Original articleAttention to the body in nonclinical somatoform dissociation depends on emotional state
Introduction
Physical symptoms that lack an identifiable medical cause [i.e., “medically unexplained symptoms” (MUS)] are common in medical settings [1] and associated with high levels of distress, disability, and resource utilization [2], [3]. While MUS can be associated with any bodily system, the worst cases often have an unusually high preponderance of unexplained neurological symptoms, such as motor (e.g., paralysis), paroxysmal (e.g., convulsions), and sensory complaints (e.g., blindness [4], [5]). Current taxonomies disagree in their classification of such “pseudoneurological” symptoms, reflecting ongoing uncertainty about the nature of these conditions [6], [7]. One influential model spans existing taxonomies by describing pseudoneurological symptoms as examples of “somatoform dissociation” [8].
Following Janet [9], most theorists suggest that disturbances in attention play a central role in unexplained neurological symptoms. Ludwig [10] argues for an impairment in attention characterized by top-down inhibition of afferent stimuli, whereas Kihlstrom [11] suggests that a disturbance in high- but not low-level attentional processes is responsible. More recently, Brown [12], [13] has argued that many unexplained neurological symptoms are disturbances in bodily awareness and control resulting from the overactivation of somatic representations in memory. In this account, attentional hypervigilance for symptom-related information and difficulties disengaging attention from somatic representations both contribute to the creation and maintenance of these symptoms. In many models, exposure to traumatic events is thought to contribute to these attentional vulnerabilities.
Attentional processes are also central in cognitive behavioral accounts of MUS more generally. Excessive body-focused attention (somatic hypervigilance) is thought to increase the salience of benign bodily sensations, causing them to be misinterpreted as evidence of serious illness; the resulting physiological and behavioral changes lead to further physical sensations, creating a vicious cycle [14]. The term “somatosensory amplification” refers to this process of focusing on and misinterpreting somatic sensations [15], individual differences in which are typically measured using the Somatosensory Amplification Scale (SSAS [15], [16]).
There is evidence that MUS are associated with deficits in attentional vigilance and mental set-shifting [17], automatic and voluntary orienting [18], and disengaging attention from stimuli [18], [19], [20], [21], [22]. Some studies have found that MUS patients attend more to illness-relevant stimuli than controls [23], [24], [25], although this has proved difficult to replicate [26], [27], [28], [29], [30]. In terms of more direct measures of body-focused attention, one study found that high SSAS scorers were poorer than low scorers at detecting physiological events (heartbeats), suggesting an attentional deficit in the former [31]; two other studies have found no relationship between SSAS scores and heartbeat detection, however [32], [33]. One recent investigation [34] found that SSAS scores were negatively correlated with a performance measure of tactile (i.e., bodily) attentional bias following exposure to body-related stimuli. After controlling for SSAS scores, however, nonclinical participants with high self-reported somatoform dissociation still showed the expected increase in tactile bias when exposed to threatening body-related pictures. These latter findings suggest that attention towards and away from the body might both be relevant in somatoform dissociation.
Despite much theorizing about attention in MUS, there remains a paucity of empirical data, particularly regarding somatoform dissociation. Moreover, the available literature is largely unable to identify the precise attentional processes in question. Accordingly, the current article describes a study investigating the relationship between somatoform dissociation and performance on a well-validated attention measure, the exogenous cue-target task [35]. There is evidence from a visual cue-target task that patients with pseudoneurological symptoms have difficulties in disengaging visual attention [18]. We used a tactile version of this task [36], [37] to measure how attention is drawn to, and disengaged from, touch stimuli, which is arguably more relevant to body-related attentional processes. Since somatoform dissociation is a trait-like phenomenon distributed across both normal and clinical populations [8], we sampled nonclinical participants with either high or low scores on the Somatoform Dissociation Questionnaire (SDQ-20) [38] and compared their performance on the task.
We investigated the association between traumatic affect and somatoform dissociation by presenting half of the participants with the task in the context of a neutral film and half in the context of a film to mimic the emotional after-effects of trauma. The trauma film paradigm has long been used as an experimental analogue to test the emotional consequences of stressful events, such as later film-related intrusive memories [39]. It has been used in a range of studies to examine forms of dissociation (e.g., Refs. [40], [41]) and, more recently, to probe the impact of somatoform dissociation on intrusive memories [42]. Film-related intrusive thoughts (typically in the form of intrusive images of scenes from the film) are monitored using a self-report diary for 1 week after the experiment, allowing a controlled test of vulnerability factors related to the impact of a stressful event.
