Early effect of NEURAPAS^®balance on current source density (CSD) of human EEG

Sixteen healthy volunteers (8 males, 48,7 ± 7 years old and 8 females, 47 ± 8 years old) were recruited by advertisement and participated in this study. They underwent a medical examination before entering the trial. Subjects reporting neurological disturbances of the central nervous system (using DSM-III criteria) were excluded from the study. Subjects with a history of drug or ethanol abuse or participation in another study within the last six months were also excluded. It was ensured that they were not on alcohol. On the day of examination no beverages containing caffeine were allowed within the last 12 hours preceding the EEG recording. The study was carried out according to the declaration of Helsinki [10] on human rights and was approved by the local ethics committee of the State Medical Association Hessen (Frankfurt, Germany) and government authorities (BfArM, Bonn, Germany). All subjects were informed about the goals of the study in detail and gave their written informed consent to participate. Each subject was randomly allocated to either the functionally active preparation or placebo within a crossover design.

Subjects sat alone in a quiet separate room in a comfortable easy chair. The light was dimmed. Baseline recording of 6 min under the condition of eyes open was followed by performance of three cognitive tests: a concentration test (d2-test) [11], a mathematical calculation test (according to [12]) and a memory test consisting in the presentation of a row of letters and numbers, which appeared for 4 s on a screen followed by 10 s black screen. After this a multiple choice of four possible answers was presented.

All recordings were repeated 0. 5 h, 1. 5 h, 3 and 4 h after the administration of the tablets (overview in Figure 1). Between the measurement sessions the subjects spent their time in the facility's break room. All experiments took place at the same time of the day (starting at 8 h a. m.).

The EEG was recorded bipolarly from 17 surface electrodes according to the international 10/20 system with Cz as physical reference electrode for calculation of the common average reference (Computer aided topographical electroencephalometry: CATEEM^® - from Mewicon GmbH, 4164 Schwarzenberg, Austria) using an electrocap [13]. The raw signals were amplified, digitalized (2048 Hz/12 bit) and transmitted via fiber optical devices to the computer. The automatic artefact rejection of the CATEEM^®-System, which eradicates EEG-alterations caused by eye blinks, swallows, respiration etc. during the recording was visually controlled and individually adjusted by the investigator. Electrocardiogram (ECG) and electrooculogram (EOG) were recorded in one channel each in order to facilitate detection of those signals superposed on to the EEG. The artefact rejection set-up was observed for about 5 minutes prior to the start of the recording to ensure, that all artefacts were correctly recognized and eliminated from further evaluation. For safety purposes the original raw data were saved on optical disk in order to allow re-evaluation of the artefact rejection mode if necessary. In these cases the experimental session was re-examined off-line with a newly adapted rejection mode. The amount of rejected data was determined automatically and given in percent of total recording time. Nevertheless the entire recording and the computer-based automatic artefact rejection were continuously supervised and adjusted by a trained technician [14]. Logarithm of CSD data was taken for discriminant analysis in order to approach normal distribution of values, which underwent frequency analysis for quantitative evaluation as proposed for the first time nearly 80 years ago [15].

