- Email: [email protected]
Equivalent dipoles of FFT data visualize drug interaction at benzodiazepine receptors
Electroencephalography and clinical Neurophysiology , 86 (1993) 231-237 © 1993 Elsevier Scientific Publishers Ireland, Ltd. 0013-4649/93/$06.00
231
EEG92530
Equivalent dipoles of FFT data visualize drug interaction at benzodiazepine receptors Thomas Dierks a, Wolfram Engelhardt b and Konrad Maurer a a Dept. of Clinical Neurophysiology, Psychiatric Hospital, Wiirzburg (Germany), and b Dept. of Anaesthesiology, University of Wiirzburg, Wiirzburg (Germany) (Accepted for publication: 2 November 1992)
Summary The aim of the present investigation was to study if the benzodiazepine receptor antagonist flumazenil could reverse the effects on the brain electrical activity induced by the benzodiazepine receptor agonist midazolam. The method of FFT approximation was used for this purpose. It allows the calculation of center of gravity equivalent dipoles of spectral EEG data. The results are reference independent and allow therefore a more unambiguous interpretation compared to conventional FFT data reports. Twelve subjects were investigated before and after 0.1 mg/kg and 0.2 mg/kg midazolam respectively, directly and 4 h after administration of 1 mg flumazenil. Our results imply that the application of flumazenil after midazolam sedation leads to an almost complete restoration of the brain electrical activity. However, especially in the beta frequencies above 20 Hz differences in depth of equivalent dipoles were found directly after flumazenil application as well as 4 h later. This could suggest that neuronal generators in different brain structures were responsible for the electrical activity after flumazenil administration compared to before. Key words: FFT approximation; Equivalent dipoles; Reference-free; EEG; Midazolam; Flumazenil
Effects of benzodiazepines on the human electroencephalogram (EEG) have been investigated for more than 20 years (Fink 1969). Midazolam, synthesized in 1976 (Walser et al. 1978), is an imidazobenzodiazepine with unique qualities, e.g., a rapid onset of action and a high metabolic clearance compared to other benzodiazepines (Haefely 1985). Since it has the anxiolytic, hypnotic, anticonvulsive, muscle relaxing and antegrade amnestic effects which are characteristic of all benzodiazepines, it is used clinically for anesthetic premedication, induction and maintenance and sedation for diagnostic and therapeutic procedures. Electrophysiological characteristics of midazolam are widespread beta activity, initially approximately at 22 Hz followed by a 15 Hz rhythm, with simultaneous attenuation and disappearance of alpha rhythm (Brown et al. 1979). Flumazenil selectively antagonizes benzodiazepine effects like sedation and amnesia due to competitive binding at benzodiazepine receptor sites (Hunkeler et
Correspondence to: Dr. Thomas Dierks, M.D., Dept. of Clinical Neurophysiology, Psychiatric Hospital, University of Wiirzburg, Fiichsleinstr. 15, D-W-8700 Wiirzburg (Germany). Tel.: + 931-203334; Fax: + 931-203425.
al. 1981). According to Sch6pf et al. (1984) flumazenil alone produces EEG effects which are similar to nootropic (Dierks et al. 1992b) and psychostimulating drugs (Herrmann 1982). However beta activity remains unmodified. On the other hand, Artru (1990) in dogs and Breimer et al. (1991b) in human volunteers reported that flumazenil effects on the EEG are negligible. Quantitative EEG (QEEG) and computerized electroencephalographic topography (CET) investigations concerning the interaction between benzodiazepines and their antagonists have been carried out (Laurian et al. 1984; Klotz et al. 1985; Breimer et al. 1991a; Engelhardt et al. 1992). Both QEEG and CET studies in the frequency domain allow only descriptive interpretations of results, since they are dependent on the reference used (Lehmann 1987). A more physiological-functional interpretation of EEG data may be achieved by reference-independent data analysis. Conventional FFT reports concerning flumazenil antagonism of benzodiazepine sedation describe a complete reversal of EEG alterations (e.g., Kochs and Schulte am Esch 1989). Consequently the hypothesis of this investigation was that a benzodiazepine antagonist would completely reverse changes in EEG activity induced by a benzodiazepine. This was investigated using
232 the F F T approximation developed by Lehmann and Michel (1989, 1990) followed by calculation of equivalent center of gravity dipoles for all frequencies.
Methods
Subjects Twelve healthy volunteers (6 female and 6 male, mean age 24.3 + 2.6 years, range: 21-30 years, mean body weight 64.2 + 7.7 kg), were investigated. No subject had a history of medical problems or a history of substance abuse or dependency. Furthermore, at the time of investigation none used regular medication. The Ethics Committee at the University of Wiirzburg approved the study. Informed written consent was obtained from the subjects. All volunteers were kept fasting overnight until 1 h after benzodiazepine antagonist administration, when water or juice was offered. Room temperature was kept constant at 25°C over the study period.
Study design E E G was recorded before any drug was administered (baseline), after i.v. administration of 0.1 m g / k g midazolam (half-midazolam), after administration of a total dose of 0.2 m g / k g midazolam (full-midazolam (20 min after half-midazolam)), immediately after 1 mg flumazenil (0 h flumazenil (20 min after full-midazolam)), and 240 min after flumazenil (4 h flumazenil). During the study electrocardiogram, oxygen saturation and blood pressure were monitored.
T. DIERKS ET AL.
Data processing Fast Fourier analysis was done on each 2 sec epoch, multiplied with a Hamming window. The resulting sine and cosine coefficients for each electrode and each frequency point (0.5 Hz resolution) were placed in a sine-cosine diagram (after Nyquist; Norcia et al. 1985; Gutowitz et al. 1986; Lehmann et al. 1986). When a single-generator model is supposed to explain such data, all entries in the diagram have to be orthogonally projected upon a line which yields the least sum of squared deviations between original positions and their projected positions on the line. The projection of the FFT constellation on this line is called F F T approximation (Lehmann and Michel 1989) and describes a map of potential distribution, which is used for estimation of the equivalent dipole source. For a detailed description and discussion of this method see Lehmann and Michel (1989, 1990). For equivalent dipole source localization we used a moving (instantaneous) dipole model (Kavanagh et al. 1978) for each frequency point in the F F T spectrum, and thus for each frequency a center of gravity localization was calculated with 4 independent parameters: (a) magnitude, (b) right-left (transverse) localization, (c) antero-posterior (sagittal) localization and (d) depth of the equivalent dipole. The two parameters describing the direction of the dipole were not used since the polarity of the potential distribution map depends on the position of the best-fit line in the sine-cosine diagram. However, the polarity does not influence the resulting dipole localization (Lehmann and Michel 1990).
Statistical analysis Data acquisition Silver-silver chloride cup electrodes were applied at 20 sites to the scalp according to the international 10-20 system (Fpl, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz, and 02). The EEGs were referred to linked mastoids. The electrode locations were cleaned to ensure low impedances and the electrodes were fastened by paste. Prior to the recording the impedances were measured, and low and similar values were ensured in all channels (in each channel lower than 3 kS2 and interelectrode difference lower than 1 kS2). The subject lay comfortably on a bed. Data were recorded with a 20-channel Bio-Logic Brain Atlas Ill Plus. The E E G was sampled at a rate of 128 H z / c h a n n e l and stored on magnetic disks for further analysis off-line. Before A-D conversion the E E G was analogue filtered with a bandpass of 1.0-30.0 Uz. Overall amplification was 20.000 times. For data analysis the first 5 successive artifact-free 2 sec epochs, 20 sec after eyes were closed, were selected off-line from the stored EEG.
For confirmatory statistical purposes, data were reduced into frequency bands. A mean localization of the equivalent dipole was calculated for delta (1.0-3.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta I (12.015.5 Hz), beta 2 (16.0-19.5 Hz), and beta 3 (20.0-23.5 Hz). Thereafter an A N O V A for repeated measurements was carried out for the 4 independent parameters in each frequency band. The calculation of significant effects was corrected for degrees of freedom (Geisser and Greenhouse (GGc) 1958). In this way, a conservative test was used to investigate group differences. Post hoc testing between the measurements was done by t test (tt) for paired samples. The calculations were performed using SPSS-PC statistical package.
