- Email: [email protected]
Dissociated changes of somatosensory evoked low-frequency scalp responses and 600 Hz bursts after single-dose administration of lorazepam
Brain Research 946 (2002) 1–11 www.elsevier.com / locate / bres
Research report
Dissociated changes of somatosensory evoked low-frequency scalp responses and 600 Hz bursts after single-dose administration of lorazepam Domenico Restuccia a , *, Massimiliano Valeriani a,b , Eugenio Grassi a , Salvatore Mazza a , Pietro Tonali a a
Department of Neurology, Catholic University, Policlinico A. Gemelli, L.go A. Gemelli 8, 00168 Rome, Italy b Division of Neurology, Ospedale Pediatrico Bambino Gesu` , IRCCS, Rome, Italy Accepted 2 April 2002
Abstract Electrical stimulation of upper limb nerves allows one to record two types of scalp responses, that is conventional low-frequency somatosensory evoked potentials (SEPs), and bursts of high-frequency (about 600 Hz) wavelets. To further clarify the functional meaning of both types of responses, we investigated whether changes of GABAergic drive could cause significant modifications of conventional as well as high-frequency SEPs. We recorded median nerve SEPs from six healthy volunteers before and after a single oral administration of lorazepam. In order to explain scalp SEP distribution before and after lorazepam administration, we performed the brain electrical source analysis of raw data. After lorazepam administration, conventional scalp SEPs showed a significant amplitude decrease of all cortical components including the primary N20 / P20 response, while the subcortical P14 response remained substantially unchanged. Similarly, dipolar analysis showed a significant strength decrease of all cortical dipoles, whereas the strength of both subcortical dipoles (possibly located at the level of the brainstem and thalamus, respectively) remained unchanged. By contrast, no significant changes of high-frequency SEPs were induced by drug intake. Therefore, our findings suggest that the inhibitory effect induced by lorazepam mainly affects intracortical circuitry. Tonic increase of the inhibitory drive, possibly mediated by GABAA receptors, can account for the reduced activity of first order deep spiny neurons generating the primary N20 / P20. Conversely, intrinsic firing properties of the cell population generating high-frequency SEP responses are unaffected by the increase of GABAergic drive. This finding lends further substance to the hypothesis that conventional and high-frequency SEPs are generated by different cell populations. 2002 Elsevier Science B.V. All rights reserved. Theme: Sensory systems Topic: Somatosensory cortex and thalamocortical relationships Keywords: Somatosensory evoked potential; Generator source; GABA; Dipolar analysis
1. Introduction Previous studies on low-frequency scalp responses to the electrical stimulation of peripheral nerves (somatosensory evoked potentials, SEPs) provided some definitive answer about their generation. It seems definitively ascertained that the frontal positive response at 20 ms of latency (P20), together with the negative parietal potential at 20 ms of latency (N20), reflect the opposite activities of an equiva*Corresponding author. Fax: 139-06-3550-1909. E-mail address: [email protected] (D. Restuccia).
lent dipole in the 3b somatosensory area [1,2,8,10]. The depolarization of pyramidal cells in deep layers, subsequent to the summation of simultaneous EPSPs, is the possible mechanism underlying the activation of this dipole [1,18]. More recently, the use of narrow bandpass filtering allowed to detect, within the same latency range of the first cortical low-frequency SEP, high-frequency (600– 700 Hz) bursts of repetitive wavelets that are more evident in the fronto–parietal region contralateral to the stimulated side [7]. Although a significant part of high-frequency responses is likely to be generated in the primary somatosensory area [6], they probably reflect the activity of a cell
0006-8993 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 02 )02818-4
2
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
population different from that generating the primary N20 response, since they are dramatically reduced during sleep, while the N20 SEP remains unaffected [29] or slightly enhanced [14]. Earlier studies hypothesized that the successive peaks of high-frequency bursts are generated by the activation of a polysynaptic cortical network; in fact, during double shock stimulation, later peaks show a slower recovery than earlier ones [11]. More recently, this different recovery for earlier and later peaks of the burst has been explained by suggesting that the whole succession of wavelets is generated by specialized cell populations, such as inhibitory interneurons or chattering cells, capable to discharge high-frequency bursts of fast sodium spikes [19]. A useful approach to study the functional characteristics of conventional as well as high-frequency SEPs is to test their respective changes after pharmacological manipulations which modifies the intracortical inhibitory drive. Benzodiazepines are known to influence both latency and amplitude of conventional SEPs by enhancing GABAergic transmission [24–26]. However, earlier studies have been performed by using a limited number of scalp leads. Moreover, the influence of benzodiazepines on high-frequency bursts has been thus far described in only one pilot study [15]. To determine whether physiologic scalp responses are influenced by inhibitory circuitry in somatosensory cortex, we recorded SEPs from six healthy volunteers before and after a single oral administration of lorazepam. Conventional as well as high-frequency SEPs have been obtained after simulation of the right median nerve. Lastly, in order to increase both temporal and spatial resolution of SEPs, we further analyzed SEP data by a spatial analysis not dependent on spatial overlapping between multiple source activities of field distributions, i.e., the brain electrical source analysis (BESA), which is useful for separating the activities of neighboring cerebral structures [4,12,22,23].
