Article Text
Abstract
Background Psychogenic movement disorders are disorders of movements that cannot be explained by a known neurological disorder and are assumed to be associated with psychiatric symptoms such as depression and anxiety.
Objective To examine the neuropsychological profile of patients with psychogenic movement disorders.
Methods We examined cognitive functioning using neuropsychological tests in 26 patients with clinically established psychogenic jerky movement disorders (PMD). We included 16 patients with Gilles de la Tourette syndrome (GTS) who served as a patient control group, in addition to 22 healthy control subjects. Non-credible test performance was detected using a Symptom Validity Test (SVT). Psychopathology was also assessed.
Results Apart from a worse performance on a verbal memory task, no evidence of neuropsychological impairments was found in our PMD sample. Interestingly however, patients with PMD reported more cognitive complaints in daily life and performed worse on the SVT than the two other groups. Patients with GTS did not report, or show, cognitive impairments. In patients with PMD, we found associations between verbal learning, SVT performance and severity of depression and anxiety complaints.
Conclusions We conclude that some patients with PMD show non-credible cognitive symptoms. In contrast, no evident cognitive impairments were present in patients with PMD or GTS. Our study underlines the importance of assessment of non-credible response in patients with PMD. Additionally, non-credible response might aid in the differentiation of PMD from other movement disorders.
- Cognitive Neuropsychology
- Gilles De La Tourette
- Movement Disorders
- Psychiatry
Statistics from Altmetric.com
Introduction
Patients with psychogenic movement disorders (PMD) have symptoms that can resemble any known neurological movement disorder, including tremor, dystonia, myoclonus, paralysis, tics and gait disorders.1 ,2 Although recent studies reported evidence of functional brain changes—for example, in frontal and subcortical motor circuits, limbic regions, thalamus, basal ganglia and parietal cortex in PMD (for a review see Nowak and Fink3), the movement disorders cannot clearly be attributed to a structural lesion of the nervous system. Instead, these disorders are assumed to be associated with an abnormal emotional state.3–5 Psychological trauma, including early-life trauma, has been related to PMD,6 although the relationship with traumatic life events is not strictly linear. PMD is commonly classified as a specific subtype of conversion disorder (DSM IV TR; American Psychiatric Association, 2000). In this study we focus on psychogenic hyperkinetic movement disorder, in particular patients with brief jerky movements.
The neuropsychological aspects of PMD have been examined in only a limited number of studies. Studying cognitive functions in these samples may provide insights into whether subjects with PMD differ from those with neurological movements disorder in certain neuropsychological characteristics, thus creating possibilities for defining neuropsychiatric ‘profiles’ of these disorders. In other hyperkinetic conversion disorders, there is some evidence of cognitive impairments. For instance, patients with psychogenic non-epileptic seizures (PNES) perform worse in cognitive domains like executive functioning (eg, mental flexibility and problem solving) and spatial memory.7 ,8 However, non-credible responding may account for the lower test performance of patients with a conversion disorder.8 ,9
It is well documented that neuropsychological testing can be invalidated by non-credible responses, that is, where a subject performs beneath the level that he or she would normally reach given a fair amount of dedication on neuropsychological tests.10–12 Credibility of performance on neuropsychological evaluation is usually assessed with Symptom Validity Tests (SVTs).13 Recent studies have yielded evidence of non-credible performance in patients with PMD as indicated by two SVTs—that is, the Structured Inventory of Malingered Symptomatology (SIMS)14 and the Amsterdam Short Term Memory Test (ASTM).15 ,16 Furthermore, low scores on cognitive tasks in patients with PNES can probably be ascribed to non-credible performance,9 with some studies reporting non-credible performance rates of up to 66%.12
Neuropsychological test performance may also be affected by psychiatric comorbidity. For instance, in patients with major depression, impaired performance has been found on tests assessing attention and executive functioning.17 Owing to the high rates of depression and anxiety in PMD, with prevalence rates ranging from 11% to 62%,1 ,18 it is important to control for psychopathology when assessing cognitive abilities in patients with PMD.
