulation in a network associated with the processing of visual, emotionally salient input [11] did not seem dysfunctional in the patients with FNS compared to HC. Initial activity modulation by emotional salience was evident, although evoked activity in posterior areas associated with emotion processing 110–150 ms following picture onset was 18
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
lower in patients with FNS than in HC. This suggests less or slower responsiveness of the visual system or a network associated with processing of emotional salience in patients.
T
Posterior activity may reflect information transfer along the ventral and dorsal path to an
IP
emotion processing network, in which thalamic nuclei (particularly the pulvinar) project to
SC R
amygdala and cortical regions like insula, cingulate, frontal, parietal cortex, and the precuneus at the transition between the parietal and central cortex [11, 18]. Reduced thalamic activity has been reported in patients with FNS compared to healthy controls [cf. 44, 45]. Thus, lower
NU
activity may indicate slower activation by visual input. Importantly, this smaller activity
MA
increase was restricted to the initial response, while there was no group difference in later activity up to 330 ms. This argues against generally reduced thalamic pulse or the lasting impact of this pulse in FNS. Importantly, the lower initial response in patients with FNS did
TE
D
not affect activity modulation by emotional salience, thus, distinction of salient stimuli succeeded in patients with FNS even at a lower activity level. Other factors might have
CE P
contributed to the blunted posterior activity: Blunted responses have been reported for patients with major depression disorder posttraumatic stress disorder, and/or in syndromes
AC
related to adverse childhood experiences [20-22]. More intense depression and higher stress load were reported in patients with FNS [44-48]. However, neither posttraumatic and depressive symptoms nor overall childhood stress load were significantly associated with the reported lower initial posterior activity and the modulation by emotional picture category. On the contrary to the assumed greater activity in sensorimotor areas in patients with FNS compared to HC (hypothesis (2)), both groups did not differ concerning to this manner in the left central ROI, but patients with FNS showed even less activity in the right central ROI. However, patients with FNS displayed activity modulation by emotional picture category in left sensorimotor areas as early as 110−150 ms following stimulus onset. Although interaction with the factor Group did not reach statistical significance, planned comparisons indicate 19
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
prominent left-sensorimotor activity modulation by picture category in patients with FNS (but not in HC). This provides an indication of an involvement of sensorimotor regions during
T
emotion processing, which can be seen in line with previous studies [12, 13, 15]. Thus,
IP
sensorimotor activity modulation may index genuine sensorimotor involvement in emotion
SC R
processing independent of the specific processing, automatic salience detection or controlled emotion regulation.
NU
The involvement of movement-related cortical areas such as SMA while emotion processing tasks has been demonstrated in patients with conversion disorder [12-14] and may
MA
indicate the activation of a cortical network spanning systems related to emotion processing and to the preparation of complex movements and motor inhibition [49, 50]. For conversion
D
disorders, it has been hypothesized that such movement-related processes are augmented in
TE
order to ward off (painful) emotional responses, which may result in the perception of symptoms [1, 14]. In our previous evaluation of neuromagnetic activity during an emotional
CE P
regulation task in a sample overlapping with the one of the present study [15], patients with FNS displayed greater sensorimotor parallel to lower frontocortical activity than controls.
AC
These results were discussed as being a potential manifestation of a dysfunctional activity shift from emotion-related to movement-related brain regions. The similarity of results between the two studies or tasks may indicate the involvement and co-activation of sensorimotor areas in FNS independent of the level of attention and control devoted to emotion processing. A reduced response in regions associated with automatic emotion detection together with the activity modulation by emotional picture category in sensorimotor regions could be conceived of as an emotional defense response, which then prompts a „conversion‟ of emotional processing, for example, to sensorimotor systems. A major impact of carry-over effects is unlikely, as the sequence of the tasks scheduled within one session varied across subjects. 20
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
It is tempting to consider the left-sensorimotor activity modulation in patients with FNS as (sign of) involvement of the hemisphere contralateral to the affected limb. However, no
T
dominance of right-sided motor weakness and/or sensory disturbances was evident in the
IP
present patient sample. Moreover, handedness did not differ between patients with FNS and
SC R
HC. The left-central focus of activity extended to the left anterior parietal region, which has been related to motor attention [51, 52]. For healthy individuals, sensorimotor activity has also been discussed as being „embodied‟ emotion – that is, as co-activation of motor systems
NU
associated with facial or limb movements involved in the expression of an emotional
MA
response, even if this response is not executed [53]. Augmented or altered embodied emotion has rarely been addressed in the context of conversion disorder or FNS, and the present results cannot resolve that dysfunctional embodied emotion affected the sensorimotor activity pattern
TE
D
in patient with FNS patients as a result of some conversion process. The limitations of the present study must also be considered. The sample only included
CE P
patients with at least one core negative functional symptom (i.e., a motor and/or sensory deficit), and no patients with e.g. functional seizures − thus limiting the applicability of the
AC
conclusions of the present study to FNS in general. These specific inclusion criteria further limited the sample size, precluding subgroup analyses for functional motor or sensory symptoms. The latter distinction should inform the specific nature of emotion processing in patients with specific FNS. While the present results inform the fast, initial and automatic detection of emotionally salient stimuli, they do not specify the subcortical-cortical network supposed to be involved in this automatic processing, so that the functional impact of (dys)functional automatic detection in the course of emotion processing cannot be identified. Additionally, this study cannot address whether these differences in emotion processing (reduced posterior neural activity and early emotion discrimination in sensorimotor regions) reflect the patient‟s current state or whether this is a longer-term trait. To answer this 21
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
question, longitudinal studies are needed to chart and monitor potential changes over time, especially before and after successful treatment.
T
In conclusion, the present study adds to the understanding of altered emotion processing in
IP
patients with FNS, verifying the functional automatic discrimination of emotional stimuli and
AC
CE P
TE
D
MA
NU
SC R
emphasizing the involvement of the sensorimotor cortex in emotion processing.
22
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [1]
Crommelinck M. Neurophysiology of conversion disorders: a historical perspective.
Neurophysiol Clin. 2014;44(4):315-21. Breuer J, Freud S. Studien über Hysterie. Leipzig and Vienna: Deutike; 1895.
[3]
Janet P. The major symptoms of hysteria. New York: The Macmillan Company; 1907.
[4]
Kozlowska K. Healing the disembodied mind: contemporary models of conversion
disorder. Harv Rev Psychiatry. 2005;13(1):1-13. [5]
SC R
IP
T
[2]
Vuilleumier P. Brain circuits implicated in psychogenic paralysis in conversion
Ban TA. Conditioning Behavior and Psychiatry. New Brunswick (U.S.A.) and London
MA
[6]
NU
disorders and hypnosis. Neurophysiol Clin. 2014;44(4):323-37.
(U.K.): AldineTransaction; 2008. [7]
Todes DP. Ivan Pavlov: A Russian Life in Science. New York: Oxford University
TE
[8]
D
Press; 2014.
Van der Hart O, Nijenhuis ERS, Steele K. The Haunted Self: Structural Dissociation
Company; 2006. [9]
Vuilleumier P. Hysterical conversion and brain function. Prog Brain Res.
AC
2005;150:309-29. [10]
CE P
and the Treatment of Chronic Traumatization. New York and London: W. W. Norton &
Windholz G. Pavlov and the mind-body problem. Integr Physiol Behav Sci.
1997;32(2):149-59. [11]
Pessoa L, Adolphs R. Emotion processing and the amygdala: from a „low road‟ to
„many roads‟ of evaluating biological significance. Nature reviews Neuroscience. 2010;11(11):773-83. [12]
Aybek S, Nicholson TR, O'Daly O, Zelaya F, Kanaan RA, David AS. Emotion-motion
interactions in conversion disorder: an FMRI study. PLoS One. 2015;10(4):e0123273. [13]
Voon V, Brezing C, Gallea C, Ameli R, Roelofs K, LaFrance WC, Jr., et al. Emotional
stimuli and motor conversion disorder. Brain. 2010;133(Pt 5):1526-36.