Results
Clinical monitoring parameters After half-midazolam application most subjects were either sedated or lightly asleep. All subjects were asleep after full-midazolam dose. Flumazenil application led
233
EQUIVALENT DIPOLES AND BENZODIAZEPINES
to an immediate, complete reversal of benzodiazepine sedation. No clinically relevant changes in blood pressure, pulse and oxygen saturation could be observed during the investigation.
a)
Supediclal
20 15 10
Analysis of frequency bands No overall significances were found for the localization of equivalent dipoles in any frequency band with regard to left-right direction (Table I).
5 0 -5 -tO
Delta band (1.0-3.5 Hz, Fig. la) The analysis in the depth direction led to an overall significant result (Table I). The Iocalizations of the equivalent dipoles at the half- and full-midazolam measurements were more superficial compared to baseline, 0 h flumazenil, 4 h flumazenil measurements
30
20
10 0 - 1 0 - 2 0 - 3 0 - 4 0 - 5 0 [r~l
id)
c)
(P(tt) < 0.03). Theta band (4.0-7.5 Hz; Fig. lb) Only in the antero-posterior direction an overall significant result was found (Table I). The localization
o)
TABLE I Resulting F values (dr: 4, 44) and significance levels (Geisser and G r e e n h o u s e corrected P(GGc)) of the A_NOVAs for the delta (1.03.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta 1 (12.0-15.5 Hz), beta a (16.0-19.5 Hz) and beta 3 bands (20.0-23.5 Hz) for magnitude (/~V) and localization in left-right direction (mm), antero-posterior direction (mm) and depth (mm) of the equivalent dipoles at 5 time points. F value
P(GGc)
1.03 0.93 0.67 7.58 4.44 2.44
0.330 0.410 0.480 0.016 * 0.026 * 0.105
1.23 0.75 0.99 0.31 2.19 1.20
0.315 0.466 0.390 0.819 0.120 0.367
1.23 4.02 45.24 46.68 34.98 14.68
0.315 0.015 0.000 0.000 0.000 0.000
6.40 1.92 13.14 13.27 4.80 4.47
0.002 0.146 0.001 0.001 0.008 0.010
Magnitude Delta Theta Alpha Beta 1 Beta e Beta3
++
+++
+
Fig. 1. Two-dimensional plot of mean values and standard errors of equivalent dipole localizations in the antero-posterior direction and for the depth in the (a) delta (1.0-3.5 Hz), (b) theta (4.0-7.5 Hz), (c) alpha (8.0-11.5 Hz), (d) beta I (12.0-15.5 Hz), (e) beta 2 (16.0-19.5 Hz) and (f) beta 3 bands (20.0-23.5 Hz) for all 5 measurements, before drug application (©), after 0.1 m g / k g midazolam (~7), after a total dose of 0.2 m g / k g midazolam ( a ), directly after administration of 1 mg flumazenil (~) and 4 h later (e). Zero value is the mid-point of a spherical head model (10% level in the 10-20 system).
Left-right Delta Theta Alpha Beta 1 Beta 2 Beta 3
of the equivalent dipole at the 0 h flumazenil measurement was the most posterior one.
Alpha band (8.0-11.5 Hz; Fig. lc)
Antero-posterior Delta Theta Alpha Beta 1 Beta e Beta 3
* ** ** ** **
Superficial depth Delta Theta Alpha Beta 1 Beta: Beta 3
** ** ** ** *
In this band there was an evident overall difference in dipole localization between the records in the antero-posterior direction (Table I). Post hoc testing revealed that the equivalent dipoles at the half- and full-midazolam dose records were more anteriorly localized compared to those before midazolam and those after flumazenil ( P ( t t ) < 0 . 0 0 1 ) . The full-midazolam equivalent dipole was more anteriorly localized compared to the half-midazolam measurement ( P ( t t ) =
0.03). In the depth direction an overall significant result was also obtained (Table I). Here too, the equivalent dipoles at the two midazolam records differed from the other ones ( P ( t t ) < 0.01). Those at the half- and full-
234
T. DIERKS ET A L
midazolam records were more superficial in contrast to all other measurements (Fig. lc).
11 [uV]
lO!
Beta~ band (12.0-15.5 Hz; Fig. ld)
Half-midazolam
In opposition to the above described frequency bands, the beta] band rendered an overall significant result when comparing the magnitude of the equivalent dipole between the measurements (Table I). The magnitudes at the midazolam records were higher than at all the other records ( P ( t t ) < 0.02; Fig. 2). Also post hoc a significant difference was found in magnitude between the baseline and the 0 h flumazenil record, the equivalent dipole magnitude at 0 h flumazenil measurement was higher compared to the baseline value (Fig. 2). With regard to the antero-posterior and the depth directions the same overall significant results were obtained as for the alpha band (Table I). Post hoc testing revealed more anterior and superficial localizations of the equivalent dipoles at the two midazolam records compared to the rest of the measurements, with the full-midazolam dose being the more anterior (Fig. ld).
CAR
•
4h-flumazenil
Fz
5
gsin
Full-midazolam Oh-flumazenil
dipoles in regard to magnitude, antero-posterior, and depth directions (Table I). The post hoc tests revealed that the magnitude at the full-midazolam measurement
Similar results as for the beta I band were obtained. Overall significance was found for the equivalent
Mastoid
• •
Delta Theta A l p h a B e t a 1 Beta2 Beta3 Fig. 2. Mean values and standard deviations of magnitude of center of gravity equivalent dipoles in the delta (1.0-3.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta I (12.0-15.5 Hz), beta 2 (16.0-19.5 Hz) and beta 3 bands (20.0-23.5 Hz) for all 5 measurements, before drug application (baseline), after 0.1 mg/kg midazolam (half-midazolam), after a total dose of 0.2 mg/kg midazolam (full-midazolam), directly after administration of l mg flumazenil (0h-flumazenil) and 4 h later (4h-flumazenil).
Beta 2 band (16.0-19.5 Hz; Fig. le)
5
I~~ Baseline
5
0 0 O
...............................
-5
. . . . -5
i 0
.
.
.
.
. 5
5
i
5
i
i
i
i
i
l
l
l
-5
0
5
i
INI'ERO~: 1 7 . 8 @ ' 7 . 2 . ~ H=
8
i
-5
1
~
1
1
1
1
1
0
k
t._J
Fig. 3. Plot of sine and cosine coefficients (from FFT; 20.0 Hz) for each channel in sine-cosine diagrams (upper part) and brain maps of beta activity (17.0-22.0 Hz) using 3 different references: linked mastoids (left), common average reference (middle), and channel Fz (right). See text for details.
EQUIVALENT DIPOLES AND BENZODIAZEPINES was higher than at all other measurements ( P ( t t ) < 0.05; Fig. 2). In the antero-posterior direction the same result could be found as for the alpha and beta I bands, with the midazolam records exhibiting more anteriorly localized equivalent dipoles. In the depth direction a different pattern was found as compared to the alpha and beta t band. Similar to the 2 frequency bands mentioned above, the equivalent dipoles at the two midazolam measurements were more superficially located compared to the others ( P ( t t ) < 0.01), but in the baseline record the equivalent dipole was more deeply localized compared to all measurements except in the 4 h flumazenil one ( P ( t t ) < 0.05; Fig. le).
Beta 3 band (20.0-23.5 Hz; Fig. If) In this frequency band the antero-posterior and the depth directions showed overall significant results (Table I), with the equivalent dipoles at the midazolam measurements more frontal compared to the other measurements ( P ( t t ) < 0.01). In the 0 h flumazenil record the equivalent dipole was still significantly more anterior compared to baseline. In the depth direction, 3 significantly separated groups of equivalent dipoles were found (P(tt) < 0.05). The two midazolam measurements were the most superficial, the baseline measurement the deepest and the 0 h and 4 h flumazenil records localized in between (Fig. lf).
