human psychopharmacology Hum. Psychopharmacol Clin Exp 2012; 27: 595–604. Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/hup.2269
The loudness dependence of auditory evoked potentials and effects of psychopathology and psychopharmacotherapy in psychiatric inpatients Julia Ostermann†, Idun Uhl†, Elke Köhler, Georg Juckel and Christine Norra* Department of Psychiatry, Psychotherapy and Preventive Medicine, Ruhr University Bochum, LWL University Hospital Bochum, Bochum, Germany
Objective Many studies have provided evidence for the loudness dependence of auditory evoked potentials (LDAEP) as a marker for central serotonergic activity but remained inconclusive for its suitability in clinical use. Methods A cross-sectional sample of 162 psychiatric inpatients (major depression N = 86, bipolar disorder N = 12, schizophrenia N = 50, and schizoaffective disorder N = 14) and 40 healthy subjects was retrospectively examined for LDAEP and effects of psychopathology and psychopharmacology. Results The LDAEP was weaker in patients with affective disorders than in healthy subjects but did not differentiate between the total patient sample and healthy controls. LDAEP correlated significantly with dimensions of the Brief Symptom Inventory in the total patient sample (depression, paranoid ideation, psychoticism, Global Symptom Index, and Positive Symptom Distress Index), in patients with affective disorders (depression) and with schizophrenia spectrum disorders (depression, psychoticism, Global Symptom Index, and Positive Symptom Distress Index). Similar correlations were found in depressed patients with a single noradrenergic and specific serotonergic antidepressant or serotonin– norepinephrine reuptake inhibitor. There was a negative correlation between dosage of typical antipsychotics and LDAEP. Hypnotics generally led to a lower LDAEP. Conclusion The LDAEP in patients is related to severity of psychopathologic syndromes irrespective of diagnosis. Chronic psychopharmacologic treatment may also differentially modulate the LDAEP, but longitudinal studies are needed. Copyright © 2012 John Wiley & Sons, Ltd. key words—event-related potentials (ERP); loudness dependence of auditory evoked potentials (LDAEP); serotonin; affective disorder; depression; schizophrenia
INTRODUCTION The loudness dependence of auditory evoked potentials (LDAEP) has been discussed as a marker for central serotonergic neurotransmission for several years. It can be used to obtain information on central serotonergic activity (Hegerl and Juckel, 1993; Juckel et al., 1999, 2003; Gallinat et al., 2000; Tuchtenhagen et al., 2000; Lee et al., 2005; Linka et al., 2005; Nathan et al., 2006). LDAEP, also known as stimulus intensity dependence, indicates the amplitude change of the auditory evoked N1/P2 component in response to different stimulus intensities. A high LDAEP indicates
*Correspondence to: C. Norra, Department of Psychiatry, Psychotherapy and Preventive Medicine, Ruhr University Bochum, LWL University Hospital Bochum, Alexandrinenstrasse 1-3, 44791 Bochum, Germany. Phone: +49234-5077 1131; Fax: +49-234-5077 1329. E-mail:
[email protected] † Both authors contributed equally to this article.
Copyright © 2012 John Wiley & Sons, Ltd.
low serotonergic activity and vice versa (Hegerl and Juckel, 1993). There is a strong evidence for the sensitivity and specificity of the LDAEP as such an indicator deriving from animal studies (Juckel et al., 1999; Manjarrez et al., 2005; Wutzler et al., 2008). In humans, a reduced LDAEP was found in schizophrenia patients (Juckel et al., 2003, 2008). In depressed patients, LDAEP was identified as a predictor of acute treatment response when a strong LDAEP exhibited a better response to selective serotonin reuptake inhibitor (SSRI) treatment (Paige et al., 1994, Gallinat et al., 2000; Linka et al., 2004; Lee et al., 2005; Juckel et al., 2007; Park et al., 2011) and other monoaminergic antidepressants (Paige et al., 1995, Linka et al., 2009b). Further patient studies investigated confounding factors of LDAEP such as substance abuse (Preuss et al., 2000; Tuchtenhagen et al., 2000), personality disorder (Norra et al., 2003), medication (Lee et al., 2005; Park et al., 2010), or Received 11 March 2012 Accepted 24 September 2012
596
j. ostermann
different stages of illness (Gudlowski et al., 2009). However, the results of these studies are still inconclusive. In this study, we aimed at examining a larger patient sample of different diagnostic entities derived from the naturalistic inpatient population of a psychiatric hospital. LDAEP should be investigated considering severity of psychopathology and differential effects of psychopharmacologic medication in the continuing establishment of the LDAEP as a general marker for central serotonergic function of use in the clinical routine. Moreover, further LDAEP data of larger patient samples are required to advance the ongoing discussion about the necessity of standards in LDAEP measurement as a neurophysiological tool.