23
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [14]
Aybek S, Nicholson TR, Zelaya F, et al. Neural correlates of recall of life events in
conversion disorder. JAMA Psychiatry. 2014;71(1):52-60. [15]
Fiess J, Rockstroh B, Schmidt R, Steffen A. Emotion regulation and functional
T
neurological symptoms: Does emotion processing convert into sensorimotor activity? J
Moser JS, Krompinger JW, Dietz J, Simons RF. Electrophysiological correlates of
SC R
[16]
IP
Psychosom Res. 2015;79(6):477-83.
decreasing and increasing emotional responses to unpleasant pictures. Psychophysiology.
[17]
NU
2009;46(1):17-27.
Junghöfer M, Bradley MM, Elbert TR, Lang PJ. Fleeting images: A new look at early
[18]
MA
emotion discrimination. Psychophysiology. 2001;38(2):175-8. Habel U, Klein M, Kellermann T, Shah NJ, Schneider F. Same or different? Neural
Peyk P, Schupp HT, Elbert T, Junghöfer M. Emotion processing in the visual brain: a
TE
[19]
D
correlates of happy and sad mood in healthy males. Neuroimage. 2005;26(1):206-14.
MEG analysis. Brain Topogr. 2008;20(4):205-15. Elbert T, Schauer M, Ruf M, Weierstall R, Neuner F, Rockstroh B, et al. The Tortured
CE P
[20]
Brain. Z Psychol. 2011;219(3):167-74. [21]
Matz K, Junghöfer M, Elbert T, Weber K, Wienbruch C, Rockstroh B. Adverse
AC
experiences in childhood influence brain responses to emotional stimuli in adult psychiatric patients. Int J Psychophysiol. 2010;75(3):277-86. [22]
Zwanzger P, Klahn AL, Arolt V, Ruland T, Zavorotnyy M, Sälzer J, et al. Impact of
electroconvulsive therapy on magnetoencephalographic correlates of dysfunctional emotional processing in major depression. Eur Neuropsychopharmacol. 2016;26(4):684-92. [23]
Ackenheil M, Stotz G, Dietz-Bauer R, Vossen A. Mini International Neuropsychiatric
Interview (German Version 5.0.0). München, Germany: Psychiatrische Universitätsklinik München; 1998. [24]
Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory.
Neuropsychologia. 1971;9(1):97-113.
24
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [25]
Lang PJ, Bradles MM, Cuthbert BN. International affective picture system (IAPS):
Instruction manual and affective ratings (Tech. Rep. No. A-6). Gainesville, FL: University of Florida, The Center for Research in Psychophysiology; 2005. Nijenhuis ERS, Spinhoven P, Van Dyck R, van der Hart O, Vanderlinden J. The
T
[26]
IP
development and psychometric characteristics of the Somatoform Dissociation Questionnaire
[27]
SC R
(SDQ-20). J Nerv Ment Dis. 1996;184(11):688-94.
Mueller-Pfeiffer C, Schumacher S, Martin-Soelch C, Pazhenkottil AP, Wirtz G,
Fuhrhans C, et al. The validity and reliability of the German version of the Somatoform
NU
Dissociation Questionnaire (SDQ-20). Journal of trauma & dissociation : the official journal of the International Society for the Study of Dissociation (ISSD). 2010;11(3):337-57. Foa EB, Tolin DF. Comparison of the PTSD Symptom Scale-Interview Version and
MA
[28]
the Clinician-Administered PTSD scale. J Trauma Stress. 2000;13(2):181-91. Foa EB, Cashman L, Jaycox L, Perry K. The validation of a self-report measure of
D
[29]
[30]
CE P
1997;9(4):445-151.
TE
posttraumatic stress disorder: The Posttraumatic Diagnostic Scale. Psychol Assess.
Derogatis LR, Savitz KL. The SCL–90–R and Brief Symptom Inventory (BSI) in
primary care. In: Maruish M, editor. Handbook of psychological assessment in primary care
[31]
AC
settings. Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers; 2000. p. 297-334. Derogatis LR. Self-report symptom inventory. Internationale Skalen für Psychiatrie.
Scalaru CIP, editor. Weinheim: Beltz; 1986. [32]
Hessel A, Schumacher J, Geyer M. Symptomcheckliste SCL-90-R: Testtheoretische
Überprüfung und Normierung an einer bevölkerungsrepräsentativen Stichprobe. Diagnostica. 2001;47:27-39. [33]
Bremner JD, Bolus R, Mayer EA. Psychometric properties of the Early Trauma
Inventory-Self Report. J Nerv Ment Dis. 2007;195(3):211-8. [34]
Wingenfeld K, Driessen M, Mensebach C, Rullkoetter N, Schaffrath C, Spitzer C, et
al. Die deutsche Version des „Early Trauma Inventory” (ETI). Diagnostica. 2011;57(1):27-38.
25
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [35]
Heim C. Deutsche Version des Early Trauma Inventory: Inventar zur Erfassung früher
traumatischer Lebensereignisse (IFTL). 2000. [36]
Fritz CO, Morris PE, Richler JJ. Effect size estimates: current use, calculations, and
IP
[37]
T
interpretation. J Exp Psychol Gen. 2012;141(1):2-18.
Oostenveld R, Fries P, Maris E, Schoffelen JM. FieldTrip: Open source software for
SC R
advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci. 2011;2011:156869.
Junghöfer M, Peyk P. Analysis of electrical potentials and magnetic fields of the brain.
NU
[38]
Matlab Select 22004. p. 24–8.
Hamalainen MS, Ilmoniemi RJ. Interpreting magnetic fields of the brain: minimum
MA
[39]
norm estimates. Med Biol Eng Comput. 1994;32(1):35-42. Demartini B, Petrochilos P, Ricciardi L, Price G, Edwards MJ, Joyce E. The role of
D
[40]
TE
alexithymia in the development of functional motor symptoms (conversion disorder). J Neurol Neurosurg Psychiatry. 2014;85(10):1132-7. Gulpek D, Kelemence Kaplan F, Kesebir S, Bora O. Alexithymia in patients with
CE P
[41]
conversion disorder. Nordic journal of psychiatry. 2014;68(5):300-5. [42]
Kaplan MJ, Dwivedi AK, Privitera MD, Isaacs K, Hughes C, Bowman M.
AC
Comparisons of childhood trauma, alexithymia, and defensive styles in patients with psychogenic non-epileptic seizures vs. epilepsy: Implications for the etiology of conversion disorder. J Psychosom Res. 2013;75(2):142-6. [43]
Vuilleumier P, Richardson MP, Armony JL, Driver J, Dolan RJ. Distant influences of
amygdala lesion on visual cortical activation during emotional face processing. Nat Neurosci. 2004;7(11):1271-8. [44]
Ejareh dar M, Kanaan RAA. Uncovering the etiology of conversion disorder: insights
from functional neuroimaging. Neuropsychiatr Dis Treat. 2016;12:143-53. [45]
Voon V, Cavanna AE, Coburn K, Sampson S, Reeve A, LaFrance WC, Jr. Functional
Neuroanatomy and Neurophysiology of Functional Neurological Disorders (Conversion Disorder). J Neuropsychiatry Clin Neurosci. 2016:appineuropsych14090217. 26
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [46]
Steffen A, Fiess J, Schmidt R, Rockstroh B. "That pulled the rug out from under my
feet!" - adverse experiences and altered emotion processing in patients with functional neurological symptoms compared to healthy comparison subjects. BMC Psychiatry.