Discussion The aim of the present investigation was to study whether the benzodiazepine receptor antagonist flumazenil can reverse changes in electrical brain activity induced by the benzodiazepine receptor agonist midazolam. Mainly, we found increased activity in the beta bands together with more anterior and superficial equivalent dipoles in the alpha and beta bands after application of midazolam. These changes induced by midazolam were mostly reversed by the application of the antagonist flumazenil. However, especially in the beta frequencies above 20 Hz, differences in depth of equivalent dipoles were found directly after flumazenil application as well as 4 h later. This could suggest that after flumazenil administration electrical brain activity was generated in different structures compared to before drug application. The assumption of a less ambiguous data interpretation by equivalent dipoles compared to conventional F F T can be justified from our data (Fig. 3). In the lower part of the figure topographic maps of the integrated amplitude between 17.0 and 22.0 Hz are dis-
235 played using: (a) linked mastoids (left), (b) common average reference (middle) and (c) channel Fz (right) as references. Obviously different references render a different distribution of values, demonstrating that F F T data provide ambiguous results (Lehmann and Skrandies 1985). The upper part of Fig. 3 displays the entries of the 20 Hz sine and cosine Fourier coefficients in sine-cosine diagrams. The entries, in contrast to FFT data, show the same configuration independent of reference. Of course the absolute values projected on the best-fit line vary, but since the same pattern is retained the interrelation between the entries will be the same regardless of reference and as such resulting in the same localization in the dipole-fit analysis (Fender 1987). For this reason our results are unambiguous with regard to the choice of reference. The method we used in the present investigation to model the equivalent dipole generators of the E E G does not allow an exact localization of the anatomical structures responsible for the electrical brain activity. However, singular center of gravity dipoles have explicit advantages over methods allowing a more exact localization when using F F T data (Liitkenh6ner 1992). The most prominent advantage is that the data may be computed in a relatively simple statistical way, since only one equivalent dipole with few parameters has to be handled (Dierks 1992). As such the method permits a substantial and sensible data reduction, allowing less ambiguous statements with regard to reference in comparison to conventional FFT power data. Our finding of increased magnitude of beta activity after midazolam application is analogous to previous conventional E E G investigations regarding beta activity (Brown et al. 1979; Kochs and Schulte am Esch 1989). Contrary to other reports we did not find a decreased magnitude in the integrated alpha band. This is most probably due to a mixture of a decrease in slow alpha magnitude and an increase in fast alpha magnitude after midazolam. Additionally, the influence of equivalent dipole localization makes a comparison of magnitude and conventional amplitude in F F T maps difficult. Midazolam seems to activate delta activity, also described by Kochs and Schulte am Esch (1989). However, our results suggest the increased activity to be dose-dependent (no increase with half-midazolam dose). The E E G is known to reflect the level of psychological arousal and in general, increased slow beta activity has been associated with increased vigilance (Herrmann 1982). Paradoxically, a similar effect can be elicited by midazolam and other benzodiazepines, which are well-known to impair psychomotor performance (Linnoila et al. 1983). At the same time increased slow beta activity over centro-frontal regions is a sign of physiological sleep (sleep spindles in sleep stage 2; Buchsbaum et al. 1982; Maurer et al. 1989). It
236 is difficult to d i f f e r e n t i a t e w h e t h e r t h e r e m a i n i n g b e t a effects a f t e r f l u m a z e n i l a d m i n i s t r a t i o n a r e m a n i f e s t a tions o f c o n t i n u e d drowsiness, which c o u l d n o t be o b s e r v e d clinically, o r of i n c r e a s e d vigilance, an effect d e s c r i b e d a f t e r f l u m a z e n i l a l o n e ( S c h 6 p f et al. 1984). T h e results o f a study on t h e a u d i t o r y e v o k e d P300 in the s a m e p o p u l a t i o n i n d i c a t e a h i g h e r vigilance directly a n d 4 h after flumazenil, d e m o n s t r a t e d by i n c r e a s e d P300 a m p l i t u d e a n d r e d u c e d P300 l a t e n c y ( E n g e l h a r d t et al. 1992). O n the o t h e r h a n d , activation (e.g., by m i d a z o l a m ) of t h e G A B A - b e n z o d i a z e p i n e r e c e p t o r c o m p l e x is well know to have inhibitory effects on c e n t r a l n e r v o u s system f u n c t i o n s ( S i m m o n d s 1984). A close c o r r e l a t i o n b e t w e e n p l a s m a levels o f b e n z o d i a z e p i n e s a n d b e t a activity in the E E G has b e e n rep o r t e d ( F i n k et al. 1976; M a n d e m a et al. 1991), supp o r t i n g the a s s u m p t i o n that b e t a activity m i r r o r s a r e d u c e d level of C N S activation. B o t h h u m a n a u t o p s y a n d p o s i t r o n emission t o m o g r a p h y ( P E T ) studies suggest t h e occipital lobe to possess the highest density of b e n z o d i a z e p i n e r e c e p t o r s in t h e h u m a n c e r e b r a l cortex ( d ' A r g y et al. 1988). G A B A n e u r o n s a r e a s s u m e d to b e local circuit n e u r o n s ( E n n a 1985) which w o u l d suggest t h a t inhibitory effects of b e n z o d i a z e p i n e s on the cortex m i g h t be m o r e local. In k e e p i n g with this a r g u m e n t is the conclusion t h a t b e n z o d i a z e p i n e effects on E E G p a r a m e t e r s s h o u l d be most p r o n o u n c e d over the p o s t e r i o r p a r t of t h e brain. O n the c o n t r a r y , the p r e s e n t d a t a suggest t h a t t h e localization of the i n c r e a s e d b e t a activity after m i d a z o l a m a d m i n i s t r a t i o n is m o r e a n t e r i o r . E x p l a n a t i o n s for t h e s e results i n c l u d e various s u b t y p e s of r e c e p t o r s with diff e r e n t r e g i o n a l d i s t r i b u t i o n s ( H a n t r a y e et al. 1988; S c h e u l e r 1991), inhibition o f s u b c o r t i c a l G A B A m e d i a t e d inhibition ( M i l l e r a n d F e r r e n d e l l i 1990), not easily assessed using P E T technology. F u r t h e r m o r e , o n e single c e n t e r of gravity d i p o l e simplifies the electrical b r a i n activity substantially, which m a y e x p l a i n the diff e r e n c e s b e t w e e n t h e v a r i o u s m e t h o d s . H o w e v e r , at the p r e s e n t t i m e all e x p l a n a t i o n s a r e very speculative. Most reports concerning flumazenil antagonism of b e n z o d i a z c p i n e s e d a t i o n d e s c r i b e a c o m p l e t e reversal of E E G a l t e r a t i o n s ( L a u r i a n et al. 1984; Kochs a n d Schulte a m E s c h 1989; D i e r k s et al. 1992a; E n g e l h a r d t et al. 1992). W e c a n n o t fully s u p p o r t this a s s u m p t i o n . F l u m a z e n i l a p p l i c a t i o n did not c o m p l e t e l y d e c r e a s e the slow b e t a m a g n i t u d e o f b a s e l i n e values, e i t h e r directly o r a f t e r 4 h. Similarly, the l o c a l i z a t i o n s of t h e equivalent d i p o l e s in the f a s t e r b e t a b a n d s d i d not r e t u r n to t h e b a s e l i n e c o n d i t i o n . C o n s e q u e n t l y , o u r hypothesis of a c o m p l e t e reversal o f b e n z o d i a z e p i n e effects on the E E G by an a n t a g o n i s t was r e j e c t e d . F u r t h e r m o r e , t h e p r e s e n t investigation d e m o n s t r a t e s the ability o f the F F T a p p r o x i m a t i o n to e s t i m a t e w h e t h e r d i f f e r e n t b r a i n s t r u c t u r e s are g e n e r a t i n g b r a i n electrical activity dep e n d i n g on d r u g a n d d r u g dosage.