Another important issue to investigate is whether the cognitive deficits are an inherent part of the phenotype or secondary to suffering from a (chronic) movement disorder. As patients with Gilles de la Tourette syndrome (GTS) share the combination of motor and psychiatric symptoms, they are a suitable control group in the investigation of neuropsychological functioning in PMD. The core GTS symptoms include multiple motor tics. Psychiatric comorbidity includes, among others, anxiety and depressive disorders.19 Cognitive deficits are not common in patients with pure GTS, although impairments have been found in patients with GTS with comorbid attention deficit hyperactivity disorder (ADHD) and obsessive compulsive disorder (OCD), especially in executive functions like attention and inhibition (for an overview see Eddy et al20).
This study aimed at investigating neuropsychological functioning in patients with PMD and GTS, as well as in a group of healthy subjects, while controlling for non-credible responding. Based on previous findings in patients with other PMDs, we hypothesised that patients with PMD would perform worse on neuropsychological tests than healthy volunteers and patients with GTS. We also expected that patients with PMD would score lower on SVTs than healthy volunteers and patients with GTS. Consequently, we hypothesised that differences between the groups on neuropsychological tests would not persist after controlling for non-credible performance. Finally, we explored the relationship between neuropsychological test performance and psychiatric symptoms.
Patients and methods
Patients
Patients with either PMD or Tourette's syndrome with jerk-like movements who had participated in a previous cross-sectional study21 ,22 were included in this study. These patients were recruited from the outpatient clinic of the Amsterdam Medical Center. In addition, we searched the movement disorders database of our clinic for patients seen between 2000 and 2007 with PMD. Furthermore, our study was announced to the Tourette Patients’ Association.
Twenty-six adults (18 men) were diagnosed with ‘clinically established’ psychogenic movement disorder according to Fahn and Williams criteria23 as determined by an experienced movement disorders specialist (MAJT). All patients with PMD had movements which might resemble either a motor tic or myoclonus, but was considered more likely to be PMD by the physician. ‘Psychogenic myoclonus’ might be the best description of this movement disorder, as described by Monday and Jankovic24; this term has been denounced by some clinical experts in the field as it is an oxymoron. Sixteen adults (13 men) with GTS, in accordance with the DSM-IV criteria, with primarily motor tics were included. For further details see van der Salm et al.21 ,22
Healthy controls
Twenty-two healthy volunteers (13 men) were included. They were matched with the PMD patients for age and education level, as determined with the International Standard Classification of Education of the United Nations Educational Scientific and Cultural Organization (Unesco; 1997, 2006; range 0–6).
For all groups, exclusion criteria were age <18 years, language barriers (eg, non-native Dutch speakers), illiteracy, neurological disorders (eg, history of cerebrovascular accident) and visual disorders and hearing loss that might hinder neuropsychological testing. Furthermore, patients using antipsychotic drugs were excluded because of the negative side effects that medication can have on certain cognitive abilities.25 Participants using other psychoactive drugs (ie, benzodiazepines) were asked to temporarily discontinue using them if possible, 24 h before testing. Finally, control subjects with psychiatric disorders as assessed with the Mini International Neuropsychiatric Interview26 were excluded.
The study design was approved by the medical ethical committee of the Amsterdam Medical Center. Written informed consent of all participants was obtained after the nature of the procedures had been fully explained.
Material
Neuropsychological assessment
Neuropsychological tests covered four cognitive domains: attention, executive functions, memory and reaction speed. Attention was examined using the Trailmaking Test (TMT) part A27 and the word reading and colour naming conditions of the Stroop Task.28 Executive functioning was assessed using the TMT part B, the colour–word interference condition of the Stroop Task and category and phonemic fluency29 tasks. To evaluate memory functions, the Dutch version of the Rey Auditory Verbal Learning Test (AVLT),30 the prose recall task of the Rivermead Behavioural Memory Test31 and the subtest visual reproduction of the Wechsler Memory Scale32 were used. The Vienna Test System33 (subtests S1, simple reaction, and S3, two-choice reaction) was used to determine reaction speed. Finally, participants were asked whether they had any cognitive problems in daily life. The answer was scored on a scale from 0 (no complaints) to 2 (moderate to severe).
Method to detect non-credible performance
The ASTM15 was used to identify patients who demonstrated non-credible performance. A set of memory stimuli is presented, followed by a brief distraction (ie, a simple addition or subtraction task) and a forced choice recognition task. A score below the cut-off point indicates that the patient probably performs below his actual ability. The ASTM has been thoroughly validated.10 ,34 The recommended cut-off score of 8515 was used to separate non-credible from credible response.