2. Materials and methods
2.1. Subjects Six healthy volunteers were included in the study (two men, four women; age 21–27 years, mean 23.5). None of the subjects had a history of neurological illness. Approval was obtained from the local ethical committee. All subjects gave their informed consent. Studies were performed according to the declaration of Helsinki. Since effective lorazepam concentrations are obtained within 30–60 min and maintained for 4–6 h [3], each subject underwent SEP recording before and 150 min after the administration of 2 mg lorazepam. Since sleep per se induces marked changes of high-frequency SEPs [14,29], subjects were instructed to stay awake.
2.2. SEP recording For SEP recording, subjects lay on a couch in a warm and semidarkened room. Each recording session included the stimulation of the right median nerve at wrist (cathode proximal, stimulus intensity just above the motor threshold, frequency 1.6 Hz, duration 0.2 ms). Disk recording electrodes (impedance below 5 KV) were placed at 19 locations of the 10–20 system (excluding Fpz and Oz). The reference electrode was at the lobe of the ear ipsilateral to the stimulated side, and the ground at Fpz. The analysis time was 62.5 ms, with a bin width of 125 ms. The amplifier bandpass was 1–3000 Hz (12 dB roll off). An automatic artifact-rejection system excluded from the average all runs containing transients exceeding 665 mV at any recording channel. In order to ensure baseline stabilization, SEPs were digitally filtered off-line by means of a digital filter with a bandpass of 20–2000 Hz. Two averages of 1500 trials each were obtained; they were superimposed to verify their reproducibility and then further averaged. Frozen maps showing the distribution of the responses over the scalp were obtained by linear interpolation from the four nearest electrodes. To selectively study high-frequency responses, a further digital filtering was performed using a narrow bandpass (500–700 Hz, 24 dB roll off).
2.3. Data analysis Wide bandpass SEPs were identified on the basis of latency, polarity and scalp distribution. Amplitudes and peak latencies were measured on the average of the two runs. Amplitudes were measured from the baseline. We evaluated the principal scalp components, labeled as in earlier reports [9,13,21,27,28], i.e., parietal N20 wave, frontal P20, central P22, frontal N24, parietal P24, and fronto–central N30 waves. To assess the distribution of responses, their amplitudes were submitted to Friedman’s test, considering scalp locations as source of variability. When statistical significance was reached, a post-hoc analysis using the Wilcoxon test was performed. To assess amplitude changes before and after lorazepam administration, response amplitudes were compared by means of Wilcoxon tests. When scalp distribution was unchanged (in all cases: see Results section), comparison was performed at the recording location where the response was maximal. High frequency bursts were evaluated over the F3, C3 and P3 locations, by measuring the latency of any reproducible wavelet; responses were labeled according to their polarity and latency. We measured: (1) the peak-topeak amplitudes of the positive frontal and negative central and parietal component whose latency was coincident to those of the wide bandpass N20 response in the same
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
subject; this response is labeled in the tables as N1 (P3 recordings) and P1 (frontal recordings); (2) the latency of reproducible wavelets following the N1 / P1; (3) the duration of the whole burst (from the N1 / P1 to the last reproducible wavelet). Latencies of any wavelet replicated in all subject in the 20–25 ms latency range, amplitudes of the wavelets occurring at about 20 ms at F3, P3 and C3 recording locations, and burst durations were compared before and after lorazepam administration by means of paired t-tests (corresponding latency values) and Wilcoxon tests (amplitude values).