T. DIERKS ET AL.
References Artru, A. The rate of CSF formation, resistance to reabsorption of CSF, and aperiodic analysis of the EEG following administration of flumazenil to dogs. Anesthesiology, 1990, 72:111-117. Breimer, L.T., Burm, A.G., Danhof, M., Hennis, P.J., Vletter, A.A., de Voogt, J.W., Spierdijk, J. and Bovill, J.G. Pharmacokineticpharmacodynamic modelling of the interaction between flumazenil and midazolam in volunteers by aperiodic EEG analysis. Clin. Pharmacokinet., 1991a, 20: 497-508. Breimer, L.T., Hennis, P.J., Burro, A.G., Danhof, M., Bovill, J.G., Spierdijk, J. and Vletter, A.A. Pharmacokinetics and EEG effects of flumazenil in volunteers. Clin. Pharmacokinet., 1991b, 20: 491-496. Brown, C.R., Sarnquist, F.H., Canup, C.A. and Pedley, T.A. Clinical, electroencephalographic, and pharmacokinetic studies of a water-soluble benzodiazepine, midazolam maleate. Anesthesiology, 1979, 50: 467-470. Buchsbaum, M.S., Mendelson, W.B., Duncan, W.C., Coppola, R., Kelsoe, J. and Gillin, J.C. Topographical cortical mapping of EEG sleep stages during daytime naps in normal subjects. Sleep, 1982, 5: 248-255. d'Argy, R., Gillberg, P.G., St~lnacke, C.G., Persson, A., Bergstr6m, M., L~ngstr6m, B., Schoeps, K.O. and Aquilonius, S.M. In vivo and in vitro receptor autoradiography of the human brain using an tiC-labeled benzodiazepine analogue. Neurosci. Len., 1988, 85: 304-310. Dierks, T. Equivalent EEG sources determined by FFT approximation in healthy subjects, schizophrenic and depressive patients. Brain Topogr., 1992, 4: 207-213. Dierks, T., Engelhardt, W. and Maurer, K. M6glichkeiten des Monitoring topographischer Effekte von zentralwirksamen Pharmaka im EEG. In: G. Laux and P. Riederer (Eds.), Plasmaspiegelbestimmung von Psychopharmaka: Therapeutisches Drug-Monitoring. WVG, Stuttgart, 1992: 75-77. Dierks, T., Ihl, R. and Maurer, K. Nootropika und Brain Mapping. In: P. Riederer, G. Laux and W. P61dinger (Eds.), Neuropsychopharmaka, Vol. 5. Springer, Vienna, 1992b: 179-186. Engelhardt, W., Friess, K., Hartung, E., Sold, M. and Dierks, T. EEG and auditory evoked potential P300 compared with psychometric tests in assessing vigilance after benzodiazepine sedation and antagonism. Br. J. Anaesth., 1992, 69: 75-80. Enna, S.J. In: R.E. Hales and A.J. Frances (Eds.), American Psychiatric Association Annual Review. American Psychiatric Press, Washington, DC, 1985: 67-82. Fender, D.H. Source localization of brain electrical activity. In: A.S. Gevins and A. R~mond (Eds.), Methods of Analysis of Brain Electrical and Magnetic Signals. Elsevier, Amsterdam, 1987: 355-403. Fink, M. EEG and human psychopharmacology. Annu. Rev. Pharmacol., 1969, 9: 241-258. Fink, M., Irwin, P. and Wienfeld, R.E. Blood levels and electroencephalographic effects of diazepam and bromazepam. Clin. Pharmacol. Ther., 1976, 20: 184. Geisser, S. and Greenhouse, S.W. An extension Box's results on the use of the F-distribution in multivariate analysis. Ann. Math. Statist., 1958, 29: 885-891. Gutowitz, H., Zemon, V. and Knight, B.W. Source geometry and dynamics of the visual evoked potential. Electroenceph. clin. Neurophysiol., 1986, 64: 308-327. Haefely, W. Tranquilizers. In: D.G. Grahame-Smith (Ed.), Psychopharmacology 2. Part 1. Preclinical Psychopharmacology. Elsevier, Amsterdam, 1985: 92-182. Hantraye, P., Brouillet, E., Fukuda, H., Chavoix, C., Guibert, B., Dodd, R.H., Prenant, C., Crouzel, M., Naquet, R. and Maziere, M. Benzodiazepine receptors studied in living primates by positron emission topography: antagonist interactions. Eur. J. Pharmacol., 1988, 153: 25-32.
EQUIVALENT DIPOLES AND BENZODIAZEPINES Herrmann, W.M. Electroencephalography and Drug Research. Fischer, Stuttgart, 1982. Hunkeler, W., M6hler, H., Pieri, L., Polc, P., Bonetti, E.P., Cumin, R., Schaffner, R. and Haefely, W. Selective antagonist of benzodiazepines. Nature, 1981, 290: 514-516. Kavanagh, R.N., Darcey, T.M., Lehmann, D. and Fender, D.H. Evaluation of methods for three-dimensional localization of electrical sources in the human brain. IEEE Trans. Biomed. Eng., 1978, 25: 421-429. Klotz, U., Ziegler, G., Ludwig, L. and Reimann, I.W. Pharmacodynamic interaction between midazolam and a specific benzodiazepine antagonist in humans. J. Clin. Pharmacol., 1985, 25: 400-406. Kochs, E. and Schulte am Esch, J. Neurophysiologisches Monitoring der Antagonisierung mit Flumazenil. In: W. Tolkdsdorf and J. Prager (Eds.), Anexate (Flumazenil), der erste spezifische Benzodiazepin-Antagonist. Roche, Basel, 1989: 217-231. Laurian, S., Galliard, J.M., Le, P.K. and Sch6pf, J. Effects of a benzodiazepine antagonist on the diazepam-induced electrical brain activity modifications. Neuropsychobiology, 1984, 11: 55-58. Lehmann, D. Principles of spatial analysis. In: A.S. Gevins and A. R6mond (Eds.), Methods of Analysis of Brain Electrical and Magnetic Signals. Elsevier, Amsterdam, 1987: 309-354. Lehmann, D. and Michel, C.M. Intracerebral dipole sources of EEG FFT power maps. Brain Topogr., 1989, 2: 155-164. Lehmann, D. and Michel, C.M. Intracerebral dipole source localization for FFT power maps. Electroenceph. clin. Neurophysiol., 1990, 76: 271-276. Lehmann, D. and Skrandies, W. Spatial analysis of evoked potentials in man - - a review. Prog. Neurobiol., 1985, 23: 227-250. Lehmann, D., Ozaki, H. and Pal, I. Averaging of spectral power and phase via vector diagram best fits without reference electrode of reference channel. Electroenceph. clin. Neurophysiol., 1986, 64: 350-363. Linnoila, M., Erwin, C.W., Brendle, A. and Simpson, D. Psychomo-
237 tor effects of diazepam in anxious patients and healthy volunteers. J. Clin. Psychopharmacol., 1983, 3: 88-96. Liitkenh6ner, B. Frequency-domain localization of intracerebral dipolar sources. Electroenceph. clin. Neurophysiol., 1992, 82: 112-118. Mandema, J.W., Sansom, L.N., Dios-Vieitez, M.C., Hollander-Jansen, M. and Danhof, M. Pharmacokinetic modeling of the electroencephalographic effects of benzodiazepines. Correlation with receptor binding and anticonvulsant activity. J. Pharmacol. Exp. Ther., 1991, 257: 472-478. Maurer, K., Dierks, T. and Rupprecht, R. Computerized electroencephalographic topography (CET) during sleep. Psychiat. Res., 1989, 29: 435-438. Miller, J.W. and Ferrendelli, J.A. Characterization of GABA-ergic seizure regulation in midline thalamus. Neuropharmacology, 1990, 29: 649-655. Norcia, A.M., Sutter, E.E. and Tyler, C.W. Electrophysiological evidence for the existence of coarse and fine disparity mechanism in humans. Vision Res., 1985, 25: 1603-1611. Scbeuler, W. EEG sleep activities react topographically different to GABAergic sleep modulation by flunitrazepam: relationship to regional distribution of benzodiazepine receptor subtypes? Neuropsychobiology, 1991, 23: 213-221. Sch6pf, J., Laurian, S., Le, P.K. and Galliard, J.M. Intrinsic activity of the benzodiazepine antagonist Ro 15-1788 in man: an electrophysiological investigation. Pharmacopsychiatry, 1984, 17: 79-83. Simmonds, M.A. Physiological and pharmacological characterizations of the actions of GABA. In: N.G. Bowery (Ed.), Actions and Interactions of GABA and Benzodiazepines. Raven Press, New York, 1984: 27-43. Walser, A., Benjamin, L.E., Flynn, T., Mason, C., Schwarz, R. and Fryer, R.I. Quinazolines and 1,4-benzodiazepines. 84, Synthesis and reactions of imidazo[1,5-a][l,4]-benzodiazepines. J. Org. Chem., 1978, 43: 936-944.