A 2×2×2×4 mixed-model design was used. Film condition (neutral; trauma) and SDQ-20 group (low; high) were between-subjects factors; time (pre-film; post-film) and stimulus-onset asynchrony (SOA; 150 ms; 350 ms; 500 ms; 1000 ms) were within-subjects factors.
Participants initially completed an online version of the SDQ-20; those meeting SDQ-20 inclusion criteria were invited to participate in the experimental session. The session began with questionnaire measures of trauma exposure, trait anxiety, somatosensory amplification, current mood, and current physical symptoms. They were then given an overview of the cue-target task, followed by a practice block and the experimental blocks. Participants then watched either a neutral or a “traumatic” film, re-rated their current mood and physical symptoms, and completed the cue-target task for a second time, comprising the experimental blocks only. They were then presented with a diary to record their physical symptoms and film-related intrusive thoughts in the week following the experimental session. Upon return of the diary, participants were debriefed and received course credit or £5 towards expenses.
One hundred twenty-four students (98 female; mean age=20.7 years, S.D.=3.2 years) were recruited. There were no significant differences in age [F(1, 124)=1.527; P>.1] or gender (all P's>0.1) between the groups and conditions. Ethical approval for the study was obtained from the relevant committee.
The “trauma” film was a 14-min compilation of three scenes from commercial feature films,1 depicting rape, murder, and torture. To provide an experimental analogue of psychological trauma [39], the film content was consistent with DSM-IV criterion A1 (i.e., an event involving actual death or threat to life, or threat to physical integrity). The neutral film was a 13-min compilation of three scenes from films depicting neutral human activity, including talking at home, talking in and around a courtroom, and eating at a café. In line with previous studies, each scene was preceded by a synopsis of the scenario, presented for approximately 40 s [41], [43], [44]. Films were shown on a laptop computer with a 14.1-in. screen. Stratified random allocation ensured that equal numbers of high- and low-SDQ participants were allocated to the conditions.
The SDQ-20 [38] was used to estimate the tendency to experience somatoform dissociation. Each of 20 items describes a symptom; respondents rate the degree to which this has applied to them in the past year, using a five-point Likert scale from 1 (not at all) to 5 (extremely). Total scores ranging from 20 to 100 were computed following Nijenhuis [45], with 20 indicating that the participant had not experienced any of the symptoms. Scores of ≥28 and ≤22 were taken as cut-offs for membership of the high- and low-SDQ groups, respectively, in line with population norms for the scale [46].
The SDQ-20 was selected in preference to other symptom measures because it focuses specifically on symptoms in the neurological domain (e.g., sensory loss, paralysis, seizures, etc.). As such, scores on the measure are less likely to be contaminated by symptoms arising from minor physical ailments, or from the somatic concomitants of depression and anxiety (e.g., autonomic arousal), and therefore provide a purer estimate of “true” somatoform dissociation.
The trait scale from the State-Trait Anxiety Inventory (STAI-T) [47] was used to measure trait negative affectivity. The scale consists of 20 statements (e.g., “I lack self confidence”) that are rated on a four-point Likert scale ranging from 1 (not at all) to 4 (very much so). Total scores range from 20 to 80. STAI-T scores were used as a covariate to control for the effects of negative affectivity, which is commonly associated with the tendency to experience physical symptoms [48].
An abridged version of the Trauma Assessment for Adults (TAA) [49] was used to ensure that participants in the neutral and trauma conditions were comparable in terms of exposure to potentially traumatic events. Scale items are based on the Potential Stressful Events Interview used in the DSM-IV field trial [50]. Thirteen items, relating to 12 specific potentially traumatic events (e.g., “Before you were age 18, has anyone ever used pressure or threats to have sexual contact with you?”) and one “other” category, are presented. Respondents indicate whether they have experienced an event or not, using a yes–no response format. A count of positively endorsed events allows for a trauma history score ranging from 0 to 13.