METHODS Subjects Two hundred fourteen psychiatric inpatients at the Department of Psychiatry, Ruhr University Bochum, Germany, who were referred for a routine electroencephalography (EEG) recording were gathered for a retrospective analysis of their event-related potential (ERP) data over a period of 2.5 years (December 2007 until May 2010). Patients were included in the crosssectional study if they met the criteria of one of the following disorders according to the International Classification of Diseases (ICD-10; WHO, 1991): schizophrenia (F20.x), schizoaffective disorder (F25.x), bipolar affective disorder (F31.x), depressive episode/ recurrent depressive disorder (F32./33.x), that is, depression. Exclusion criteria were age below 18 years, a personal history of neurological or severe somatic illness, hearing impairment, or substance intoxication
Table 1.
ET AL.
prior to admission. Patients were excluded if their routine EEG data showed pathological patterns, for example, epileptic activity. Because of higher artifact rates, 35 patients had to be excluded from the analysis and so were 17 because of further exclusion criteria (refer to the aforementioned criteria). Data from 162 patients (84 men and 78 women) remained in the final analysis (Table 1). Main diagnostic groups were depression (N = 86) and schizophrenia (N = 50). The smaller group of 12 bipolar patients comprised two patients with a manic episode, six patients with a depressive episode, and four patients with a mixed episode. Fourteen schizoaffective patients were included: 10 with a depressive episode and four with a mixed episode. In addition, ERP data from 40 healthy control subjects were included in the dataset. Exclusion criteria for healthy control subjects were a previous or family history of psychiatric and neurological diseases, other chronic or severe somatic diseases, or any continuing medication except for contraceptive pills. Regarding their psychiatric histories, 99 patients reported previous treatments of psychiatric disorders, mostly affective disorders (N = 44) and schizophrenia spectrum disorders (schizophrenia/schizoaffective disorders, N = 21). Other reasons for prior psychiatric treatment including secondary diagnoses were alcohol abuse, anxiety disorder, aggression, posttraumatic stress disorder, social phobia, adjustment disorder, borderline personality disorder, and attention deficit hyperactivity disorder. Sixty-five patients reported a positive family history of neurological or psychiatric illness including affective disorders (N = 27), schizophrenia (N = 14), alcohol and drug abuse (N = 11), and to a lesser extent dementia, Parkinson’s disease, borderline personality disorder, and eating disorder.
Demographic dataset of patients and healthy control subjects
Age: years (SD) Gender: m/f Handedness: right/left Education: ≥12/9–11/≤8 years Smoking: non-smoker (>6 months smoke free)/smoker Marital status: married/unmarried/separated or divorced/widowed
Healthy controls
Depression
Schizophrenia
Bipolar disorder
Schizoaffective disorder
Group differences
N = 40
N = 86
N = 50
N = 12
N = 14
p
45.0 (+14.2) 8/4 6/1 (N = 7) 4/3/0 (N = 7) 3/4 (N = 7)
44.2 (+11.4) 7/7 4/2 (N = 6) 3/4/1 (N = 8) 2/6 (N = 8)
0.001a 0.012b 0.532b 10. The Brief Symptom Inventory (BSI; Derogatis, 1993) consists of 53 self-rated questions covering different dimensions of subjective somatic and psychological impairments, also with respect to non-patient populations. The global scales Global Severity Index (GSI) and Positive Symptom Distress Index (PSDI) of the BSI as well as the subscores G4 (depression), G8 (paranoid ideation), and G9 (psychoticism) were used in this study with reference to the main diagnostic entities. As there are no normative data on LDAEP values and we aimed for a comparison with BSI data, the subscores were not T-transformed. Loudness dependence of auditory evoked potentials Investigation of auditory evoked potentials took place in an electrically shielded and sound-attenuated room adjacent to the recording apparatus (BrainVision BrainAmpW, MR, Brain Products GmbH, Munich, Germany). The subjects were seated in a slightly reclined chair with a head rest. They were asked to keep their eyes open during the entire testing. Auditory evoked potentials were recorded with 32 non-polarizable Ag–AgCl electrodes referred to as FCz, placed according to the international 10/20 system. Auditory stimuli were presented binaurally via