231
EEG92530
Equivalent dipoles of FFT data visualize drug interaction at benzodiazepine receptors Thomas Dierks a, Wolfram Engelhardt b and Konrad Maurer a a Dept. of Clinical Neurophysiology, Psychiatric Hospital, Wiirzburg (Germany), and b Dept. of Anaesthesiology, University of Wiirzburg, Wiirzburg (Germany) (Accepted for publication: 2 November 1992)
Summary The aim of the present investigation was to study if the benzodiazepine receptor antagonist flumazenil could reverse the effects on the brain electrical activity induced by the benzodiazepine receptor agonist midazolam. The method of FFT approximation was used for this purpose. It allows the calculation of center of gravity equivalent dipoles of spectral EEG data. The results are reference independent and allow therefore a more unambiguous interpretation compared to conventional FFT data reports. Twelve subjects were investigated before and after 0.1 mg/kg and 0.2 mg/kg midazolam respectively, directly and 4 h after administration of 1 mg flumazenil. Our results imply that the application of flumazenil after midazolam sedation leads to an almost complete restoration of the brain electrical activity. However, especially in the beta frequencies above 20 Hz differences in depth of equivalent dipoles were found directly after flumazenil application as well as 4 h later. This could suggest that neuronal generators in different brain structures were responsible for the electrical activity after flumazenil administration compared to before. Key words: FFT approximation; Equivalent dipoles; Reference-free; EEG; Midazolam; Flumazenil
Effects of benzodiazepines on the human electroencephalogram (EEG) have been investigated for more than 20 years (Fink 1969). Midazolam, synthesized in 1976 (Walser et al. 1978), is an imidazobenzodiazepine with unique qualities, e.g., a rapid onset of action and a high metabolic clearance compared to other benzodiazepines (Haefely 1985). Since it has the anxiolytic, hypnotic, anticonvulsive, muscle relaxing and antegrade amnestic effects which are characteristic of all benzodiazepines, it is used clinically for anesthetic premedication, induction and maintenance and sedation for diagnostic and therapeutic procedures. Electrophysiological characteristics of midazolam are widespread beta activity, initially approximately at 22 Hz followed by a 15 Hz rhythm, with simultaneous attenuation and disappearance of alpha rhythm (Brown et al. 1979). Flumazenil selectively antagonizes benzodiazepine effects like sedation and amnesia due to competitive binding at benzodiazepine receptor sites (Hunkeler et
Correspondence to: Dr. Thomas Dierks, M.D., Dept. of Clinical Neurophysiology, Psychiatric Hospital, University of Wiirzburg, Fiichsleinstr. 15, D-W-8700 Wiirzburg (Germany). Tel.: + 931-203334; Fax: + 931-203425.
al. 1981). According to Sch6pf et al. (1984) flumazenil alone produces EEG effects which are similar to nootropic (Dierks et al. 1992b) and psychostimulating drugs (Herrmann 1982). However beta activity remains unmodified. On the other hand, Artru (1990) in dogs and Breimer et al. (1991b) in human volunteers reported that flumazenil effects on the EEG are negligible. Quantitative EEG (QEEG) and computerized electroencephalographic topography (CET) investigations concerning the interaction between benzodiazepines and their antagonists have been carried out (Laurian et al. 1984; Klotz et al. 1985; Breimer et al. 1991a; Engelhardt et al. 1992). Both QEEG and CET studies in the frequency domain allow only descriptive interpretations of results, since they are dependent on the reference used (Lehmann 1987). A more physiological-functional interpretation of EEG data may be achieved by reference-independent data analysis. Conventional FFT reports concerning flumazenil antagonism of benzodiazepine sedation describe a complete reversal of EEG alterations (e.g., Kochs and Schulte am Esch 1989). Consequently the hypothesis of this investigation was that a benzodiazepine antagonist would completely reverse changes in EEG activity induced by a benzodiazepine. This was investigated using
232 the F F T approximation developed by Lehmann and Michel (1989, 1990) followed by calculation of equivalent center of gravity dipoles for all frequencies.
Methods
Subjects Twelve healthy volunteers (6 female and 6 male, mean age 24.3 + 2.6 years, range: 21-30 years, mean body weight 64.2 + 7.7 kg), were investigated. No subject had a history of medical problems or a history of substance abuse or dependency. Furthermore, at the time of investigation none used regular medication. The Ethics Committee at the University of Wiirzburg approved the study. Informed written consent was obtained from the subjects. All volunteers were kept fasting overnight until 1 h after benzodiazepine antagonist administration, when water or juice was offered. Room temperature was kept constant at 25°C over the study period.
Study design E E G was recorded before any drug was administered (baseline), after i.v. administration of 0.1 m g / k g midazolam (half-midazolam), after administration of a total dose of 0.2 m g / k g midazolam (full-midazolam (20 min after half-midazolam)), immediately after 1 mg flumazenil (0 h flumazenil (20 min after full-midazolam)), and 240 min after flumazenil (4 h flumazenil). During the study electrocardiogram, oxygen saturation and blood pressure were monitored.
T. DIERKS ET AL.
Data processing Fast Fourier analysis was done on each 2 sec epoch, multiplied with a Hamming window. The resulting sine and cosine coefficients for each electrode and each frequency point (0.5 Hz resolution) were placed in a sine-cosine diagram (after Nyquist; Norcia et al. 1985; Gutowitz et al. 1986; Lehmann et al. 1986). When a single-generator model is supposed to explain such data, all entries in the diagram have to be orthogonally projected upon a line which yields the least sum of squared deviations between original positions and their projected positions on the line. The projection of the FFT constellation on this line is called F F T approximation (Lehmann and Michel 1989) and describes a map of potential distribution, which is used for estimation of the equivalent dipole source. For a detailed description and discussion of this method see Lehmann and Michel (1989, 1990). For equivalent dipole source localization we used a moving (instantaneous) dipole model (Kavanagh et al. 1978) for each frequency point in the F F T spectrum, and thus for each frequency a center of gravity localization was calculated with 4 independent parameters: (a) magnitude, (b) right-left (transverse) localization, (c) antero-posterior (sagittal) localization and (d) depth of the equivalent dipole. The two parameters describing the direction of the dipole were not used since the polarity of the potential distribution map depends on the position of the best-fit line in the sine-cosine diagram. However, the polarity does not influence the resulting dipole localization (Lehmann and Michel 1990).
Statistical analysis Data acquisition Silver-silver chloride cup electrodes were applied at 20 sites to the scalp according to the international 10-20 system (Fpl, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz, and 02). The EEGs were referred to linked mastoids. The electrode locations were cleaned to ensure low impedances and the electrodes were fastened by paste. Prior to the recording the impedances were measured, and low and similar values were ensured in all channels (in each channel lower than 3 kS2 and interelectrode difference lower than 1 kS2). The subject lay comfortably on a bed. Data were recorded with a 20-channel Bio-Logic Brain Atlas Ill Plus. The E E G was sampled at a rate of 128 H z / c h a n n e l and stored on magnetic disks for further analysis off-line. Before A-D conversion the E E G was analogue filtered with a bandpass of 1.0-30.0 Uz. Overall amplification was 20.000 times. For data analysis the first 5 successive artifact-free 2 sec epochs, 20 sec after eyes were closed, were selected off-line from the stored EEG.
For confirmatory statistical purposes, data were reduced into frequency bands. A mean localization of the equivalent dipole was calculated for delta (1.0-3.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta I (12.015.5 Hz), beta 2 (16.0-19.5 Hz), and beta 3 (20.0-23.5 Hz). Thereafter an A N O V A for repeated measurements was carried out for the 4 independent parameters in each frequency band. The calculation of significant effects was corrected for degrees of freedom (Geisser and Greenhouse (GGc) 1958). In this way, a conservative test was used to investigate group differences. Post hoc testing between the measurements was done by t test (tt) for paired samples. The calculations were performed using SPSS-PC statistical package.