The Somatosensory Amplification Scale (SSAS) [15] is a measure of the tendency to notice ambiguous, but largely benign, sensory events and to experience them as unpleasant. The scale consists of 10 statements (e.g., “I am often aware of various things happening within my body”), rated on a five-point Likert scale ranging from 1 (not at all true) to 5 (extremely true). Total scores range from 10 to 50. Following Ref. [34], the SSAS was used to control for individual differences in the tendency to experience tactile stimulation as aversive, rather than as a trait measure of body focus per se.
A modified Symptom Checklist [51] was used to measure participants' experience of a range of common physical symptoms before and after the films. Symptoms were anchored on seven-point unipolar scales, ranging from the absence of a symptom through increasing degrees of its presence (e.g., “no headache” to “headache”). In the version adapted here, “pain” and “fatigue” were added as these are two of the most frequently observed MUS [52], as was an “other” category. A total sum score from 0 to 90 was obtained by summing scores for each symptom. We used ANCOVA to control for post-film symptom totals (when film conditions were considered in isolation during the analysis), allowing us to establish whether any attentional effects were due to between-group differences in physical symptoms.
We measured current mood (“happy,” “anxious,” “depressed,” “angry”) before and after the films using an 11-point scale ranging from 0 (not at all) to 10 (extremely). “Distress” was measured post-film only. We measured current mood to check that our trauma manipulation had the anticipated effect on emotional state.
Following Ref. [41], participants completed a Symptoms and Intrusions Diary comprising daily Symptom Checklists and sections for identifying intrusive thoughts (and their associated distress) related to the films during the week following the experiment. Total number of intrusive thoughts and total intrusion-related distress were multiplied to produce a composite intrusion burden variable. We anticipated that participants in the trauma condition would report significantly higher intrusion burden than those in the neutral condition, analogous to the effect of traumatic events on intrusive symptom experience in patients with posttraumatic stress disorder. We also predicted that this effect would be significantly greater for the high SDQ-20 participants, based on the assumption that the trauma film would have a greater emotional impact on these participants. In order to assess whether attention to the body predicted intrusion burden following the trauma film, we looked at the correlation between post-film cueing effects and intrusion burden for participants in the trauma condition.
A tactile cue-target task [36], [37] was used to measure attention to the body. Participants made a speeded judgment about the frequency of a tactile target (high- or low-frequency vibration) presented to one of the hands, which was preceded by a cue vibration presented to either the same or the opposite hand. The time between cue and target (SOA) was varied to measure the time course of attention to the cue. At short SOAs, responses are quicker when the cue and target are presented at the same location (cued trials) than at the opposite location (uncued trials), reflecting automatic attention to the cued location [37]. Conversely, at longer SOAs, participants are slower to respond on cued than on uncued trials [36], [37]; this “inhibition of return” effect is thought to arise because attention is slower to return to a location it has recently disengaged from [53].
To minimize the impact of speed/accuracy trade-offs, performance in each condition was calculated using inverse efficiency (IE) scores [IE=reaction time/(1−proportion of trials in which the wrong response was made)]. IE scores for targets presented to the same hand as the cue (cued) were subtracted from IE scores for targets presented to the opposite hand (uncued) to produce a cueing effect estimating the influence of the cue on exogenous (i.e., “stimulus driven”) attention. Higher scores on this measure indicate greater benefits of being cued to the location of the target and/or lower costs of being cued to an irrelevant location.
Participants were centered 35 cm in front of a 17-in. monitor with their arms approximately 15 cm on either side of this midline. A 1-cm cross at the screen center (approximately 1.6° of visual angle) was used as a fixation point. Tactile stimuli were delivered via vibrotactile bone conductors (B/C 2-PIN, 100 Ω, Oticon Limited, Hamilton, UK) within two 9×7×4-cm foam cubes. Stimuli were produced using digital sound files run through an amplifier (Dancer Design TactAmp 4.2). The “buzz” cue was a 10-ms stream of white noise. The targets were 300-ms trains of square waves. The “high” and “low” target vibrations had frequencies of 200 and 40 Hz, respectively. Participants matched the subjective intensities of the high vibrations across left and right conductors before the task.
Participants placed their right foot on two pedals, so their toes rested on the foremost pedal and their heel, the rearmost. Half the participants were instructed to raise their toes in response to a high vibration and their heel in response to a low vibration. The other half were given the opposite instruction. Participants were played white noise through a sound-attenuating headset to prevent them from hearing the vibrations.