Psychopathology
To assess general anxiety, the Beck Anxiety Inventory (BAI35) was used. Severity of depressive symptoms was determined using the Beck Depression Inventory (BDI36).
Medication use
To further rule out possible confounding effects of medication, we examined differences in neuropsychological test performance between groups of subjects prescribed different types of medication (table 1). For the analyses, subjects were assigned to one of the following categories: (1) psychoactive drugs (eg, benzodiazepines, antidepressants); (2) other (eg, immunosuppressant agents, non-steroidal anti-inflammatory drugs); (3) no medication.
Procedure
Neuropsychological and psychiatric functioning was tested by trained (neuro)psychologists and supervised neuropsychology interns in sessions of about 2 h. All participants were assessed individually. Before testing, a brief (neuro)psychological history was taken.
Statistical analysis
First, normality of the data was checked both visually (ie, histogram) and with the Kolmogorov–Smirnov test. If variables were not normally distributed, a stabilising z-score transformation of cumulative proportions (estimated by Van der Waerden's formula) was applied. Group differences on the four neuropsychological domains were tested using multivariate analysis of variance (MANOVA). If a significant effect was found, post hoc Tukey Honest Significant Difference (HSD) tests were applied to determine differences between groups. With this strategy of statistical analysis we reduced the number of comparisons, thereby decreasing the probability of type 1 errors. Associations between the performance on cognitive measures, psychopathology and SVT were evaluated using non-parametric correlation tests (Spearman's ρ). Group differences in age were assessed using one-way analysis of variance, and group differences in gender were investigated using a χ2 test. Differences between the groups in medication use and education level were tested using the Kruskal–Wallis test. This test was also used to assess differences in neuropsychological test performance between the three medication categories. Finally, differences between patients with PMD and those with GTS in disease duration were tested with a Mann–Whitney U test. For all tests, α≤0.05 was considered statistically significant. We chose not to correct for multiple comparisons (except Tukey HSD), as the occurrence of a type II error (failing to detect a difference) is increased in a small research sample like ours, and corrections to prevent type I error (detecting a difference where there is not) such as Bonferroni correction, will probably inflate type II errors.37 ,38 Values missing owing to administrative reasons (eg, defect devices; 0.6% of neuropsychological data points) were replaced by the corresponding group mean. Furthermore, outliers were replaced by the second lowest or highest value in the particular group (1.5% of neuropsychological data points).
Results
Twenty-six patients with PMD, 16 patients with GTS and 22 healthy volunteers were willing to participate in our study. Demographic and clinical characteristics of these subjects are presented in table 1. There were no differences between the groups in education level and gender. However, the groups differed in age (F(2,61)=12.72; p<0.001). Post hoc tests showed that patients with GTS were younger than patients with PMD (p<0.001) and healthy controls (p=0.001). Because of these age differences, we only used demographically corrected T scores (ie, age, or age and education corrected). Also, as might be expected in light of the differences between the patient groups in age at onset, disease duration was longer in GTS than in patients with PMD (p=0.002). As preliminary analyses yielded no significant associations between disease length and neuropsychological test performance, we did not control for disease length in the analyses. Furthermore, we detected differences between groups in medication prescription (χ2=16.75, p=0.002). However, Kruskal–Wallis tests showed no significant differences between the medication categories on the neuropsychological measures (all p values>0.06) for patients with PMD. Moreover, we found no differences between subjects who did or did not use psychoactive medication (ie, category 1 vs categories 2 and 3 combined). Therefore, we decided not to include medication use as a covariate in the statistical analyses.
Neuropsychological and psychopathological findings
Neuropsychological test results and anxiety and depression scores are presented in table 2. When performing the MANOVA, assumptions of equal covariance matrices were met (Box's test, all p values ≥0.12). Although no overall group differences were found in the memory domain, we found group differences on one of the memory subtests—that is, the AVLT immediate reproduction. Groups seemed to differ significantly on this task, with the PMD group performing about one standard deviation below the expected level. Furthermore, groups also seemed to differ on the colour condition of the Stroop, where post hoc analyses yielded lower T scores in patients with GTS than in controls (p=0.007). Finally, group differences were found in the rate of self-reported cognitive complaints, where patients with PMD reported the highest level of cognitive complaints in daily life.