Results
Clinical monitoring parameters After half-midazolam application most subjects were either sedated or lightly asleep. All subjects were asleep after full-midazolam dose. Flumazenil application led
233
EQUIVALENT DIPOLES AND BENZODIAZEPINES
to an immediate, complete reversal of benzodiazepine sedation. No clinically relevant changes in blood pressure, pulse and oxygen saturation could be observed during the investigation.
a)
Supediclal
20 15 10
Analysis of frequency bands No overall significances were found for the localization of equivalent dipoles in any frequency band with regard to left-right direction (Table I).
5 0 -5 -tO
Delta band (1.0-3.5 Hz, Fig. la) The analysis in the depth direction led to an overall significant result (Table I). The Iocalizations of the equivalent dipoles at the half- and full-midazolam measurements were more superficial compared to baseline, 0 h flumazenil, 4 h flumazenil measurements
30
20
10 0 - 1 0 - 2 0 - 3 0 - 4 0 - 5 0 [r~l
id)
c)
(P(tt) < 0.03). Theta band (4.0-7.5 Hz; Fig. lb) Only in the antero-posterior direction an overall significant result was found (Table I). The localization
o)
TABLE I Resulting F values (dr: 4, 44) and significance levels (Geisser and G r e e n h o u s e corrected P(GGc)) of the A_NOVAs for the delta (1.03.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta 1 (12.0-15.5 Hz), beta a (16.0-19.5 Hz) and beta 3 bands (20.0-23.5 Hz) for magnitude (/~V) and localization in left-right direction (mm), antero-posterior direction (mm) and depth (mm) of the equivalent dipoles at 5 time points. F value
P(GGc)
1.03 0.93 0.67 7.58 4.44 2.44
0.330 0.410 0.480 0.016 * 0.026 * 0.105
1.23 0.75 0.99 0.31 2.19 1.20
0.315 0.466 0.390 0.819 0.120 0.367
1.23 4.02 45.24 46.68 34.98 14.68
0.315 0.015 0.000 0.000 0.000 0.000
6.40 1.92 13.14 13.27 4.80 4.47
0.002 0.146 0.001 0.001 0.008 0.010
Magnitude Delta Theta Alpha Beta 1 Beta e Beta3
++
+++
+
Fig. 1. Two-dimensional plot of mean values and standard errors of equivalent dipole localizations in the antero-posterior direction and for the depth in the (a) delta (1.0-3.5 Hz), (b) theta (4.0-7.5 Hz), (c) alpha (8.0-11.5 Hz), (d) beta I (12.0-15.5 Hz), (e) beta 2 (16.0-19.5 Hz) and (f) beta 3 bands (20.0-23.5 Hz) for all 5 measurements, before drug application (©), after 0.1 m g / k g midazolam (~7), after a total dose of 0.2 m g / k g midazolam ( a ), directly after administration of 1 mg flumazenil (~) and 4 h later (e). Zero value is the mid-point of a spherical head model (10% level in the 10-20 system).
Left-right Delta Theta Alpha Beta 1 Beta 2 Beta 3
of the equivalent dipole at the 0 h flumazenil measurement was the most posterior one.
Alpha band (8.0-11.5 Hz; Fig. lc)
Antero-posterior Delta Theta Alpha Beta 1 Beta e Beta 3
* ** ** ** **
Superficial depth Delta Theta Alpha Beta 1 Beta: Beta 3
** ** ** ** *
In this band there was an evident overall difference in dipole localization between the records in the antero-posterior direction (Table I). Post hoc testing revealed that the equivalent dipoles at the half- and full-midazolam dose records were more anteriorly localized compared to those before midazolam and those after flumazenil ( P ( t t ) < 0 . 0 0 1 ) . The full-midazolam equivalent dipole was more anteriorly localized compared to the half-midazolam measurement ( P ( t t ) =
0.03). In the depth direction an overall significant result was also obtained (Table I). Here too, the equivalent dipoles at the two midazolam records differed from the other ones ( P ( t t ) < 0.01). Those at the half- and full-
234
T. DIERKS ET A L
midazolam records were more superficial in contrast to all other measurements (Fig. lc).
11 [uV]
lO!
Beta~ band (12.0-15.5 Hz; Fig. ld)
Half-midazolam
In opposition to the above described frequency bands, the beta] band rendered an overall significant result when comparing the magnitude of the equivalent dipole between the measurements (Table I). The magnitudes at the midazolam records were higher than at all the other records ( P ( t t ) < 0.02; Fig. 2). Also post hoc a significant difference was found in magnitude between the baseline and the 0 h flumazenil record, the equivalent dipole magnitude at 0 h flumazenil measurement was higher compared to the baseline value (Fig. 2). With regard to the antero-posterior and the depth directions the same overall significant results were obtained as for the alpha band (Table I). Post hoc testing revealed more anterior and superficial localizations of the equivalent dipoles at the two midazolam records compared to the rest of the measurements, with the full-midazolam dose being the more anterior (Fig. ld).
CAR
•
4h-flumazenil
Fz
5
gsin
Full-midazolam Oh-flumazenil
dipoles in regard to magnitude, antero-posterior, and depth directions (Table I). The post hoc tests revealed that the magnitude at the full-midazolam measurement
Similar results as for the beta I band were obtained. Overall significance was found for the equivalent
Mastoid
• •
Delta Theta A l p h a B e t a 1 Beta2 Beta3 Fig. 2. Mean values and standard deviations of magnitude of center of gravity equivalent dipoles in the delta (1.0-3.5 Hz), theta (4.0-7.5 Hz), alpha (8.0-11.5 Hz), beta I (12.0-15.5 Hz), beta 2 (16.0-19.5 Hz) and beta 3 bands (20.0-23.5 Hz) for all 5 measurements, before drug application (baseline), after 0.1 mg/kg midazolam (half-midazolam), after a total dose of 0.2 mg/kg midazolam (full-midazolam), directly after administration of l mg flumazenil (0h-flumazenil) and 4 h later (4h-flumazenil).
Beta 2 band (16.0-19.5 Hz; Fig. le)
5
I~~ Baseline
5
0 0 O
...............................
-5
. . . . -5
i 0
.
.
.
.
. 5
5
i
5
i
i
i
i
i
l
l
l
-5
0
5
i
INI'ERO~: 1 7 . 8 @ ' 7 . 2 . ~ H=
8
i
-5
1
~
1
1
1
1
1
0
k
t._J
Fig. 3. Plot of sine and cosine coefficients (from FFT; 20.0 Hz) for each channel in sine-cosine diagrams (upper part) and brain maps of beta activity (17.0-22.0 Hz) using 3 different references: linked mastoids (left), common average reference (middle), and channel Fz (right). See text for details.
EQUIVALENT DIPOLES AND BENZODIAZEPINES was higher than at all other measurements ( P ( t t ) < 0.05; Fig. 2). In the antero-posterior direction the same result could be found as for the alpha and beta I bands, with the midazolam records exhibiting more anteriorly localized equivalent dipoles. In the depth direction a different pattern was found as compared to the alpha and beta t band. Similar to the 2 frequency bands mentioned above, the equivalent dipoles at the two midazolam measurements were more superficially located compared to the others ( P ( t t ) < 0.01), but in the baseline record the equivalent dipole was more deeply localized compared to all measurements except in the 4 h flumazenil one ( P ( t t ) < 0.05; Fig. le).
Beta 3 band (20.0-23.5 Hz; Fig. If) In this frequency band the antero-posterior and the depth directions showed overall significant results (Table I), with the equivalent dipoles at the midazolam measurements more frontal compared to the other measurements ( P ( t t ) < 0.01). In the 0 h flumazenil record the equivalent dipole was still significantly more anterior compared to baseline. In the depth direction, 3 significantly separated groups of equivalent dipoles were found (P(tt) < 0.05). The two midazolam measurements were the most superficial, the baseline measurement the deepest and the 0 h and 4 h flumazenil records localized in between (Fig. lf).