Roelofs K, Keijsers GP, Hoogduin KA, Naring GW, Moene FC. Childhood abuse in
IP
[47]
T
2015;15:133.
[48]
SC R
patients with conversion disorder. Am J Psychiatry. 2002;159(11):1908-13. Sar V, Akyüz G, Kundakçi T, Kiziltan E, Dogan O. Childhood trauma, dissociation,
and psychiatric comorbidity in patients with conversion disorder. Am J Psychiatry.
[49]
NU
2004;161(12):2271-6.
Sagaspe P, Schwartz S, Vuilleumier P. Fear and stop: a role for the amygdala in motor
[50]
MA
inhibition by emotional signals. Neuroimage. 2011;55(4):1825-35. Shibasaki H, Hallett M. What is the Bereitschaftspotential? Clin Neurophysiol.
TE
[51]
D
2006;117(11):2341-56.
Rushworth MF, Krams M, Passingham RE. The attentional role of the left parietal
CE P
cortex: the distinct lateralization and localization of motor attention in the human brain. J Cogn Neurosci. 2001;13(5):698-710. [52]
Rushworth MF, Johansen-Berg H, Gobel SM, Devlin JT. The left parietal and
[53]
AC
premotor cortices: motor attention and selection. Neuroimage. 2003;20 Suppl 1:S89-100. Laricchiuta D, Petrosini L, Picerni E, Cutuli D, Iorio M, Chiapponi C, et al. The
embodied emotion in cerebellum: a neuroimaging study of alexithymia. Brain structure & function. 2015;220(4):2275-87.
27
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS FUNDING
Research was supported by the Deutsche Forschungsgemeinschaft (STE 2263/2-1) and the
IP
T
University of Konstanz.
SC R
ACKNOWLEDGEMENTS
We thank David Schubring for his helpful advice on data analysis and Katharina Haag,
NU
Johanna Kienle, Ursula Lommen, Eva-Maria Schlichtmann and Sebastian Schuster for their support in data collection. Additionally, we thank the medical and psychotherapeutical teams
MA
of the Psychotherapeutic Neurology section, Kliniken Schmieder Konstanz and Gailingen for their support in patient recruitment. Thank you also to the Deutsche Forschungsgemeinschaft,
D
the University of Konstanz and the Lurija Institute of the Schmieder Foundation for making
AC
CE P
TE
this project possible.
28
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Highlights
Emotionally salient pictures modulate initial (110−330 ms) posterior neuromagnetic
T
responses.
This correlate of salience detection is similar in patients with FNS and controls.
In patients with FNS, sensorimotor areas are involved in this automatic salience
SC R
IP
AC
CE P
TE
D
MA
NU
processing.
29
S0022-3999(16)30434-2 doi:10.1016/j.jpsychores.2016.10.007 PSR 9229
To appear in:
Journal of Psychosomatic Research
Received date: Revised date: Accepted date:
19 February 2016 20 September 2016 20 October 2016
Please cite this article as: Fiess Johanna, Rockstroh Brigitte, Schmidt Roger, Wienbruch Christian, Steffen Astrid, Functional neurological symptoms modulate processing of emotionally salient stimuli, Journal of Psychosomatic Research (2016), doi:10.1016/j.jpsychores.2016.10.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Functional neurological symptoms modulate processing of emotionally salient stimuli
IP
T
Running title: Automatic emotion processing and FNS
Johanna Fiess, M.Sc., Department of Psychology, University of Konstanz, Germany,
SC R
[email protected]
Brigitte Rockstroh, Ph.D., Department of Psychology, University of Konstanz, Germany,
NU
[email protected]
Roger Schmidt, M.D., Neurological Rehabilitation Center Kliniken Schmieder, Konstanz,
MA
Germany, [email protected]
Christian Wienbruch, Ph.D., Department of Psychology, University of Konstanz, Germany, [email protected]
CE P
Corresponding author:
TE
[email protected]
D
Astrid Steffen, Ph.D., Department of Psychology, University of Konstanz, Germany,
Johanna Fiess, Department of Psychology, University of Konstanz,
AC
P.O. Box 905, D-78457 Konstanz, Germany Phone: +49-7531-884604, Fax: +49-7531-884601 E-mail: [email protected]
Keywords: functional neurological symptoms (FNS); affective pictures; rapid serial visual presentation (RSVP); magnetoencephalography (MEG); conversion; dissociative movement and sensibility disorders.
Number of tables: 2; number of figures: 3; supplementary materials: 2
1
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Abstract
Objective: Dysfunctional emotion processing has been discussed as a contributing factor to
T
functional neurological symptoms (FNS) in the context of conversion disorder, and refers to
IP
blunted recognition and the expression of one‟s own feelings. However, the emotion
SC R
processing components characteristic for FNS and/or relevant for conversion remain to be specified. With this goal, the present study targeted the initial, automatic discrimination of
NU
emotionally salient stimuli.
Methods: The magnetoencephalogram (MEG) was monitored in 21 patients with functional
MA
weakness and/or sensory disturbance subtypes of FNS and 21 healthy comparison participants (HC) while they passively watched 600 emotionally arousing, pleasant, unpleasant or neutral
D
stimuli in a rapid serial visual presentation (RSVP) design. Neuromagnetic activity was
CE P
central regions of interest.
TE
analyzed 110−330 ms following picture onset in source space for prior defined posterior and
Results: As early as 110 ms and across presentation interval, posterior neural activity modulation by picture category was similar in both groups, despite smaller initial (110−150
AC
ms) overall and posterior power in patients with FNS. The initial activity modulation by picture category was also evident in the left sensorimotor area in patients with FNS, but not significant in HC. Conclusions: Similar activity modulation by emotional picture category in patients with FNS and HC suggests that the fast, automatic detection of emotional salience is unchanged in patients with FNS, but involves an emotion-processing network spanning posterior and sensorimotor areas.
2
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Introduction
Altered emotion processing has been discussed as a factor that potentially contributes to
T
the generation of functional neurological symptoms (FNS) ever since FNS have been
IP
conceptualized in the context of conversion or dissociative disorders [cf. 1, 2-7]: Breuer and
SC R
Freud linked FNS to a conversion of intrapsychic distress into physical symptoms [2], and Janet linked psychoform and somatoform symptoms (like FNS) to the dissociation of psychobiological systems (thoughts, sensations, and behavior) consequent upon psychological
NU
trauma and individual predispositions [1, 3-5, 8]. Pavlov connected emotion processing and
MA
FNS when proposing that excessive cortical inhibitory control over (subcortically generated) emotional activities (centered in subcortical areas) flooded other neurological pathways,
D
thereby causing FNS−like paralysis or anesthesia [cf. 7, 9, 10].
TE
Within the framework of psychological emotion processing theories, emotion processing is defined by distinct processes such as the initial, automatic detection of emotional and salient
CE P
“ecologically important” [11, p. 773] stimuli and/or controlled processing, the conscious processing of one‟s own emotional responses or their regulation. Evaluating dysfunctional
AC
emotion processing in individuals with FNS indicated a different initial detection of emotional stimulus category: in an incidental affective task, patients with FNS responded with a greater amygdala activity to unpleasant stimuli (measured by functional magnetic resonance imaging, fMRI) and also showed less activity habituation across trials compared to healthy control participants [12]. Moreover, patients with FNS exhibited less modulation of amygdala activation by stimulus valence (unpleasant – pleasant) than HC [13]. Aybek et al. [12] further showed that emotionally salient stimuli prompted greater activation in the supplementary motor area (SMA) in patients with FNS than in controls, and a greater functional connectedness between amygdala and SMA [13, 14], thus, between brain areas associated with emotion processing and movement preparation.