In this study the EEG was computed not in the potential mode measured as voltage, but in a surface charge mode obtained by Laplacian estimates also known as current source density (CSD) analysis [16]. The charge is the 2^nd derivative of the potential and gives the spatial curvature of the potential. All calculations are based on the standard set-up of the 10/20 system of recording. Under the condition of using a homogenous, steadily conducting medium the surface charge mode provides the source density of the electrical flow on the cortex surface. Whereas the EEG in the potential mode tends to produce a more extensive and diffuse picture of changes, the Laplacian estimate acts as a spatial filter emphasizing local sources over distant sources. Others [16] were able to demonstrate, that spectral parameters obtained from the CSD showed higher correlations with computer tomography measures than those calculated from the potential mode of the EEG. We therefore used this methodology in order to describe the focal changes of brain activity. Using a Lagrange interpolation, signals from 82 additional virtual electrodes were calculated to provide high-resolution topographical maps. The signals of all 99 electrode positions (17 real and 82 virtual) underwent the Fast Fourier Transformation (FFT) based on 4-second sweeps of data epochs (Hanning window). Data were analysed from 0. 86 to 35 Hz using the CATEEM^® software. In this software the resulting frequency spectra are divided into six frequency bands: delta (1. 25-4. 50 Hz), theta (4. 75-6. 75 Hz), alpha1 (7. 00-9. 50 Hz), alpha2 (9. 75-12. 50 Hz), beta1 (12. 75-18. 50 Hz) and beta2 (18. 75-35. 00 Hz). This frequency analysis is based on absolute spectral power values. Colour coding of the maps is achieved by transforming the content of the power spectrum into the so-called colour vision mode pictures (i. e. frequencies are converted into spectral colours with a resolution of 0. 25 Hz, which gives 140 colours which are then mixed according to spectral content to give one map similar to the RGB mode of TV pictures). Data acquisition and analysis were carried out simultaneously and provide topographical maps displayed on-line on the computer screen. The maps show the relative, time-averaged, functional changes of electric brain activity of each specific recording condition in comparison to the reference period at the beginning of each recording session during the relaxed state with open eyes.

Band powers are firstly compared between placebo and verum during relaxation, then the effects of different mental loads are determined in the absence of either placebo or verum, and finally the results of verum and placebo are compared under different mental loads.

EEG data from the first recording session before drug intake are given as absolute numbers (μV²). By setting the absolute electric power of this first pre-drug recording to 100%, changes produced by the preparations are given as percentage from this pre-drug condition. Except for the construction of brain maps only those particular electrode positions were used for the numeric comparison of verum to placebo during mental load, which have been found to change during the particular mental challenge. For explorative statistical evaluation the nonparametric sign test was used. For mathematical classification of drug effects the linear discriminant analysis according to Fischer was used. Results from the first three discriminant functions were projected into space (x, y and z coordinates), whereas results from the fourth to sixth discriminant functions were coded into red, green and blue, respectively, followed by an additive colour mixture (so-called RGB-mode). Reference preparations used for comparison have been tested in our lab earlier under identical conditions. These data were processed by linear discriminant analysis. Thereafter we used the same transformation by using this projection without any further discriminant analysis procedure for depiction of Neurapas^® balance.

Psychometric performance was evaluated by the mathematical product of "quantity × quality", where quantity was defined as the number of correct answers and quality as the number of correct answers divided by the number of tasks tackled.

NEURAPAS^® balance is a film-coated tablet containing a combination of special dried extracts: 60 mg "St John's wort Herb" (4. 6-6. 5:1), extractant: ethanol 38% (m/m), 28 mg "Valerian root" (3. 8-5. 6:1), extractant: ethanol 40% (m/m), and 32 mg "passion flower herb" (6. 25-7. 1:1), extractant: ethanol 60% (m/m). NEURAPAS^® balance is registered in different European countries for the treatment of mild depression, low mood, anxiety, and sleep disorders and marketed by Pascoe Pharmazeutische Präparate GmbH, Giessen, Germany. The verum preparation consisted of the maximum daily dosage of 6 tablets of NEURAPAS^® balance. At least one week passed between administrations within the crossover design.

Source density analysis of the primary recording of the reference EEG before drug administration revealed similar absolute power values for the first day of placebo administration (Pl) in comparison to the first day of active drug administration (Neu) for all electrode positions as documented in Table 1. Following drug administration no changes were documented between placebo and verum within the first 1. 5 hours (not shown). Three hours after administration lower spectral power values were seen in general for the verum condition (please compare median values in Table 2) compared to placebo values. Only one highly statistically significant value was recognized at the electrode position P4. At four hours after administration again lower values were obtained in general but only alpha waves-particularly alpha2 frequencies at the electrode positions C3, T3, P4 and O1- showed differences to placebo values (Table 3).