Discussion The aim of the present investigation was to study whether the benzodiazepine receptor antagonist flumazenil can reverse changes in electrical brain activity induced by the benzodiazepine receptor agonist midazolam. Mainly, we found increased activity in the beta bands together with more anterior and superficial equivalent dipoles in the alpha and beta bands after application of midazolam. These changes induced by midazolam were mostly reversed by the application of the antagonist flumazenil. However, especially in the beta frequencies above 20 Hz, differences in depth of equivalent dipoles were found directly after flumazenil application as well as 4 h later. This could suggest that after flumazenil administration electrical brain activity was generated in different structures compared to before drug application. The assumption of a less ambiguous data interpretation by equivalent dipoles compared to conventional F F T can be justified from our data (Fig. 3). In the lower part of the figure topographic maps of the integrated amplitude between 17.0 and 22.0 Hz are dis-
235 played using: (a) linked mastoids (left), (b) common average reference (middle) and (c) channel Fz (right) as references. Obviously different references render a different distribution of values, demonstrating that F F T data provide ambiguous results (Lehmann and Skrandies 1985). The upper part of Fig. 3 displays the entries of the 20 Hz sine and cosine Fourier coefficients in sine-cosine diagrams. The entries, in contrast to FFT data, show the same configuration independent of reference. Of course the absolute values projected on the best-fit line vary, but since the same pattern is retained the interrelation between the entries will be the same regardless of reference and as such resulting in the same localization in the dipole-fit analysis (Fender 1987). For this reason our results are unambiguous with regard to the choice of reference. The method we used in the present investigation to model the equivalent dipole generators of the E E G does not allow an exact localization of the anatomical structures responsible for the electrical brain activity. However, singular center of gravity dipoles have explicit advantages over methods allowing a more exact localization when using F F T data (Liitkenh6ner 1992). The most prominent advantage is that the data may be computed in a relatively simple statistical way, since only one equivalent dipole with few parameters has to be handled (Dierks 1992). As such the method permits a substantial and sensible data reduction, allowing less ambiguous statements with regard to reference in comparison to conventional FFT power data. Our finding of increased magnitude of beta activity after midazolam application is analogous to previous conventional E E G investigations regarding beta activity (Brown et al. 1979; Kochs and Schulte am Esch 1989). Contrary to other reports we did not find a decreased magnitude in the integrated alpha band. This is most probably due to a mixture of a decrease in slow alpha magnitude and an increase in fast alpha magnitude after midazolam. Additionally, the influence of equivalent dipole localization makes a comparison of magnitude and conventional amplitude in F F T maps difficult. Midazolam seems to activate delta activity, also described by Kochs and Schulte am Esch (1989). However, our results suggest the increased activity to be dose-dependent (no increase with half-midazolam dose). The E E G is known to reflect the level of psychological arousal and in general, increased slow beta activity has been associated with increased vigilance (Herrmann 1982). Paradoxically, a similar effect can be elicited by midazolam and other benzodiazepines, which are well-known to impair psychomotor performance (Linnoila et al. 1983). At the same time increased slow beta activity over centro-frontal regions is a sign of physiological sleep (sleep spindles in sleep stage 2; Buchsbaum et al. 1982; Maurer et al. 1989). It
236 is difficult to d i f f e r e n t i a t e w h e t h e r t h e r e m a i n i n g b e t a effects a f t e r f l u m a z e n i l a d m i n i s t r a t i o n a r e m a n i f e s t a tions o f c o n t i n u e d drowsiness, which c o u l d n o t be o b s e r v e d clinically, o r of i n c r e a s e d vigilance, an effect d e s c r i b e d a f t e r f l u m a z e n i l a l o n e ( S c h 6 p f et al. 1984). T h e results o f a study on t h e a u d i t o r y e v o k e d P300 in the s a m e p o p u l a t i o n i n d i c a t e a h i g h e r vigilance directly a n d 4 h after flumazenil, d e m o n s t r a t e d by i n c r e a s e d P300 a m p l i t u d e a n d r e d u c e d P300 l a t e n c y ( E n g e l h a r d t et al. 1992). O n the o t h e r h a n d , activation (e.g., by m i d a z o l a m ) of t h e G A B A - b e n z o d i a z e p i n e r e c e p t o r c o m p l e x is well know to have inhibitory effects on c e n t r a l n e r v o u s system f u n c t i o n s ( S i m m o n d s 1984). A close c o r r e l a t i o n b e t w e e n p l a s m a levels o f b e n z o d i a z e p i n e s a n d b e t a activity in the E E G has b e e n rep o r t e d ( F i n k et al. 1976; M a n d e m a et al. 1991), supp o r t i n g the a s s u m p t i o n that b e t a activity m i r r o r s a r e d u c e d level of C N S activation. B o t h h u m a n a u t o p s y a n d p o s i t r o n emission t o m o g r a p h y ( P E T ) studies suggest t h e occipital lobe to possess the highest density of b e n z o d i a z e p i n e r e c e p t o r s in t h e h u m a n c e r e b r a l cortex ( d ' A r g y et al. 1988). G A B A n e u r o n s a r e a s s u m e d to b e local circuit n e u r o n s ( E n n a 1985) which w o u l d suggest t h a t inhibitory effects of b e n z o d i a z e p i n e s on the cortex m i g h t be m o r e local. In k e e p i n g with this a r g u m e n t is the conclusion t h a t b e n z o d i a z e p i n e effects on E E G p a r a m e t e r s s h o u l d be most p r o n o u n c e d over the p o s t e r i o r p a r t of t h e brain. O n the c o n t r a r y , the p r e s e n t d a t a suggest t h a t t h e localization of the i n c r e a s e d b e t a activity after m i d a z o l a m a d m i n i s t r a t i o n is m o r e a n t e r i o r . E x p l a n a t i o n s for t h e s e results i n c l u d e various s u b t y p e s of r e c e p t o r s with diff e r e n t r e g i o n a l d i s t r i b u t i o n s ( H a n t r a y e et al. 1988; S c h e u l e r 1991), inhibition o f s u b c o r t i c a l G A B A m e d i a t e d inhibition ( M i l l e r a n d F e r r e n d e l l i 1990), not easily assessed using P E T technology. F u r t h e r m o r e , o n e single c e n t e r of gravity d i p o l e simplifies the electrical b r a i n activity substantially, which m a y e x p l a i n the diff e r e n c e s b e t w e e n t h e v a r i o u s m e t h o d s . H o w e v e r , at the p r e s e n t t i m e all e x p l a n a t i o n s a r e very speculative. Most reports concerning flumazenil antagonism of b e n z o d i a z c p i n e s e d a t i o n d e s c r i b e a c o m p l e t e reversal of E E G a l t e r a t i o n s ( L a u r i a n et al. 1984; Kochs a n d Schulte a m E s c h 1989; D i e r k s et al. 1992a; E n g e l h a r d t et al. 1992). W e c a n n o t fully s u p p o r t this a s s u m p t i o n . F l u m a z e n i l a p p l i c a t i o n did not c o m p l e t e l y d e c r e a s e the slow b e t a m a g n i t u d e o f b a s e l i n e values, e i t h e r directly o r a f t e r 4 h. Similarly, the l o c a l i z a t i o n s of t h e equivalent d i p o l e s in the f a s t e r b e t a b a n d s d i d not r e t u r n to t h e b a s e l i n e c o n d i t i o n . C o n s e q u e n t l y , o u r hypothesis of a c o m p l e t e reversal o f b e n z o d i a z e p i n e effects on the E E G by an a n t a g o n i s t was r e j e c t e d . F u r t h e r m o r e , t h e p r e s e n t investigation d e m o n s t r a t e s the ability o f the F F T a p p r o x i m a t i o n to e s t i m a t e w h e t h e r d i f f e r e n t b r a i n s t r u c t u r e s are g e n e r a t i n g b r a i n electrical activity dep e n d i n g on d r u g a n d d r u g dosage.