3
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Referring to emotion regulation, we [15, this journal] found less frontocortical but more sensorimotor neuromagnetic activity in patients with FNS than in healthy control participants, when subjects activated trained strategies of cognitive reappraisal upon presentation of
IP
T
unpleasant pictures [16]. When subjects passively watched the unpleasant and neutral pictures
SC R
in a second condition within the same design, patients and controls displayed similar activity in posterior areas. While the latter activity was associated with the processing of the emotional salience of visual stimuli [11, 17, 18], sensorimotor activity was discussed as an
NU
indication of the close association of emotion and sensorimotor processing in patients with
MA
(dominant sensory and motor) FNS.
Complementing these findings, the present study scrutinized the initial, automatic discrimination of emotionally salient stimuli using a rapid serial visual presentation (RSVP)
TE
D
design to enforce automatic detection and high-resolution magnetoencephalography (MEG) to depict the fast cortical processing. The RSVP design has been shown to prompt greater
CE P
posterior activity to arousing pleasant and unpleasant compared to neutral stimuli as early as 150 ms after stimulus onset [17, 19-21] and has been used to study dysfunctional emotion
AC
processing in patients with posttraumatic stress disorder (PTSD), major depression disorder (MDD) or stress-related syndromes [20-22]. In the present study, neuromagnetic activity modulation by picture category was compared between patients with FNS and healthy comparison participants (HC) with the hypotheses that: (1) patients with FNS show enhanced processing of emotionally salient relative to neutral stimuli (referring to evidence from fMRI studies [12, 13]) in posterior areas associated with a visual emotional processing network [e.g. 17]); and that (2) patients with FNS display greater activity in sensorimotor areas than HC (referring to evidence of simultaneous activity in emotion- and movement-related areas in emotion recognition [12, 13] and emotion regulation [15] tasks). 4
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Methods Participants
T
The study involved 21 patients with FNS and 21 healthy comparison participants (HC).
IP
Samples overlapped with those reported in Fiess et al. [15; this journal]. Patients with an ICD
SC R
diagnosis of conversion disorder (ICD-10 codes F44.4, F44.6, F44.7) were recruited from the local rehabilitation center (Kliniken Schmieder) where they were following comprehensive treatment protocols involving individual and group psychotherapy, physiotherapy and
NU
occupational therapy. Diagnoses were given by two or more experienced psychiatrists and
MA
neurologists following standardized ICD-10 guidelines, with at least one core negative somatoform dissociative symptom (e.g., motor disorders, hypesthesia) required for a
D
diagnosis of conversion disorder. Symptom duration varied between 7 and 41.5 months
TE
(interquartile range) around the median of 12 months. Dominant symptoms were motor
AC
CE P
weakness and/or sensory disturbances on the left and/or right side of the body (see Table 1A).
5
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Table 1A: Dominant symptoms in patients with FNS (n = 21) Right-sided
Both sides
None
Motor weakness (n)
6
1
12
2
Sensory disturbances (n)
5
3
9
IP
T
Left-sided
4
Patients with
SC R
Table 1B: Sociodemographic and clinical information of study samples HC
vs. HC
21
21
Gender (f/m)
13/8
11/10
n. s.
Age (M±SD)
42±13.4
48±14.3
n. s.
Years schooling (M±SD)
11±3.2
11±1.6
n. s.
SDQ-20 (median (range))
33 (28–40)
21 (20–22.5)
U = 26, z = -4.9; p < 0.001, r = 0.76
1.4 (0.3–1.9)
0.1 (0–0.3)
U = 82.5, z = -3.5; p < 0.001, r = 0.54
8 (0.5–24)
0 (0–4)
U = 108, z = -2.9; p < 0.01, r = 0.45
213 (15–585)
34 (8–75.5)
U = 78, z = -2.6; p < 0.01, r = 0.39
18
20
Left-handed
1
1
Ambidextrous
2
0
D
CE P
PSSI (median (range)) ETI (median (range)) Handedness (n)
AC
Right-handed
TE
SCL-90-R (median (range)) Depression
MA
n
NU
FNS
Patients with FNS
n. s.
Note. FNS = patients with functional neurological symptoms; HC = healthy control participants; SDQ20 = Somatoform Dissociation Questionnaire (scores range from 20−100); SCL-90-R = 90-item Symptom Checklist (scores range from 0−4); PSSI = Posttraumatic Stress Scale-Interview severity score (scores range from 0−86); ETI = Early Trauma Inventory.
6
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
HC were recruited from the local community through flyers and verbal advertising in order to be demographically comparable to the patient sample. HC were screened using the Mini International Neuropsychiatric Interview [MINI; 23] to exclude any psychiatric disorders.
IP
T
From n = 24 individuals screened with the MINI, n = 3 were excluded because of a
SC R
hypomanic episode in the past, a current major depressive episode, and alcohol abuse, respectively. Recruitment was continued until the control group consisted of n = 21 participants. As evident from Table 1B, groups did not differ with respect to mean age, gender
NU
distribution or years of school education. Patients with FNS and HC with central nervous lesions (e.g., degenerative disorders, tumors) were not included. For patients with FNS, this
MA
screening was accomplished upon admission following a standard protocol that included screening for neuropathology, clinical structural MRI and electromyography as clinically
D
indicated. All participants had normal or corrected-to-normal vision. The Fisher-Freeman-
TE
Halton test (an extension of the Fisher exact test for r x c contingency tables) confirmed that
CE P
groups did not differ significantly in handedness, as assessed using the Edinburgh Handedness Inventory [24; see Table 1B].
AC
Setting
The study protocol was approved by the ethics committee of the University of Konstanz. Participants were informed about the design and procedures and provided their written informed consent. Thereafter, demographic and self-report data (see below) were assessed. A separate session comprised the rapid serial visual presentation (RSVP) protocol [17] while the MEG was being monitored. Stimulus material For the RSVP, highly arousing pleasant (n = 100), unpleasant (n = 100) and neutral pictures (n = 100) were chosen from the International Affective Picture System [IAPS; 25] on the basis of their normative ratings. Brightness, contrast and color spectra of the stimuli were 7
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
matched across picture categories. Each stimulus was presented once within each of two series of 300 pictures (total 600 stimuli) without a perceivable gap for 333 ms each (3 Hz, 60 Hz refresh rate) in a pseudorandom sequence. The two picture series were presented without a
IP
T
break, the presentation time was approximately 4 minutes. Participants were instructed to
SC R
maintain their focus on a small, centrally located fixation cross overlaying each picture and to watch the sequence of pictures without any specific task. Data assessment and analysis
NU
Clinical data: The severity of somatoform dissociation symptoms was determined using
MA
the Somatoform Dissociation Questionnaire [SDQ-20; 26, German version by 27]. Participants were further characterized by comorbid psychopathology using the Posttraumatic
D
Stress Scale-Interview [PSSI; 28, 29] for screening trauma-related symptoms, and the German
TE
version of the 90-item Symptom Checklist [SCL-90-R; 30, 31, 32]. Moreover, adverse childhood experiences, which have been related to psychopathology, were screened using the
CE P
German version of the Early Trauma Inventory [ETI; 33, German version by 34, see 35]. Given that self-report data of HC were not normally distributed, group differences were
AC
evaluated with non-parametric Mann-Whitney U-tests reporting exact significance values. Effect sizes were calculated as suggested by Fritz et al. [36, p. 12] by the formula r = z / √N. MEG data acquisition and analysis: Magnetic fields were continuously recorded using a 148-channel whole-head magnetometer (MAGNES 2500 WH, 4D Neuroimaging, San Diego, USA) with a sampling rate of 678.16 Hz, and filtered with an analog bandpass filter from 0.1 to 200 Hz. The participant‟s head position was monitored using five coils (nasion, inion, Cz, left and right ear canal) and headshape, digitized with a Polhemus 3Space® Fasttrack. Signals recorded by eleven MEG reference sensors were used to remove external, non-biological noise.