Participants were informed that they would receive a “quick buzz” followed by a “high or low vibration” and instructed to identify the vibration as high or low as quickly and accurately as possible. The fixation point then appeared and, after a random interval between 700 and 1000 ms, the cue was presented to one of the participant's index fingers. The target was presented following a variable delay from the start of the cue (150, 350, 550, or 1000 ms SOA), to enable both attentional facilitation and disengagement to be measured [37].The target was presented to the cued index finger on 50% of trials (i.e., the cues did not predict target location). If participants correctly identified the frequency, the trial terminated and the screen went blank. If they did not respond within 1400 ms, “no response” was displayed. If they responded incorrectly, “wrong” was displayed. An interval of 1000 ms separated the response to one trial and the start of the next.
A practice block of 30 trials was presented followed by 160 experimental trials split into three blocks; participants could rest between blocks. Twenty trials were presented in each of the eight conditions [Cue location (cued, uncued)×SOA (150, 350, 550, 1000)]. In each condition, target type (high vs. low frequency) and location (left vs. right hand) were balanced but not analyzed.
Prior to analysis, errors were removed from the tactile-attention data. Trials where participants made the wrong response or responded prematurely (reaction times <150 ms) or very late (>1200 ms) were deemed invalid and excluded (9.6% of all data). In order to remove comparatively long and short reaction times, a within-subjects outlier procedure was used to remove reaction times n standard deviations from the mean, where n was adjusted as a factor of sample size [54]. This avoids potential distortions when sample size (i.e., number of errors) differs in different conditions and participants. IE scores were then calculated for each participant in each condition. Finally, cueing effects (IE uncued trials−IE cued trials) were computed for each participant in each condition.
Nonnormal variables (post-film anxiety, depression, anger, and distress) were transformed using square root and log transformations as appropriate, following the recommendations of Tabachnick and Fiddell [55]. Intrusion burden remained nonnormal following transformation so was analyzed untransformed. Corrections were applied in cases where the assumption of sphericity for repeated measures ANOVA was violated [56]. Analyses were conducted first without and then with STAI-T, TAA, and SSAS scores as covariates. Where repeated measures ANCOVAs were performed, the relevant correction [57] was applied to all covariates.
Section snippets
Group comparability
Table 1 presents descriptive statistics for the questionnaires. Between-subjects MANOVA revealed a significant main effect of the SDQ-20 group [F(4, 117)=70.63, P<.001, ηp2=0.707]; the main effect of film [F(4, 117)=1.39, P=.241, ηp2=.045] and the Group×Film interaction [F(4, 117)=1.43, P<.230, ηp2=0.046] were nonsignificant. F tests revealed that participants in the high SDQ-20 group scored significantly higher on all of the measures: SDQ-20, F(1, 120)=270.43, P<.001, ηp2=.693; STAI-T, F(1,
Discussion
Low-SDQ participants showed a typical pattern on the cue-target task, characterized by a significant positive cueing effect at the 150-ms SOA, followed by a linear decrease in cueing effect across SOAs [53]. A similar pattern was observed in both film conditions for this group, with cueing effects being greater across all SOAs following the trauma film when controlling for covariates. This adds to the literature indicating that attentional cueing effects, even to neutral stimuli, are influenced
References (67)
- et al.
Medically unexplained symptoms. An epidemiological study in seven specialities
J Psychosom Res
(2001) - et al.
Number of bodily symptoms predicts outcome more accurately than health anxiety in patients attending neurology, cardiology, and gastroenterology clinics
J Psychosom Res
(2006) - et al.
A hierarchical classes analysis (HICLAS) of primary care patients with medically unexplained somatic symptoms
Psychiatry Res
(1998) - et al.
The Somatosensory Amplification Scale and its relationship to hypochondriasis
J Psychiatr Res
(1990) - et al.
Is somatization a habituation disorder? Physiological reactivity in somatization syndrome
Psychiatry Res
(2001) - et al.
Attentional bias towards health-threat information in chronic fatigue syndrome
J Psychosom Res
(2008) - et al.
Personalized pain words and Stroop interference in chronic pain patients
Eur J Pain
(2003) - et al.
Do patients with chronic pain attend to pain-related information? Preliminary evidence for the mediating role of fear
Pain
(1997) - et al.