Different mental tasks produce changes of electric brain activity within several anatomically defined brain regions. According to earlier experience a concentration test induces spectral frequency changes which can be differentiated from those in the presence of memory testing or performance of mathematical tasks. If one compares the spectral changes during the performance of the d2 test with the relaxed eyes open situation, brain regions react in a different reproducible manner also in this study. Fronto-temporal electrode positions show increases of theta power whereas in other regions alpha1, alpha2 and central beta1 power provide considerably lower values (Figure 2). These changes in comparison to the eyes open situation are statistically significant with p < 0. 05 for particular electrode positions (labelled by stars in the upper part of Figure 2). Documentation of these changes within a brain map containing all frequencies by additive colour mixture (for details s. methods) reveals more increases of theta power within left frontal cortex which leads to orange colour dominance on this side (middle part of Figure 2). Details of the statistical evaluation are given in the lower part of Figure 2 corresponding to the map. Highest significances for frequency changes during performance of the d2 test can be documented for increased fronto-temporal theta power and decreased alpha2 and beta1 power in other regions. Statistically significant decreases of alpha2 power were observed for the electrode positions Cz, C3, C4, Fz, P3, P4 and O2.

A similar picture arises during the performance of the concentration performance test (CPT), where mathematical calculations have to be done. Large increases of left and middle frontal theta power are accompanied by massive decreases mainly in centro-parietal alpha2 spectral power involving also some areas with decrease in beta1 power. These changes in comparison to the relaxed eyes open status are also statistically highly significant (labelled by stars in Figure 3). Increases of left frontal theta power are also recognized in the map (middle part of Figure 3). Statistically significant decreases of alpha2 power in other regions were observed for the electrode positions Cz, C3, Fz, P3, Pz, P4 and O1.

During the performance of the memory test, where a row of 8 combined letters and numbers have to be memorized within 4 seconds, also left and right frontal delta and theta power increase in a highly significant manner (labelled by stars in the upper part of Figure 4) which also leads to dominant red-yellow colour increases in the brain map (middle part of Figure 4). Details of the statistically significant differences in comparison to the relaxed status are documented in the lower part of Figure 4. Statistically significant decreases of alpha2 power were observed for the electrode positions Cz, Fz, P3, P4 and O1.

Thus, all three mental loads lead to somewhat different changes of frequency patterns in comparison to the relaxed state but have in common a fronto-temporal increase in delta and/or theta power with significant decreases of alpha and beta1 power in other brain regions.

No significant differences between verum and placebo were seen with regard to fronto-temporal delta/theta power increases. In order to better interpret the effects of the drug on electric spectral power, medium power of those electrode positions involved in the statistically significant (at least by p < 0. 05) decrease of alpha2 power during mental work were unified to give one parameter set. They represent a brain region heavily involved in the process of a particular mental task performance.

Recording under the three different mental performance conditions succeeded in uncovering drug-induced changes within the alpha and beta1 spectral frequencies during the 3^rd and 4^th hour after drug administration. Results for the d2 test are given in Figure 5. Under this recording condition statistically significant lower power values of alpha1, alpha2 and beta1 power were observed 3 and/or 4 hours after administration of verum in comparison to placebo. Regarding the data during performance of the CPT again statistically significantly lower values were obtained during the 4^th hour after administration for theta, alpha1 and beta1 spectral power (Figure 6). During the 3^rd and 4^th hour performance of the memory test led to significantly diminished values of alpha1 and alpha2 spectral power in comparison to placebo, too (Figure 7).

Thus, during all mental tasks alpha and beta1 power was statistically significantly lower in the presence of verum in comparison to placebo. Analysis of the psychometric results revealed either no difference (Table 4) or a trend of better performance for the CPT at 3 and 4 hours after administration (Table 5), which, however, was not statistically significant, or no difference during the memory test (Table 6).