8
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Preprocessing of the MEG data was conducted with FieldTrip, a MATLAB-based toolbox [37]. Trials containing movement artifacts and superconducting quantum interference device (SQUID) jumps were rejected based on visual inspection. Five dysfunctional sensors were
IP
T
excluded from further analysis. Independent component analysis (ICA) was used to visually
SC R
detect eye or cardiac artifacts, which were subsequently removed from the data. Frequencies above 40 Hz were filtered with a FIR low pass filter using a Kaiser window. The number of artifact-free trials retained did not differ between groups or picture category (means ±
NU
standard deviation, FNS: neutral – 196.7±2.78, unpleasant – 197±3.3, pleasant – 197.1±2.19; HC: neutral – 196±2.3, unpleasant – 196.3±2.72, pleasant – 197.1±3). Average event-related
MA
magnetic fields were calculated for each picture category and subject. Source analysis. The L2 minimum-norm was calculated based on a one-shell spherical
D
head model with evenly distributed 2 (azimuthal and polar direction) × 350 dipoles as a
TE
source model using the MATLAB-based software EMEGS© [38]. The L2 minimum-norm
CE P
estimate enables an enhanced resolution of brain activations generating the magnetic field without a priori assumptions regarding the location and number of current sources [39]. This
AC
calculation was based on information on the center of a sphere fitted to the digitized head shape and the positions of the MEG sensors relative to the head. A source shell radius of 87% of the spherical volume conductor head radius was chosen, which roughly corresponded with the grey matter volume. Across all participants and conditions, the Tikhonov regularization parameter λ was set to 0.2. Source data were used for hypothesis testing: In a first step, the global field power of the L2 minimum-norm across time was plotted separately for each group and each stimulus category. In a second step, time windows and regions distinguishing groups and stimulus categories were determined based on (a) previous evidence of dominant activity modulation by stimulus valence in posterior cortical regions [17, 19, 21], (b) on previous reports of
9
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
sensorimotor activity distinguishing patients with FNS and controls in an emotion regulation task [15], and (c) on visual inspection (see also Figure 2A, 3B, and Supplementary Figure). Because previous research indicated a right-hemispheric dominance of activity modulation by
IP
T
picture category in healthy controls [e.g. 17, 21], hemisphere (left vs. right) was taken into
SC R
account as additional factor.
The modulation of neural activity by group and by stimulus category was evaluated in factorial repeated-measures ANOVAs with the between-subject factor Group (patients with
NU
FNS vs. HC) and the within-subject factor Emotion (pleasant vs. neutral vs. unpleasant
MA
stimuli) separately for the three time windows of interest (110–150 ms, 150–220 ms, and 220–300 ms). In three separate ANOVAs dependent variables of neural activity were overall activity (global field power), and activity in the two regions of interest (ROIs), the latter
D
including the within-subject factor Hemisphere (left vs. right). For interactions, degrees of
TE
freedom were corrected using Greenhouse-Geisser‟s estimates of sphericity where
CE P
appropriate; adjusted p values are reported. Confounds by multiple comparisons were controlled by Bonferroni-corrected post-hoc analyses. A potential confound by comorbid
AC
symptoms of depression, trauma-related symptoms, amount of adverse childhood experiences and the modulation by stimulus salience was evaluated by Spearman's rank-order correlations including the individual questionnaire scores (depression score of the SCL-90-R, PSSI symptom score, amount of adverse childhood experiences, ETI) and the overall neural activity across all emotional picture categories (pleasant, neutral and unpleasant). In addition (in patients with FNS), the questionnaire scores were considered as covariates of no interest in the repeated measures ANOVA.
10
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Results
Figure 1 illustrates stimulus-evoked brain activation as global field power across time
T
separately for the two groups and the three picture categories. Activity differences between
IP
groups and stimulus conditions are evident in the three time windows (110–150 ms, 150–220
SC R
ms and 220–330 ms, see Table 2A). Table 2 summarizes statistical effects for overall activity
AC
CE P
TE
D
MA
NU
(Table 2A), the posterior, and the central ROI (Table 2B, 2C).
11
ACCEPTED MANUSCRIPT
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
D
Figure 1: Global field power averaged separately for patients with FNS (functional
TE
neurological symptoms; black) and HC (healthy comparison participants; grey) as well as for the three picture categories pleasant (dotted), neutral (dash-dotted) and unpleasant (solid).
CE P
Grey bars mark the three time windows selected for analyses (110−150 ms, 150−220 ms, and
AC
220 −330 ms after stimulus onset).
12
ACCEPTED MANUSCRIPT Automatic emotion processing and FNS
CR
IP
T
Table 2: Statistical effects on neural activity in the three latency windows, separately for the global power, the posterior and central regions of interest. 110–150 150–220 220–330 F(1,40) = 5.9; p = 0.02 n. s. n. s.
B: Posterior areas (FNS x HC) Group Emotion Hemisphere Group x Emotion Group x Hemisphere Emotion x Hemisphere Group x Emotion x Hemisphere
F(1,40) = 4.28; p = 0.04 F(1.57,62.95) = 3.11; p = 0.05 n. s. n. s. n. s. n. s. n. s.
C: Central areas (FNS x HC) Group Emotion Hemisphere Group x Emotion Group x Hemisphere Emotion x Hemisphere Group x Emotion x Hemisphere
F(1,40) = 4.72; p = 0.04 F(2,80) = 3.68; p = 0.03 F(1,40) = 10.05; p < 0.01 n. s. n. s. F(2,80) = 4.18; p = 0.02 n. s.
n. s. F(2,80) = 6.18; p < 0.001 n. s.
n. s. F(1.54,61.72) = 5.94; p < 0.01 F(1,40) = 7.02; p = 0.01 n. s. n. s. n. s. n. s.
n. s. F(2,80) = 6.97; p < 0.01 n. s. n. s. n. s. F(2,80) = 3.71; p = 0.03 n. s.
n. s. n. s. F(1,40) = 20.99; p < 0.001 n. s. n. s. n. s. n. s.
n. s. n. s. F(1,40) = 15.68; p < 0.001 n. s. n. s. n. s. n. s.
MA N
TE D
CE P
AC
n. s. F(2,80) = 2.63; p = 0.08 n. s.
US
A: Global power (FNS x HC) Group Emotion Group x Emotion
Note. FNS = functional neurological symptoms; HC = healthy comparison participants.
13
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
As evident in Figure 1, patients with FNS showed smaller overall activity than HC mainly in the initial time window. This was confirmed by a main effect Group in the initial (110-150 ms) interval, while groups did not differ in the two subsequent time windows (see Table 2A).
IP
T
Table 2 further indicates that modulation of activity by stimulus category is area-specific (in
SC R
the posterior ROI) and only marginally evident in global activity. In line with previous evidence [17, 21], main effects Emotion are verified for the posterior ROI, but evident in global activity only in the 220−330 ms time window.
NU
Falsifying hypothesis (1), the modulation of dipole activity by emotional picture content in the posterior ROI was similar in both groups, that is, the Group x Emotion interaction did not
MA
reach significance (p > .05; main effects Emotion in all three time windows p ≤ .05, see Table 2B and Figure 2A). In line with previous results [17, 21], a right-hemispheric dominance of
D
posterior dipole activity was verified for the 150−220 ms interval independent of group and
TE
stimulus content (main effect Hemisphere, p ≤ .05; see Table 2B), and a greater modulation
CE P
by stimulus content in the right compared to the left hemisphere in the 220−300 ms window (Emotion x Hemisphere interaction; see Table 2B). Planned group-specific analyses verified
AC
the modulation of dipole activity by emotional picture content (main effect Emotion) for patients with FNS (F(2,40) = 5.28; p = 0.01) and HC (F(2,40) = 4.45; p = 0.02) in the late time window (220−330 ms), and additionally for HC in the intermediate time window (150−220 ms, F(2,40) = 4.15; p = 0.02).