Relationship between neuropsychological test performance and non-credible performances
Group differences were found on the ASTM (F(2,61)=3.09; p=0.05). Post hoc analyses showed that patients with PMD scored significantly lower on this task than patients with GTS (p=0.02; one-tailed) and healthy controls (p=0.04; one-tailed). Because of these findings, we ran the MANOVAs again excluding subjects who performed below the ASTM cut-off point (N=9). Similar results were obtained. Additionally, although the effect was less pronounced (F=3.96, p=0.03), group differences in the AVLT immediate production persisted after covarying for ASTM scores. Correlation analyses showed that a lower score on the ASTM was related to a lower score on the immediate recall of the AVLT (rho=0.41, p=0.001). The mean ASTM test scores for the three groups are presented in figure 1.
Relationship between neuropsychological test performance and severity of psychopathology
The psychiatric results were used to rule out a confounding effect of psychopathology on neuropsychological test performances. In order to minimise the number of variables entered into the correlation analyses we used only tests on which differences were found (ie, ASTM, AVLT, Stroop). Within the PMD group, lower performances on the ASTM were related to higher scores on the BDI (rho=−0.42, p=0.04). Moreover, a lower T score on the immediate reproduction of the AVLT was related to a higher score on the BDI (rho=−0.32; p=0.01) and the BAI (rho=−0.25; p=0.05) within the entire sample. However, the group effect for the AVLT immediate production persisted when covarying for BAI and BDI scores (F=4.04, p=0.02). The ASTM effect reduced to a statistical trend when covarying for the BDI and BAI scores (F=2.80, p=0.07).
Discussion
Our study aimed at investigating the cognitive profile of patients with PMD while controlling for non-credible performance and psychiatric features—in particular, depression and anxiety. Our results showed that patients with PMD and those with GTS do not present evident impairments on neuropsychological examination. However, patients with PMD did report more cognitive complaints in daily life than healthy controls. Moreover, we found that patients with PMD obtained a lower score on the SVT than patients with GTS and controls. Together with previous studies in patients with other types of somatoform disorders,9 ,16 our findings thus provide further evidence that patients with PMD tend to show non-credible performance on neuropsychological examination.
We can only speculate on the causes of the tendency of patients with PMD to demonstrate non-credible performances. In a recent review39 it was hypothesised that in patients with traumatic brain injury factors such as patients’ expectations of outcome, stress and anxiety may influence performance on neuropsychological assessment and on tasks that evaluate effort. Failing the SVT has also often been ascribed to secondary gain, such as a financial incentive.12 ,40 However, research shows that this is mostly not the case with patients with PNES, but that other factors, such as emotional, physical or sexual abuse, predict performance on the SVT.41 However, as some of the PMD patients in our sample were indeed involved in lawsuits or other financial compensation matters, we cannot exclude the possibility that this accounted for their SVT failure. Future studies in patients with PMD should further explore the effects of personality characteristics and behavioural factors on (over)reporting of neuropsychological complaints and performance on the SVT. The finding that patients tend to show non-credible performance also has important implications for previous studies that did not use SVTs. Observed differences in neuropsychological profiles of patients with PMD in these studies should be interpreted with caution, because no claims can be made on the presence of actual cognitive impairments.
We found an inverse relation between the severity of depressive symptoms and credibility of performance in patients with PMD. Although this might indicate that patients with higher rates of depressive symptoms have a higher tendency to demonstrate non-credible cognitive response, an alternative explanation is that patients with PMD also exaggerated their self-reported depressive symptoms. Indeed, the mean BAI and BDI scores in our PMD sample are well below the cut-off point and therefore not clinically significant, making a confounding influence of symptoms of anxiety or depression on neuropsychological test results unlikely. The hypothesis that patients with PMD tend to over-report both cognitive and psychiatric symptoms is confirmed in a study by van Beilen et al,16 demonstrating evidence of symptom exaggeration in patients with PMD on the SIMS—that is, a questionnaire that assesses biased responding on both psychopathology and neuropsychological symptoms. In addition, evidence of over-reporting symptoms in relation to non-credible performance on cognitive tests has been described in a mixed psychiatric patient sample.42 We therefore hypothesise that non-credible performance itself is not due to psychopathological symptoms, such as depression. Rather, both depressive and cognitive symptoms seem to be over-reported in patients with PMD. This finding is important for diagnostic purposes such as to prevent drawing false-positive conclusions, but also has implications for the choice of treatment.