headphones. Pure sinus tones (1000 Hz, 40 ms, 10 ms r/f, ISI 1800–2200 ms) of five different intensities (60, 70, 80, 90, 100 dB sound pressure level) were presented in a pseudo-randomized way using Presentation 11.3W (Neurobehavioral Systems Inc., Albany, CA, USA). Eye movements were controlled for with an electrode located 1 cm below the left outer canthus. Impedances were kept below 5 kΩ. Data were collected with a sampling rate of 500 Hz and an analogous band-pass filter (0.16–70 Hz). Seventy stimuli per intensity were presented. Subjects were included in further analysis if at least 40 artifact-free sweeps, that is, maximum amplitude range of 100 mV, during a 350-ms prestimulus and an 800-ms poststimulus period were available. Using BrainVision AnalyzerW (Brain Products GmbH, Munich, Germany), artifact-free stimuli were averaged. The N1 amplitude was regarded as the nadir between 50 and 150 ms after the stimulus and P2 as the peak between 100 and 250 ms poststimulus. These amplitudes were collected semi-automatically. The N1/P2 amplitude was then calculated as the difference of peak amplitudes between N1 and P2. The LDAEP Copyright © 2012 John Wiley & Sons, Ltd.
597
of the scalp data (Cz) was calculated as a median exponential slope of the amplitudes (taking all possible 10 connections between the five different amplitude values) of the single loudness levels (Hegerl and Juckel, 1993). Various studies have shown that N1 and P2 reach the maximum amplitude at Cz, and scalp data of N1 and P2 usually are analyzed at this point with the most reliable results. Several studies also employed dipole source analysis to display the N1/P2 component at its origin in the primary auditory cortex. However, the superiority of dipole source analysis is still controversial (Hensch et al., 2008), and the present study made the priority of use of LDAEP in clinical practice. Test protocol Healthy control subjects and patients were all examined in the same laboratory between 9 and 12 AM. Prior to the EEG recordings, they obtained information about the ERP study and subsequently gave their written informed consent to participate. Additional ERP measurements were in accordance with the Helsinki Declaration as revised in 1989. After routine EEG recording that lasted for 20 min, the subsequent auditory ERP measurement was performed for another 12 min. After this session, participants completed the psychometric questionnaires and a standardized demographic questionnaire. Psychopharmacologic medication was documented along with the dosages on the EEG recording day and the day before testing (Table 2). Statistics Statistical calculations were first carried out involving all four psychiatric diagnostic groups. To obtain larger subsamples, we performed clustering of diagnoses for affective disorders (ICD-10: F3x.x, i.e., unipolar depressive disorder and bipolar disorder, N = 98), as well as for schizophrenia spectrum disorders (ICD-10: F2x.x, i.e., schizophrenia and schizoaffective disorder, N = 64). Tests for normal distribution of all variables were conducted prior to the statistical analysis. If test scores failed to show a normal distribution, non-parametric tests were used. The LDAEP, that is, median slope of the Cz electrode, was used as a dependent variable and the diagnostic groups as independent variables. Differences between two groups were analyzed by t-tests or Mann– Whitney U tests. Differences between three or more groups were analyzed by analyses of variance with additional Bonferroni post hoc tests. With regard to treatment effects, dosage of antidepressant medication was taken as an independent variable and LDAEP as a dependent variable. To warrant comparability and to reduce possible interacting effects of medication, we included only patients who were administered a Hum. Psychopharmacol Clin Exp 2012; 27: 595–604. DOI: 10.1002/hup
598 Table 2.
j. ostermann
ET AL.
Psychopharmacologic medication of all four patient groups and the total patient sample Depression Schizophrenia Bipolar disorder Schizoaffective disorder
Medication Antidepressants Antipsychotics Benzodiazepines, hypnotics, and antihistamines Mood stabilizers SSRI (citalopram/escitalopram/sertraline/ fluoxetine/paroxetine)
NaSSA Mirtazapine Other antidepressants (tricyclics, tetracyclics, MAO inhibitors, NDRI) Antipsychotics Atypical Typical
Group differences
N = 86
N = 50
N = 12
N = 14
N = 162
p
74 29 35 11
14 46 13 1
10 8 2 6
4 13 4 4
102 96 54 22