2.4. Brain electric source analysis Detailed description of BESA is reported elsewhere [23]. The BESA program calculates potential distributions over the scalp from preset voltage dipoles within a threeshell model of the head. It also evaluates the fit between the recorded and calculated field distributions. The percentage of data that cannot be explained by the calculated field distribution is expressed as residual variance (RV). Of course, the lower the RV the better the dipolar model and, in an ideal case, the RV should only be due to the recorded noise. In general, RV values lower than 10% are considered acceptable, particularly when obtained from individual recordings. However, it should also be clear that even zero RV does not prove the model to be correct, on account of the infinite number of solutions to the ‘inverse problem’ of deriving intracranial sources from the extracranial potential field. BESA uses a spherical three-shell model with an 85-mm radius and assumes that the brain surface is at 70 mm from the center of the sphere. The
3
spatial position of each dipole is described on the basis of three axes: (1) the line through T3 and T4 (x-axis); (2) the line through Fpz and Oz ( y-axis); (3) the line through Cz (z-axis). The three axes have their intersection point at the center of the sphere. The spatial orientation of the dipoles is described by two angles: (1) f is the angle in the x–y plane measured counterclockwise from the nearest x-axis; (2) u is the vertical angle that is measured from the z-axis and is positive for the right hemisphere. The strength is given in ‘mVeff’, 1 mVeff being the strength of a horizontal dipole, located at y550 mm, which produces a voltage difference of 0.5 mV between C3 and C4. Strengths of any dipole, before and after lorazepam administration, were compared by means of paired t-tests.
3. Results
3.1. Wide bandpass SEPs 3.1.1. Before lorazepam In all our subjects, we could identify N20, P20, P22, P24 and N30 SEP components in parietal, central and frontal traces (see Table 1 and Fig. 1). After median nerve stimulation, the N30 response was well identifiable over central and frontal locations. It was preceded by a negative N24 frontal wave which appeared as a shoulder on the rising phase of the N30 potential in five out of six subjects (see Fig. 1). Concerning scalp distribution, parietal N20 and P24 responses had their maximal amplitude at P3 leads, while a P22 central response was maximal at C3 locations. The N24 frontal wave was well evident at all frontal locations. Its mean amplitude was higher at F3 and
Table 1 Sub.
Age
SEPs right median nerve (before lorazepam) Latencies (ms)
1 2 3 4 5 6
23 27 21 21 24 25
Amplitudes (mV)
P14
N20
P22
P24
N24
N30
P14
N20
P22
P24
N24
N30
14.77 13.55 14.4 14.4 14.4 15.26
19.17 17.58 19.04 18.55 19.04 20.26
21.73 19.87 21.24 21 21.97 22.83
25.88 24.66 30.27 24.9 25.15 25
25.02 26.25 26.12 25.15 24.17 –
28.93 30.4 29.79 29.3 30.52 30.52
0.46 0.37 0.54 1.46 0.81 0.56
1.27 0.77 1.51 3.2 2.81 2.85
0.73 1.1 1.22 2.47 0.97 0.25
1.1 1.12 1.98 3.25 1.05 1.39
1.51 0.81 0.61 2.88 1.54 –
1.42 0.59 2.37 4.54 1.03 1.54
SEPs right median nerve (after lorazepam) Latencies (ms)
1 2 3 4 5 6
23 27 21 21 24 25
Amplitudes (mV)
P14
N20
P22
P24
N24
N30
P14
N20
P22
P24
N24
N30
14.77 13.18 14.16 14.4 14.4 15.63
18.8 17.21 18.55 18.31 19.29 19.9
20.93 19.9 21 20.5 20.75 22.71
25.39 24.66 25.39 24.66 26.37 25
26 25.88 25.63 25.15 25.63 –
28.08 30 29.76 30 29.32 30
0.44 0.37 0.78 1.17 0.98 0.44
1.25 0.63 1.5 2.66 1.68 2.2
0.63 0.73 0.71 1.64 0.73 0.12
1.03 1.03 0.12 2.69 0.1 0.68
1.1 0.56 0.49 2.29 1.17 –
0.98 0.56 1.22 3.37 0.68 1
N24 amplitudes are measured at F3 locations; N30 amplitudes are measured at Fz locations
4
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
Fig. 1. Low-frequency SEPs to right median nerve stimulation in subject Nos. 2 (A), 4 (B), and 6 (C). SEPs obtained before (thin traces) and after (thick traces) drug administration are superimposed. P14 subcortical component does not show any difference before and after drug intake, while all cortical responses, including the primary N20 / P20 component, show a slight but significant amplitude reduction. SEP latencies are substantially unchanged after lorazepam administration.