Hypervigilance and attentional fixedness in chronic musculoskeletal pain: Consistency of findings across modified Stroop and dot-probe tasks
J Pain
(2005) - et al.
Dot-probe evaluation of selective attentional processing of pain cues in patients with chronic headaches
Pain
(2005)
Feeling your body or feeling badly: Evidence for the limited validity of the Somatosensory Amplification Scale as an index of somatic sensitivity
J Psychosom Res
Accurate awareness of heartbeat in hypochondriacal and non-hypochondriacal patients
J Psychosom Res
Somatoform dissociation and somatosensory amplification are differentially associated with attention to the tactile modality following exposure to body-related stimuli
J Psychosom Res
Investigating the time-course of tactile reflexive attention using a non-spatial discrimination task
Acta Psychol
Inducing and modulating intrusive emotional memories: A review of the trauma film paradigm
Acta Psychol
The influence of a visuospatial grounding task on intrusive images of a traumatic film
Behav Res Ther
Effects of threat cues on attentional shifting, disengagement and response slowing in anxious individuals
Behav Res Ther
A cognitive-motivational analysis of anxiety
Behav Res Ther
Trauma and medically unexplained symptoms: Towards an integration of cognitive and neuro-biological accounts
Clin Psychol Rev
Somatization increases medical utilization and costs independent of psychiatric and medical comorbidity
Arch Gen Psychiatry
The value of pseudoneurological symptoms for assessing psychopathology in primary care
Psychosom Med
Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM–IV)
The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines
Somatoform Dissociation: Phenomena, Measurement, & Theoretical Issues
L'automatisme psychologique [Psychological automatisms]
Hysteria: A neurobiological theory
Arch Gen Psychiatry
Dissociative and conversion disorders
The cognitive psychology of dissociative states
Cognit Neuropsychiatry
Psychological mechanisms of medically unexplained symptoms: An integrative conceptual model
Psychol Bull
Somatoform disorders
The amplification of somatic symptoms
Psychosom Med
Cognitive performance in conversion hysteria
Arch Gen Psychiatry
Endogenous and exogenous attention in patients with conversion paresis
Cogn Neuropsychol
Cited by (26)
Electrophysiological evidence for changes in attentional orienting and selection in functional somatic symptoms
2019, Clinical NeurophysiologyCitation Excerpt :These statistical results confirm the presence of attentional modulations at later latencies in both groups; however, these were stronger and more widespread for the high group. Previous research has suggested that changes in attention contribute to high somatic symptom reporting in the absence of a medical cause (Brown et al., 2010; Horvath e al., 1980; Roelofs, 2003). To investigate the neural basis of attentional changes we invited participants who scored high on the SDQ indicating the presence of a number of functional somatic symptoms and, in addition, we invited a control group who scored low on the SDQ indicating absence or very weak presence of somatic symptoms to perform a tactile exogenous attention task while their brain activity was recoded.
Attending away from the body predicts increased physical symptom reports at six months in primary care patients
2018, Journal of Psychosomatic ResearchCitation Excerpt :There has been less research investigating attentional biases in more body-relevant sensory modalities, such as touch. Brown et al. [47] found that high symptom reporters were slower to disengage their attention from tactile cues than low symptom reporters under neutral conditions, but then displayed avoidance of tactile stimuli following a negative mood induction. In contrast, Brown et al. [48] found that non-clinical participants with high symptom reports were disproportionately faster than controls when responding to tactile versus visual targets (suggesting body bias), but only following presentation of threatening stimuli.
Psychological characteristics of dissociation in general population
2017, Clinica y SaludCitation Excerpt :The profile described in somatization was similar to that obtained in dissociation, but differences between low and high scores in somatoform dissociation are less evident in the different facets explored. These data suggest that this type of dissociation is related to more anxious profiles, which is consistent with those studies that related somatoform disorders with anxiety processes, the arousal, the hypervigilance to bodily sensations, and more avoidant profiles (Baslet, 2011; Brown et al., 2010; Novakova, Howlett, Baker, & Reuber, 2015). In terms of sleep, the profile found in somatization is the same as in the psychoform dissociation, highlighting the importance of these processes and symptoms in both types of dissociation.
Cognitive behavioral therapy for somatoform disorders and pain
2017, The Science of Cognitive Behavioral TherapySymptoms and the body: Taking the inferential leap
2017, Neuroscience and Biobehavioral Reviews