14
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
Figure 2A: Upper rows: Statistical parametric topographical maps of the main effect Emotion (pleasant, neutral, and unpleasant picture category) separately averaged over all participants for the three time windows (110–150 ms, 150–220 ms and 220–330 ms following stimulus onset). Lower row: Schematic layout of the dipole distribution in the posterior areas used for statistical analyses per time window. B Estimated neural strength (ordinate: dipole activity expressed as minimum norm estimates, mean ± standard error in nA/m) in posterior areas plotted separately for groups (grey: patients with functional neurological symptoms, FNS; black: healthy comparison participants, HC) and picture category (pleasant, neutral, unpleasant) within each of the three time windows (abscissa) as well as for the left and right hemisphere. 15
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Referring to hypothesis (2), Figure 3 illustrates the modulation of neural activity by emotional picture category in the central ROI. The group difference reported for initial overall
T
activity (110−150 ms) was indicated for the central ROI (main effect Group, Table 2C and
IP
Figure 3). However, initial activity modulation was evident only in the left-hemispheric
SC R
central ROI (main effect Emotion, Emotion x Hemisphere interaction, see Table 2C and Figure 3B). Although the Group x Emotion interaction did not reach significance (F(2,80) = 1.26; p = .29), planned comparisons verified the prominent left-hemispheric neural activity
NU
modulation for patients with FNS (F(2,40) = 6.47; p < 0.01), but not for HC (F(2,40) = 1.65;
MA
p = .24).
A potential impact of comorbid psychopathology on the reported neural activity was
D
evaluated by correlation analyses and repeated measures ANOVA. Patients with FNS
TE
experienced more somatoform dissociation, more comorbid psychopathology including trauma-related symptoms, and a greater childhood stress load than HC (see Table 1B and
CE P
Supplementary Table). Yet, neither depression, trauma-related symptoms scores nor childhood stress load revealed significant correlations with the global power (p > .1) or ROI
AC
specific activity (p > .1). Similarly, controlling for comorbid symptoms and childhood stress load as factor in the respective ANOVAs in patients with FNS had no impact on the Emotion main effect (F(2,43) = 4.71; p = 0.02). These results do not indicate a major influence of comorbid psychopathology on the reported neural activity modulation.
16
ACCEPTED MANUSCRIPT
AC
CE P
TE
D
MA
NU
SC R
IP
T
Automatic emotion processing and FNS
Figure 3: A Estimated neural strength (ordinate: dipole activity expressed as minimum norm estimates, mean ± standard error in nA/m) in the central region of interest (upper right corners: schematic layout of the dipole distribution per hemisphere used for statistical analyses) plotted separately for groups (grey: patients with functional neurological symptoms, FNS; black: healthy comparison participants, HC) and picture category (pleasant, neutral, unpleasant) within each of the three time windows (abscissa: 110–150 ms, 150–220 ms and 220–330 ms following picture onset) as well as for the left and right hemisphere. B Topographies representing statistical parametric maps of the contrasts between picture categories (pleasant-neutral and unpleasant-neutral) per group (patients with FNS and HC) in the time window 110–150 ms following picture onset. 17
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS Discussion
Dysfunctional emotion processing has been discussed as a contributing factor to functional neurological symptoms [FNS; 12-14, 40, 41, 42], but the specific process or
IP
T
processes that are dysfunctional in or relevant for FNS remain unclear. The present study
SC R
addressed the capability of patients with FNS to automatically distinguish emotionally salient and neutral stimuli in the rapid flow of visual stimuli with the hypothesis that cortical signs of this initial recognition of emotionally salient information differ from the pattern usually
NU
observed in healthy individuals, suggesting abnormal emotion processing at a very early
MA
processing stage.
Falsifying the hypothesis (1) of enhanced neuromagnetic activity in response to
D
emotionally salient (relative to neutral) pictures, patients with FNS and HC exhibited similar
TE
activity modulation by emotional picture category in posterior areas within 110−330 ms following stimulus onset. Whereas fMRI studies reported augmented subcortical (amygdala)
CE P
responses to emotionally salient stimuli [12, 13], the present MEG-study added evidence that cortical responses to visual emotional input within the first 110−330 ms distinguish
AC
emotionally salient from neutral stimuli in patients with FNS and healthy controls alike. Whether and in which way such early and fast cortical responding was related to subcortical (amygdala) activation could not be determined from the MEG-measurement with sufficient precision. A relationship might be assumed, as Vuilleumier et al. [43] showed that visual cortex activation by different facial expressions (sad, happy) can be modified by amygdala activity, potentially via fast feedback loops [see also 11]. The present results suggest that activity modulation in a network associated with the processing of visual, emotionally salient input [11] did not seem dysfunctional in the patients with FNS compared to HC. Initial activity modulation by emotional salience was evident, although evoked activity in posterior areas associated with emotion processing 110–150 ms following picture onset was 18
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
lower in patients with FNS than in HC. This suggests less or slower responsiveness of the visual system or a network associated with processing of emotional salience in patients.
T
Posterior activity may reflect information transfer along the ventral and dorsal path to an
IP
emotion processing network, in which thalamic nuclei (particularly the pulvinar) project to
SC R
amygdala and cortical regions like insula, cingulate, frontal, parietal cortex, and the precuneus at the transition between the parietal and central cortex [11, 18]. Reduced thalamic activity has been reported in patients with FNS compared to healthy controls [cf. 44, 45]. Thus, lower
NU
activity may indicate slower activation by visual input. Importantly, this smaller activity
MA
increase was restricted to the initial response, while there was no group difference in later activity up to 330 ms. This argues against generally reduced thalamic pulse or the lasting impact of this pulse in FNS. Importantly, the lower initial response in patients with FNS did
TE
D
not affect activity modulation by emotional salience, thus, distinction of salient stimuli succeeded in patients with FNS even at a lower activity level. Other factors might have
CE P
contributed to the blunted posterior activity: Blunted responses have been reported for patients with major depression disorder posttraumatic stress disorder, and/or in syndromes
AC
related to adverse childhood experiences [20-22]. More intense depression and higher stress load were reported in patients with FNS [44-48]. However, neither posttraumatic and depressive symptoms nor overall childhood stress load were significantly associated with the reported lower initial posterior activity and the modulation by emotional picture category. On the contrary to the assumed greater activity in sensorimotor areas in patients with FNS compared to HC (hypothesis (2)), both groups did not differ concerning to this manner in the left central ROI, but patients with FNS showed even less activity in the right central ROI. However, patients with FNS displayed activity modulation by emotional picture category in left sensorimotor areas as early as 110−150 ms following stimulus onset. Although interaction with the factor Group did not reach statistical significance, planned comparisons indicate 19
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
prominent left-sensorimotor activity modulation by picture category in patients with FNS (but not in HC). This provides an indication of an involvement of sensorimotor regions during
T
emotion processing, which can be seen in line with previous studies [12, 13, 15]. Thus,
IP
sensorimotor activity modulation may index genuine sensorimotor involvement in emotion
SC R
processing independent of the specific processing, automatic salience detection or controlled emotion regulation.