Fz locations, but there were no significant differences among the frontal electrodes (Friedmann test, Z50.4, P. 0.05). The N30 mean amplitude was higher at Fz, F3 and
Cz locations after median nerve stimulation, but there were no significant differences among the frontal electrodes (Friedmann test Z57.3, P.0.05).
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
3.1.2. After lorazepam By comparing SEP amplitudes before and after lorazepam administration, we found a significant difference, caused by a clear-cut amplitude reduction after lorazepam intake, concerning all cortical components (N20, Wilcoxon test, Z52.201, P,0.05; P20, Wilcoxon test, Z52.201, P,0.05; P22, Wilcoxon test, Z52.201, P,0.05; N24, Wilcoxon test, Z52.023, P,0.05; P24, Wilcoxon test, Z52.201, P,0.05; N30 Wilcoxon test, Z52.201, P,0.05; see Table 1 and Fig. 1). Concerning scalp SEP distribution, it remained substantially unchanged after lorazepam administration: the mean amplitude of both N24 and N30 was higher at F3 and Fz locations, but there were no significant differences among the frontal electrodes (Friedmann test, P.0.05). 3.2. High frequency SEPs 3.2.1. Before lorazepam After the N1 and P1 components, F3, P3 and C3 traces showed a number (ranging from 4 to 8) of high frequency (about 600 Hz) wavelets. Table 2 shows the mean values of the three parameters we considered. 3.2.2. After lorazepam Latencies of all wavelets remained substantially unchanged after lorazepam administration (paired t-tests, P. 0.05). Moreover, no significant changes were observed for
5
both amplitude (Wilcoxon tests, P.0.05) and duration (paired t-test, P.0.05) values (see Table 2 and Fig. 2).
3.3. Dipolar analysis: wide bandpass SEPs To build dipolar models of wide-range SEPs in our subjects, we employed a ‘sequential strategy’, as described in detail elsewhere [27,28]. We divided the analysis time (from the subcortical P14 to the N30 response) into two intervals, choosing the peak of the N20 response as the division point. In the earlier interval, which was analyzed first, two subcortical and two cortical dipolar sources were activated. Two subcortical sources were believed necessary according to other studies [5], showing that subcortical activity can be explained by two different sources possibly located in the brainstem and in the thalamus, respectively. Moreover, previous results suggested that two different cortical generators contribute to the SEP topography in the 20 ms latency range [27]. When we added to the analysis the interval after the N20 peak latency, another dipole was needed to explain the scalp SEP topography. This fivedipole model explained well the SEP distribution in individual traces (RV values ranging from 1 to 6.22%; see Table 3 and Figs. 3, 4 and 5). The first dipole (No. 1), whose peaking activity had the same latency as the P14, was placed at the base of the skull, while the second dipole (No. 2) was placed near the center of the head peaking at 16 ms. The other three
Table 2 High frequency bursts: values before lorazepam administration P3 N1 Latency (ms) mean 18.82 SD 0.66
F3
C3
N2
N3
N4
P1
P2
P3
P4
N1
N2
N3
20.62 1.02
22.22 1.37
22.92 0.89
18.94 0.84
20.42 0.72
22.42 0.98
23.38 0.7
19.26 0.52
20.96 0.65
22.64 0.71
Amplitude (mV) mean 0.039 SD 0.011
0.046 0.017
Burst duration (ms) mean SD
0.035 0.014
7.04 1.45
6.86 1.28
6.56 0.84
High frequency bursts: values after lorazepam administration P3 N1 Latency (ms) mean 18.8 SD 0.64
F3 N2
N3
N4
P1
P2
P3
P4
N1
N2
N3
20.58 1.15
22.34 1.2
23.28 1.09
18.92 0.75
20.3 0.83
22.2 0.83
23.68 1.02
19.26 0.52
20.84 0.55
22.64 0.66
Amplitude (mV) mean 0.042 SD 0.016 Burst duration (ms) mean SD
C3
0.044 0.019 6.42 1.15
0.037 0.014 6.42 1.15
6.62 0.73
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
6
potentials. Dipole No. 4 showed an inconstant early peak of activity slightly preceding the first peak of dipole No. 3; later, it was activated with opposite polarity at the same latency as the P22 response. The fifth dipole (No. 5) reached a radial orientation and a medial location and showed a late peak of activity at the latency of the fronto–central N30. After lorazepam administration (see Figs. 3, 4, and 5), the strengths of subcortical dipoles remained substantially unchanged (paired t-tests, P.0.05), while the strengths of the remaining three perirolandic dipoles were significantly decreased (paired t-tests, P,0.05). RV values remained substantially unchanged (individual values ranging from 1.35 to 7.89%).