T. DIERKS ET AL.
References Artru, A. The rate of CSF formation, resistance to reabsorption of CSF, and aperiodic analysis of the EEG following administration of flumazenil to dogs. Anesthesiology, 1990, 72:111-117. Breimer, L.T., Burm, A.G., Danhof, M., Hennis, P.J., Vletter, A.A., de Voogt, J.W., Spierdijk, J. and Bovill, J.G. Pharmacokineticpharmacodynamic modelling of the interaction between flumazenil and midazolam in volunteers by aperiodic EEG analysis. Clin. Pharmacokinet., 1991a, 20: 497-508. Breimer, L.T., Hennis, P.J., Burro, A.G., Danhof, M., Bovill, J.G., Spierdijk, J. and Vletter, A.A. Pharmacokinetics and EEG effects of flumazenil in volunteers. Clin. Pharmacokinet., 1991b, 20: 491-496. Brown, C.R., Sarnquist, F.H., Canup, C.A. and Pedley, T.A. Clinical, electroencephalographic, and pharmacokinetic studies of a water-soluble benzodiazepine, midazolam maleate. Anesthesiology, 1979, 50: 467-470. Buchsbaum, M.S., Mendelson, W.B., Duncan, W.C., Coppola, R., Kelsoe, J. and Gillin, J.C. Topographical cortical mapping of EEG sleep stages during daytime naps in normal subjects. Sleep, 1982, 5: 248-255. d'Argy, R., Gillberg, P.G., St~lnacke, C.G., Persson, A., Bergstr6m, M., L~ngstr6m, B., Schoeps, K.O. and Aquilonius, S.M. In vivo and in vitro receptor autoradiography of the human brain using an tiC-labeled benzodiazepine analogue. Neurosci. Len., 1988, 85: 304-310. Dierks, T. Equivalent EEG sources determined by FFT approximation in healthy subjects, schizophrenic and depressive patients. Brain Topogr., 1992, 4: 207-213. Dierks, T., Engelhardt, W. and Maurer, K. M6glichkeiten des Monitoring topographischer Effekte von zentralwirksamen Pharmaka im EEG. In: G. Laux and P. Riederer (Eds.), Plasmaspiegelbestimmung von Psychopharmaka: Therapeutisches Drug-Monitoring. WVG, Stuttgart, 1992: 75-77. Dierks, T., Ihl, R. and Maurer, K. Nootropika und Brain Mapping. In: P. Riederer, G. Laux and W. P61dinger (Eds.), Neuropsychopharmaka, Vol. 5. Springer, Vienna, 1992b: 179-186. Engelhardt, W., Friess, K., Hartung, E., Sold, M. and Dierks, T. EEG and auditory evoked potential P300 compared with psychometric tests in assessing vigilance after benzodiazepine sedation and antagonism. Br. J. Anaesth., 1992, 69: 75-80. Enna, S.J. In: R.E. Hales and A.J. Frances (Eds.), American Psychiatric Association Annual Review. American Psychiatric Press, Washington, DC, 1985: 67-82. Fender, D.H. Source localization of brain electrical activity. In: A.S. Gevins and A. R~mond (Eds.), Methods of Analysis of Brain Electrical and Magnetic Signals. Elsevier, Amsterdam, 1987: 355-403. Fink, M. EEG and human psychopharmacology. Annu. Rev. Pharmacol., 1969, 9: 241-258. Fink, M., Irwin, P. and Wienfeld, R.E. Blood levels and electroencephalographic effects of diazepam and bromazepam. Clin. Pharmacol. Ther., 1976, 20: 184. Geisser, S. and Greenhouse, S.W. An extension Box's results on the use of the F-distribution in multivariate analysis. Ann. Math. Statist., 1958, 29: 885-891. Gutowitz, H., Zemon, V. and Knight, B.W. Source geometry and dynamics of the visual evoked potential. Electroenceph. clin. Neurophysiol., 1986, 64: 308-327. Haefely, W. Tranquilizers. In: D.G. Grahame-Smith (Ed.), Psychopharmacology 2. Part 1. Preclinical Psychopharmacology. Elsevier, Amsterdam, 1985: 92-182. Hantraye, P., Brouillet, E., Fukuda, H., Chavoix, C., Guibert, B., Dodd, R.H., Prenant, C., Crouzel, M., Naquet, R. and Maziere, M. Benzodiazepine receptors studied in living primates by positron emission topography: antagonist interactions. Eur. J. Pharmacol., 1988, 153: 25-32.
EQUIVALENT DIPOLES AND BENZODIAZEPINES Herrmann, W.M. Electroencephalography and Drug Research. Fischer, Stuttgart, 1982. Hunkeler, W., M6hler, H., Pieri, L., Polc, P., Bonetti, E.P., Cumin, R., Schaffner, R. and Haefely, W. Selective antagonist of benzodiazepines. Nature, 1981, 290: 514-516. Kavanagh, R.N., Darcey, T.M., Lehmann, D. and Fender, D.H. Evaluation of methods for three-dimensional localization of electrical sources in the human brain. IEEE Trans. Biomed. Eng., 1978, 25: 421-429. Klotz, U., Ziegler, G., Ludwig, L. and Reimann, I.W. Pharmacodynamic interaction between midazolam and a specific benzodiazepine antagonist in humans. J. Clin. Pharmacol., 1985, 25: 400-406. Kochs, E. and Schulte am Esch, J. Neurophysiologisches Monitoring der Antagonisierung mit Flumazenil. In: W. Tolkdsdorf and J. Prager (Eds.), Anexate (Flumazenil), der erste spezifische Benzodiazepin-Antagonist. Roche, Basel, 1989: 217-231. Laurian, S., Galliard, J.M., Le, P.K. and Sch6pf, J. Effects of a benzodiazepine antagonist on the diazepam-induced electrical brain activity modifications. Neuropsychobiology, 1984, 11: 55-58. Lehmann, D. Principles of spatial analysis. In: A.S. Gevins and A. R6mond (Eds.), Methods of Analysis of Brain Electrical and Magnetic Signals. Elsevier, Amsterdam, 1987: 309-354. Lehmann, D. and Michel, C.M. Intracerebral dipole sources of EEG FFT power maps. Brain Topogr., 1989, 2: 155-164. Lehmann, D. and Michel, C.M. Intracerebral dipole source localization for FFT power maps. Electroenceph. clin. Neurophysiol., 1990, 76: 271-276. Lehmann, D. and Skrandies, W. Spatial analysis of evoked potentials in man - - a review. Prog. Neurobiol., 1985, 23: 227-250. Lehmann, D., Ozaki, H. and Pal, I. Averaging of spectral power and phase via vector diagram best fits without reference electrode of reference channel. Electroenceph. clin. Neurophysiol., 1986, 64: 350-363. Linnoila, M., Erwin, C.W., Brendle, A. and Simpson, D. Psychomo-
237 tor effects of diazepam in anxious patients and healthy volunteers. J. Clin. Psychopharmacol., 1983, 3: 88-96. Liitkenh6ner, B. Frequency-domain localization of intracerebral dipolar sources. Electroenceph. clin. Neurophysiol., 1992, 82: 112-118. Mandema, J.W., Sansom, L.N., Dios-Vieitez, M.C., Hollander-Jansen, M. and Danhof, M. Pharmacokinetic modeling of the electroencephalographic effects of benzodiazepines. Correlation with receptor binding and anticonvulsant activity. J. Pharmacol. Exp. Ther., 1991, 257: 472-478. Maurer, K., Dierks, T. and Rupprecht, R. Computerized electroencephalographic topography (CET) during sleep. Psychiat. Res., 1989, 29: 435-438. Miller, J.W. and Ferrendelli, J.A. Characterization of GABA-ergic seizure regulation in midline thalamus. Neuropharmacology, 1990, 29: 649-655. Norcia, A.M., Sutter, E.E. and Tyler, C.W. Electrophysiological evidence for the existence of coarse and fine disparity mechanism in humans. Vision Res., 1985, 25: 1603-1611. Scbeuler, W. EEG sleep activities react topographically different to GABAergic sleep modulation by flunitrazepam: relationship to regional distribution of benzodiazepine receptor subtypes? Neuropsychobiology, 1991, 23: 213-221. Sch6pf, J., Laurian, S., Le, P.K. and Galliard, J.M. Intrinsic activity of the benzodiazepine antagonist Ro 15-1788 in man: an electrophysiological investigation. Pharmacopsychiatry, 1984, 17: 79-83. Simmonds, M.A. Physiological and pharmacological characterizations of the actions of GABA. In: N.G. Bowery (Ed.), Actions and Interactions of GABA and Benzodiazepines. Raven Press, New York, 1984: 27-43. Walser, A., Benjamin, L.E., Flynn, T., Mason, C., Schwarz, R. and Fryer, R.I. Quinazolines and 1,4-benzodiazepines. 84, Synthesis and reactions of imidazo[1,5-a][l,4]-benzodiazepines. J. Org. Chem., 1978, 43: 936-944.