NU
The involvement of movement-related cortical areas such as SMA while emotion processing tasks has been demonstrated in patients with conversion disorder [12-14] and may
MA
indicate the activation of a cortical network spanning systems related to emotion processing and to the preparation of complex movements and motor inhibition [49, 50]. For conversion
D
disorders, it has been hypothesized that such movement-related processes are augmented in
TE
order to ward off (painful) emotional responses, which may result in the perception of symptoms [1, 14]. In our previous evaluation of neuromagnetic activity during an emotional
CE P
regulation task in a sample overlapping with the one of the present study [15], patients with FNS displayed greater sensorimotor parallel to lower frontocortical activity than controls.
AC
These results were discussed as being a potential manifestation of a dysfunctional activity shift from emotion-related to movement-related brain regions. The similarity of results between the two studies or tasks may indicate the involvement and co-activation of sensorimotor areas in FNS independent of the level of attention and control devoted to emotion processing. A reduced response in regions associated with automatic emotion detection together with the activity modulation by emotional picture category in sensorimotor regions could be conceived of as an emotional defense response, which then prompts a „conversion‟ of emotional processing, for example, to sensorimotor systems. A major impact of carry-over effects is unlikely, as the sequence of the tasks scheduled within one session varied across subjects. 20
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
It is tempting to consider the left-sensorimotor activity modulation in patients with FNS as (sign of) involvement of the hemisphere contralateral to the affected limb. However, no
T
dominance of right-sided motor weakness and/or sensory disturbances was evident in the
IP
present patient sample. Moreover, handedness did not differ between patients with FNS and
SC R
HC. The left-central focus of activity extended to the left anterior parietal region, which has been related to motor attention [51, 52]. For healthy individuals, sensorimotor activity has also been discussed as being „embodied‟ emotion – that is, as co-activation of motor systems
NU
associated with facial or limb movements involved in the expression of an emotional
MA
response, even if this response is not executed [53]. Augmented or altered embodied emotion has rarely been addressed in the context of conversion disorder or FNS, and the present results cannot resolve that dysfunctional embodied emotion affected the sensorimotor activity pattern
TE
D
in patient with FNS patients as a result of some conversion process. The limitations of the present study must also be considered. The sample only included
CE P
patients with at least one core negative functional symptom (i.e., a motor and/or sensory deficit), and no patients with e.g. functional seizures − thus limiting the applicability of the
AC
conclusions of the present study to FNS in general. These specific inclusion criteria further limited the sample size, precluding subgroup analyses for functional motor or sensory symptoms. The latter distinction should inform the specific nature of emotion processing in patients with specific FNS. While the present results inform the fast, initial and automatic detection of emotionally salient stimuli, they do not specify the subcortical-cortical network supposed to be involved in this automatic processing, so that the functional impact of (dys)functional automatic detection in the course of emotion processing cannot be identified. Additionally, this study cannot address whether these differences in emotion processing (reduced posterior neural activity and early emotion discrimination in sensorimotor regions) reflect the patient‟s current state or whether this is a longer-term trait. To answer this 21
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
question, longitudinal studies are needed to chart and monitor potential changes over time, especially before and after successful treatment.
T
In conclusion, the present study adds to the understanding of altered emotion processing in
IP
patients with FNS, verifying the functional automatic discrimination of emotional stimuli and
AC
CE P
TE
D
MA
NU
SC R
emphasizing the involvement of the sensorimotor cortex in emotion processing.
22
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [1]
Crommelinck M. Neurophysiology of conversion disorders: a historical perspective.
Neurophysiol Clin. 2014;44(4):315-21. Breuer J, Freud S. Studien über Hysterie. Leipzig and Vienna: Deutike; 1895.
[3]
Janet P. The major symptoms of hysteria. New York: The Macmillan Company; 1907.
[4]
Kozlowska K. Healing the disembodied mind: contemporary models of conversion
disorder. Harv Rev Psychiatry. 2005;13(1):1-13. [5]
SC R
IP
T
[2]
Vuilleumier P. Brain circuits implicated in psychogenic paralysis in conversion
Ban TA. Conditioning Behavior and Psychiatry. New Brunswick (U.S.A.) and London
MA
[6]
NU
disorders and hypnosis. Neurophysiol Clin. 2014;44(4):323-37.
(U.K.): AldineTransaction; 2008. [7]
Todes DP. Ivan Pavlov: A Russian Life in Science. New York: Oxford University
TE
[8]
D
Press; 2014.
Van der Hart O, Nijenhuis ERS, Steele K. The Haunted Self: Structural Dissociation
Company; 2006. [9]
Vuilleumier P. Hysterical conversion and brain function. Prog Brain Res.
AC
2005;150:309-29. [10]
CE P
and the Treatment of Chronic Traumatization. New York and London: W. W. Norton &
Windholz G. Pavlov and the mind-body problem. Integr Physiol Behav Sci.
1997;32(2):149-59. [11]
Pessoa L, Adolphs R. Emotion processing and the amygdala: from a „low road‟ to
„many roads‟ of evaluating biological significance. Nature reviews Neuroscience. 2010;11(11):773-83. [12]
Aybek S, Nicholson TR, O'Daly O, Zelaya F, Kanaan RA, David AS. Emotion-motion
interactions in conversion disorder: an FMRI study. PLoS One. 2015;10(4):e0123273. [13]
Voon V, Brezing C, Gallea C, Ameli R, Roelofs K, LaFrance WC, Jr., et al. Emotional
stimuli and motor conversion disorder. Brain. 2010;133(Pt 5):1526-36.
23
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [14]
Aybek S, Nicholson TR, Zelaya F, et al. Neural correlates of recall of life events in
conversion disorder. JAMA Psychiatry. 2014;71(1):52-60. [15]
Fiess J, Rockstroh B, Schmidt R, Steffen A. Emotion regulation and functional
T
neurological symptoms: Does emotion processing convert into sensorimotor activity? J
Moser JS, Krompinger JW, Dietz J, Simons RF. Electrophysiological correlates of
SC R
[16]
IP
Psychosom Res. 2015;79(6):477-83.
decreasing and increasing emotional responses to unpleasant pictures. Psychophysiology.
[17]
NU
2009;46(1):17-27.
Junghöfer M, Bradley MM, Elbert TR, Lang PJ. Fleeting images: A new look at early
[18]
MA
emotion discrimination. Psychophysiology. 2001;38(2):175-8. Habel U, Klein M, Kellermann T, Shah NJ, Schneider F. Same or different? Neural
Peyk P, Schupp HT, Elbert T, Junghöfer M. Emotion processing in the visual brain: a
TE
[19]
D
correlates of happy and sad mood in healthy males. Neuroimage. 2005;26(1):206-14.
MEG analysis. Brain Topogr. 2008;20(4):205-15. Elbert T, Schauer M, Ruf M, Weierstall R, Neuner F, Rockstroh B, et al. The Tortured
CE P
[20]
Brain. Z Psychol. 2011;219(3):167-74. [21]
Matz K, Junghöfer M, Elbert T, Weber K, Wienbruch C, Rockstroh B. Adverse
AC
experiences in childhood influence brain responses to emotional stimuli in adult psychiatric patients. Int J Psychophysiol. 2010;75(3):277-86. [22]
Zwanzger P, Klahn AL, Arolt V, Ruland T, Zavorotnyy M, Sälzer J, et al. Impact of
electroconvulsive therapy on magnetoencephalographic correlates of dysfunctional emotional processing in major depression. Eur Neuropsychopharmacol. 2016;26(4):684-92. [23]
Ackenheil M, Stotz G, Dietz-Bauer R, Vossen A. Mini International Neuropsychiatric
Interview (German Version 5.0.0). München, Germany: Psychiatrische Universitätsklinik München; 1998. [24]
Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory.
Neuropsychologia. 1971;9(1):97-113.