3.4. Dipolar analysis: high frequency SEPs
Fig. 2. SEPs to right median nerve stimulation in subject No. 1. SEPs obtained before (thin traces) and after (thick traces) drug administration are superimposed. Traces from P3, C3 and Fz locations have been digitally filtered using narrow bandpass (500–700 Hz; lower rows). The highest row shows the P3 recording with a wide (20–2000 Hz) bandpass filtering. The vertical line indicates the latency of the N20 / P20 component, possibly coincident with the arrival of the afferent volley to the somatosensory cortex. High frequency bursts are clearly unaffected by the drug intake.
dipoles had perirolandic locations. Dipole No. 3 was oriented tangentially and was activated at the latencies of both the N20 / P20 and, with inverted polarity, P24 / N24
We further built dipolar models of traces of high frequency SEPs after median nerve stimulation, using the same sequential strategy. SEPs were well explained by a four-dipole model, since no significant activation was found in the 30 ms range of latency. The locations of these four dipoles were similar to those demonstrated for the first four dipoles in low-frequency SEPs. This four-dipole model explained well the SEP distribution (RV values ranging from 6.56 to 12.6%; see Table 3 and Fig. 6); RV values remained substantially unchanged after lorazepam administration (individual values ranging from 6.99 to 13.8%). Dipole strengths of high-frequency models remained unchanged after lorazepam administration (paired t-tests, P.0.05; see Figs. 6 and 7).
Table 3 Coordinates of dipoles: Low frequency median nerve SEP model Dipole 1 x Mean SD
0.21 2.86
Dipole 2 z
theta
phi
x
y
z
theta
phi
x
y
z
theta
phi
210.78 16.97
254.83 4.40
2165.00 13.80
227.63 66.85
214.54 10.84
213.40 6.30
211.80 15.16
282.13 76.85
14.35 52.35
237.33 7.15
8.14 4.33
43.60 6.62
241.33 9.50
212.75 6.75
x
y
z
theta
phi
233.33 7.53
223.94 65.01
232.67 9.67
0.43 14.05
Dipole 4 x Mean SD
Dipole 3
y
Dipole 5 y
238.83 4.83
z 2.76 4.70
theta 42.30 8.75
243.17 7.47
phi 23.57 7.44
53.03 6.76
Coordinates of dipoles: High frequency median nerve SEP model Dipole 1 x 20.67 11.66
Mean SD
Dipole 2 y
z
theta
219.60 18.12
242.60 10.29
y
z
225.00 169.13
phi 50.98 59.44
Dipole 4 x Mean SD
239.80 6.02
4.06 1.39
theta 37.02 6.88
247.40 9.48
phi 25.76 1.73
Dipole 3
x
y
213.46 7.24
210.44 8.22
z 11.02 7.83
theta
phi
x
y
z
theta
phi
260.20 22.75
36.20 30.67
237.20 4.09
8.04 7.35
43.96 2.78
241.40 4.22
212.47 11.43
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
7
Fig. 3. Five-dipole spatiotemporal solution for wide bandpass (20–2000 Hz) median nerve SEPs before and after drug intake; patient No. 4. The residual variance is 1 and 1.9 before and after drug intake, respectively. The source potentials of the dipoles are shown on the left (thick traces: before drug intake. Thin traces: after drug intake); analysis time starts from 8 ms. On the right, three views of the head illustrate the location and orientation of the dipoles. The two upper rows show source potential and location of the subcortical dipoles (dipoles 1 and 2). Source potential and location of the tangential perirolandic dipole (dipole 3) are shown in the third row. The fourth and fifth rows show source potentials and locations of the other two perirolandic dipoles. Note that all perirolandic dipoles (Nos. 3, 4, and 5) are notably weaker after lorazepam administration.