Although we found no significant overall group differences in the memory domain, patients with PMD seemed to perform worse on the immediate reproduction condition of a verbal learning task. As confirmed by our correlation analyses, this result can probably be ascribed to higher rates of psychopathology and non-credible performances. However, as group differences in verbal learning persisted even after controlling for non-credible performance and psychopathology, we cannot entirely exclude the possibility of actual memory impairments in PMD. These impairments might be due to attentional deficits. For instance, patients with psychogenic tremor seem to show excessive visual attention towards the affected limb when performing a motor task.43 However, we found no indication of attentional dysfunction in our patient sample. During the past few years, functional MRI brain imaging studies have shown that patients with PMD may present functional abnormalities in several brain structures, including the hippocampus,3 ,44 providing us with a possible cerebral correlate of the observed memory impairments. The detected functional changes can either be primary or secondary, so no implications can yet be made on causality as these cerebral correlates might just be an epiphenomenon of PMD. To the best of our knowledge, no structural abnormalities have yet been reported.
The observation that patients with PMD tend to over-report cognitive and perhaps psychiatric symptoms adds to the debate on conversion disorders. Based on SVTs we cannot make any assumptions about whether non-credible responding is an intentional, conscious process. The only conclusion one may draw from abnormal scores on (cognitive) SVTs, assuming that severe cognitive impairment is ruled out, is that the patient probably did not put forth a reasonable amount of effort to doing the task; the reason why remains unknown. Thus, the SVT cannot help us decide whether or not a patient is feigning.
Apart from a worse performance on one task that assesses attention, no evident cognitive impairments were seen in our GTS sample. More specifically, no executive dysfunctions were observed. Previous studies did report cognitive deficits in patients with GTS, impairments being most pronounced in executive functions—for example, cognitive inhibition.45 During the past decade, evidence is emerging that these deficits can often be ascribed to comorbid psychiatric disorders, specifically ADHD and OCD.20 ,46 Although comorbidity is not always an explanation, the absence of cognitive impairments in our patients with GTS is probably explained by the absence of ADHD and OCD in this sample.
We acknowledge several limitations of our study. First, the sample size was small. Moreover, there were demographic differences (ie, age) between the groups. As we controlled for these differences by using demographically corrected T scores, this should not have biased our results for the cognitive tasks. Nevertheless, the higher rate of reported cognitive complaints in the PMD group might be partly due to the older age of patients with PMD. Third, no measure of (estimated) IQ was administered in our study. As we found no differences in education between groups, it is unlikely that the differences on neuropsychological tests between groups can be attributed to possible differences in IQ. Nevertheless, we cannot completely rule out IQ as a possible confounding effect. Future research should therefore include a measure of IQ. Furthermore, although we found no evident effects of medication on neuropsychological test performance, we cannot rule out the possibility that medication use has biased our results. Therefore, future studies should aim at including medication-naïve subjects. In addition, we used only one test to detect non-credible performance, whereas the use of at least two SVTs has been recommended.7 ,16 Finally, we tested a specific subgroup of PMD and our findings cannot be generalised to hypokinetic movement disorders.
We conclude that patients with PMD do not present with evident cognitive impairments. They did show evidence of non-credible responding and possible over-reporting of cognitive complaints. We cannot exclude the possibility that this is related to depressive or other psychiatric symptoms, although the association between SVT performance and depressive symptoms can probably be ascribed to non-credible responding on both cognitive and psychopathology scales. Our study underlines the importance of evaluating credibility of performance in patients with somatoform disorders and psychogenic movement disorders, in particular. Therefore, the use of SVTs in future research is strongly recommended.
References
Footnotes
-
CEJH, MJvT contributed equally.
-
Contributors CEJH and MJvT collected, analysed and interpreted the data and drafted the manuscript. SMAvdS, AFvR and DC were involved in the overall design and conception of the study and critical revision of the article. BS was involved in the supervision of the neuropsychological assessments, the statistical analyses and in writing and revising the manuscript. MAJT was responsible for the overall design and conception of the study and critically revised the manuscript. All authors approved the final version of this paper. MAJT stands as guarantor for this work.
-
Funding This work was supported by The Netherlands Organization for Health Research and Development with grant number NWO VIDI 016.056.333.
-
Competing interest None.
-
Patient consent Obtained.
-
Ethics approval Medical ethical committee of the Academic Medical Center.
-
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Editorial commentary