24
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [25]
Lang PJ, Bradles MM, Cuthbert BN. International affective picture system (IAPS):
Instruction manual and affective ratings (Tech. Rep. No. A-6). Gainesville, FL: University of Florida, The Center for Research in Psychophysiology; 2005. Nijenhuis ERS, Spinhoven P, Van Dyck R, van der Hart O, Vanderlinden J. The
T
[26]
IP
development and psychometric characteristics of the Somatoform Dissociation Questionnaire
[27]
SC R
(SDQ-20). J Nerv Ment Dis. 1996;184(11):688-94.
Mueller-Pfeiffer C, Schumacher S, Martin-Soelch C, Pazhenkottil AP, Wirtz G,
Fuhrhans C, et al. The validity and reliability of the German version of the Somatoform
NU
Dissociation Questionnaire (SDQ-20). Journal of trauma & dissociation : the official journal of the International Society for the Study of Dissociation (ISSD). 2010;11(3):337-57. Foa EB, Tolin DF. Comparison of the PTSD Symptom Scale-Interview Version and
MA
[28]
the Clinician-Administered PTSD scale. J Trauma Stress. 2000;13(2):181-91. Foa EB, Cashman L, Jaycox L, Perry K. The validation of a self-report measure of
D
[29]
[30]
CE P
1997;9(4):445-151.
TE
posttraumatic stress disorder: The Posttraumatic Diagnostic Scale. Psychol Assess.
Derogatis LR, Savitz KL. The SCL–90–R and Brief Symptom Inventory (BSI) in
primary care. In: Maruish M, editor. Handbook of psychological assessment in primary care
[31]
AC
settings. Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers; 2000. p. 297-334. Derogatis LR. Self-report symptom inventory. Internationale Skalen für Psychiatrie.
Scalaru CIP, editor. Weinheim: Beltz; 1986. [32]
Hessel A, Schumacher J, Geyer M. Symptomcheckliste SCL-90-R: Testtheoretische
Überprüfung und Normierung an einer bevölkerungsrepräsentativen Stichprobe. Diagnostica. 2001;47:27-39. [33]
Bremner JD, Bolus R, Mayer EA. Psychometric properties of the Early Trauma
Inventory-Self Report. J Nerv Ment Dis. 2007;195(3):211-8. [34]
Wingenfeld K, Driessen M, Mensebach C, Rullkoetter N, Schaffrath C, Spitzer C, et
al. Die deutsche Version des „Early Trauma Inventory” (ETI). Diagnostica. 2011;57(1):27-38.
25
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [35]
Heim C. Deutsche Version des Early Trauma Inventory: Inventar zur Erfassung früher
traumatischer Lebensereignisse (IFTL). 2000. [36]
Fritz CO, Morris PE, Richler JJ. Effect size estimates: current use, calculations, and
IP
[37]
T
interpretation. J Exp Psychol Gen. 2012;141(1):2-18.
Oostenveld R, Fries P, Maris E, Schoffelen JM. FieldTrip: Open source software for
SC R
advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci. 2011;2011:156869.
Junghöfer M, Peyk P. Analysis of electrical potentials and magnetic fields of the brain.
NU
[38]
Matlab Select 22004. p. 24–8.
Hamalainen MS, Ilmoniemi RJ. Interpreting magnetic fields of the brain: minimum
MA
[39]
norm estimates. Med Biol Eng Comput. 1994;32(1):35-42. Demartini B, Petrochilos P, Ricciardi L, Price G, Edwards MJ, Joyce E. The role of
D
[40]
TE
alexithymia in the development of functional motor symptoms (conversion disorder). J Neurol Neurosurg Psychiatry. 2014;85(10):1132-7. Gulpek D, Kelemence Kaplan F, Kesebir S, Bora O. Alexithymia in patients with
CE P
[41]
conversion disorder. Nordic journal of psychiatry. 2014;68(5):300-5. [42]
Kaplan MJ, Dwivedi AK, Privitera MD, Isaacs K, Hughes C, Bowman M.
AC
Comparisons of childhood trauma, alexithymia, and defensive styles in patients with psychogenic non-epileptic seizures vs. epilepsy: Implications for the etiology of conversion disorder. J Psychosom Res. 2013;75(2):142-6. [43]
Vuilleumier P, Richardson MP, Armony JL, Driver J, Dolan RJ. Distant influences of
amygdala lesion on visual cortical activation during emotional face processing. Nat Neurosci. 2004;7(11):1271-8. [44]
Ejareh dar M, Kanaan RAA. Uncovering the etiology of conversion disorder: insights
from functional neuroimaging. Neuropsychiatr Dis Treat. 2016;12:143-53. [45]
Voon V, Cavanna AE, Coburn K, Sampson S, Reeve A, LaFrance WC, Jr. Functional
Neuroanatomy and Neurophysiology of Functional Neurological Disorders (Conversion Disorder). J Neuropsychiatry Clin Neurosci. 2016:appineuropsych14090217. 26
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS [46]
Steffen A, Fiess J, Schmidt R, Rockstroh B. "That pulled the rug out from under my
feet!" - adverse experiences and altered emotion processing in patients with functional neurological symptoms compared to healthy comparison subjects. BMC Psychiatry.
Roelofs K, Keijsers GP, Hoogduin KA, Naring GW, Moene FC. Childhood abuse in
IP
[47]
T
2015;15:133.
[48]
SC R
patients with conversion disorder. Am J Psychiatry. 2002;159(11):1908-13. Sar V, Akyüz G, Kundakçi T, Kiziltan E, Dogan O. Childhood trauma, dissociation,
and psychiatric comorbidity in patients with conversion disorder. Am J Psychiatry.
[49]
NU
2004;161(12):2271-6.
Sagaspe P, Schwartz S, Vuilleumier P. Fear and stop: a role for the amygdala in motor
[50]
MA
inhibition by emotional signals. Neuroimage. 2011;55(4):1825-35. Shibasaki H, Hallett M. What is the Bereitschaftspotential? Clin Neurophysiol.
TE
[51]
D
2006;117(11):2341-56.
Rushworth MF, Krams M, Passingham RE. The attentional role of the left parietal
CE P
cortex: the distinct lateralization and localization of motor attention in the human brain. J Cogn Neurosci. 2001;13(5):698-710. [52]
Rushworth MF, Johansen-Berg H, Gobel SM, Devlin JT. The left parietal and
[53]
AC
premotor cortices: motor attention and selection. Neuroimage. 2003;20 Suppl 1:S89-100. Laricchiuta D, Petrosini L, Picerni E, Cutuli D, Iorio M, Chiapponi C, et al. The
embodied emotion in cerebellum: a neuroimaging study of alexithymia. Brain structure & function. 2015;220(4):2275-87.
27
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS FUNDING
Research was supported by the Deutsche Forschungsgemeinschaft (STE 2263/2-1) and the
IP
T
University of Konstanz.
SC R
ACKNOWLEDGEMENTS
We thank David Schubring for his helpful advice on data analysis and Katharina Haag,
NU
Johanna Kienle, Ursula Lommen, Eva-Maria Schlichtmann and Sebastian Schuster for their support in data collection. Additionally, we thank the medical and psychotherapeutical teams
MA
of the Psychotherapeutic Neurology section, Kliniken Schmieder Konstanz and Gailingen for their support in patient recruitment. Thank you also to the Deutsche Forschungsgemeinschaft,
D
the University of Konstanz and the Lurija Institute of the Schmieder Foundation for making
AC
CE P
TE
this project possible.
28
ACCEPTED MANUSCRIPT
Automatic emotion processing and FNS
Highlights
Emotionally salient pictures modulate initial (110−330 ms) posterior neuromagnetic
T
responses.
This correlate of salience detection is similar in patients with FNS and controls.
In patients with FNS, sensorimotor areas are involved in this automatic salience
SC R
IP
AC
CE P
TE
D
MA
NU
processing.
29