4. Discussion The main finding in this paper is represented by the dissociate behavior of conventional and high-frequency SEPs after single oral dose of lorazepam. While conventional SEPs showed a slight but significant amplitude reduction, high-frequency bursts did not showed any measurable modification. The notion that conventional SEPs are modified by benzodiazepine intake is not entirely new. In past years, in order to verify the reliability of SEP monitoring during surgery, several studies tested the action of anesthetic drugs including benzodiazepines on SEP recordings. Although these earlier studies were performed by using the same simplified recording technique usually utilized in operating room, they were able to demonstrate that benzodiazepines mainly affect the amplitude of cortical SEP components, while latency values showed minor changes or were not affected at all [24,25]. Our present findings suggest that lorazepam administration induces slight but significant amplitude reduction not only of the primary N20 / P20, but also of all following cortical
components. The N20 / P20 response probably reflects the initial excitation of deep layers of the 3b somatosensory area [1,2,8,10,18]. Lorazepam, such as other benzodiazepines, probably exerts its action by enhancing the activity of the BZ1 subtype of GABAA receptors [20]. Tonic and sustained enhancement of such an inhibitory drive could therefore reduce the sharp and synchronous summation of EPSPs that generates the N20 / P20 response. Moreover, this tonic decrease of spiny cell excitability can also explain the reduction of physiologic excitability changes which possibly generate later SEP components [1,2]. Data obtained by dipolar analysis of our traces confirm that lorazepam influence the SEP recording at a cortical level, since both subcortical dipoles, possibly located at the level of brainstem and thalamus, respectively, did not show any significant change. By contrast, high-frequency bursts seem unaffected by lorazepam administration. Besides the visual analysis of raw data, also dipolar analysis clearly showed that cortical generators of high frequency bursts are unaffected by lorazepam administration. In fact, high frequency SEPs in
8
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
Fig. 4. Five-dipole spatiotemporal solution for wide bandpass (20–2000 Hz) median nerve SEPs before and after drug intake; patient No. 2. The residual variance is 1.98 and 2.3 before and after drug intake, respectively. Same presentation as in Fig. 3. All perirolandic dipoles (Nos. 3, 4, and 5) are weaker after lorazepam administration.
Fig. 5. The histogram shows the average strength of each dipole in low-frequency median nerve SEP modeling from all six volunteers, before (black columns) and after (ragged columns) lorazepam administration. Bars above each column represent standard deviations. Strengths of all cortical dipoles were significantly reduced after drug administration (asterisk: paired t-test, P,0.05).
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
9
Fig. 6. Four-dipole spatiotemporal solution for high frequency (500–700 Hz) median nerve SEPs before and after drug intake; patient No. 1. The residual variance is 7.3 and 8.2 before and after drug intake, respectively. The source potentials of the dipoles are shown on the left (black traces: before drug intake. Red traces: after drug intake); analysis time starts from 8 ms. On the right, three views of the head illustrate the location and orientation of the dipoles. The two upper rows show source potential and location of the subcortical dipoles (dipoles 1 and 2). The third and fourth rows show source potentials and locations of the perirolandic dipoles (dipoles 3 and 4). The time courses of all dipoles are unaffected by the drug intake.
Fig. 7. The histogram shows the average strength of each dipole in high-frequency median nerve SEP modeling from all six volunteers, before (black columns) and after (ragged columns) lorazepam administration. Bars above each column represent standard deviations. Strengths of all dipoles remained substantially unchanged after drug administration.
10
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
the 20–25 ms range of latency are explained by two perirolandic sources, which remained substantially unmodified after drug intake. Although this part of highfrequency burst is recorded at the same sites and in the same latency range of the N20 / P20 response, previous studies provided several evidences that it is generated by different cell populations. High-frequency bursts disappear during sleep while the N20 increases or remains unchanged [14,29]; moreover, 600 Hz bursts exhibit a significantly higher short-term variability than low-frequency responses, so that it has been suggested that they might correlate to more variable aspects of the somatosensory processing, such as floating attentional shifts [17]. Many classes of specialized cortical cells, such as pyramidal chattering cells and fast-spiking inhibitory interneurons, are capable to discharge high-frequency bursts [7]. Our present data cannot obviously specify what class of cell account for high-frequency SEP generation, but they clearly indicate that 600 Hz burst are insensitive to the increase of intracortical GABAergic drive. This finding is in agreement with another report suggesting that the enhancement of GABAergic transmission subsequent to valproate administration has no effect on high frequency SEPs [17]. On the contrary, slight but significant modifications of high frequency SEPs after lorazepam administration, mainly concerning their latency, have been found in another study [15]. This apparent discrepancy with our data can be explained by considering the higher dose of lorazepam utilized by Haueisen et al. in the above-mentioned report [15]. It is well known that latency shifts of postsynaptic cortical responses are mainly caused by conduction slowing preceding the arrival of the afferent volley to the cortex [24–26]; it is thus conceivable that higher doses of lorazepam might act also at subcortical levels, thus causing slight latency delays of genuine cortical responses. Previous literature provided evidence that both cortical and subcortical generators contribute to scalp-evoked high-frequency SEPs [6,7]. Also in our study, part of scalp-recorded SEPs was explained by subcortical sources. Lastly, a recent magnetoelectroencephalographic study [16] suggests that burst whose latency is coincident to the ascending slope of the primary N20 probably reflects presynaptic activity. Therefore, it is conceivable that subcortical generators of bursts might be affected by administration of lorazepam doses higher than that we used in the present study, thus entailing slight latency shifts of the scalp-recorded response. It remains to be explained why lorazepam administration does not induce the same changes of high-frequency SEPs as those observed during sleep. It should be considered that the finding of decreased high frequency bursts during lowered vigilance may not reflect uniquely a decreased capacity of single cell bursting, but it could be due to increased burst timing jitter leading to partial cancellation with averaging [7]. This hypothesis agrees with the increased latency of high frequency SEPs by high doses of
lorazepam [15] and explains the lack of any high frequency burst modification in our subjects who received lower doses of this drug. Moreover, sleep involves also mechanisms different from the increase of the GABAergic tone, which were not probably activated by lorazepam administration in our subjects, explicitly asked to stay awake. In conclusion, our study confirms that high frequency SEPs occurring in the 20–25 ms latency range and conventional low-frequency cortical SEPs are probably generated by different cell populations. Moreover, it strongly indicates that specialized cell populations generating cortical high frequency SEPs are substantially insensitive to GABAergic inhibition.
References [1] T. Allison, G. Mc Carthy, C.C. Wood, T.M. Darcey, D.D. Spencer, P.D. Williamson, Human cortical potentials evoked by stimulation of the median nerve. I. Cytoarchitectonic areas generating short-latency activity, J. Neurophysiol. 62 (1989) 694–710. [2] T. Allison, G. Mc Carthy, C.C. Wood, S.J. Jones, Potentials evoked in human and monkey cerebral cortex by stimulation of the median nerve, Brain 114 (1991) 2465–2503. [3] E.G. Bradshaw, A.A. Ali, B.A. Mulley, R.M. Rye, Plasma concentrations and clinical effects of lorazepam after oral administration, Br. J. Anaesth. 53 (5) (1981) 517–522. ¨ [4] H. Buchner, L. Adams, A. Muller, I. Ludwig, A. Knepper, A. Thron, K. Niemann, M. Scherg, Somatotopy of human hand somatosensory cortex revealed by dipole source analysis of early somatosensory evoked potentials and 3D-NMR tomography, Electroencephalogr. Clin. Neurophysiol. 96 (1995) 121–134. [5] H. Buchner, T.D. Waberski, M. Fuchs, H.A. Wishmann, R. Beck¨ mann, A. Rienacker, Origin of P16 median nerve SEP component identified by dipole source analysis. Subthalamic or within the thalamo–cortical radiations?, Exp. Brain Res. 104 (1995) 511–518. [6] G. Curio, B.M. Mackert, M. Burghoff, R. Koetitz, K. Abraham¨ Fuchs, W. Harer, Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system, Electroencephalogr. Clin. Neurophysiol. 91 (1994) 483–487. [7] G. Curio, Linking 600 Hz ‘spikelike’ EEG / MEG wavelets (‘sbursts’) to cellular substrates. Concepts and caveats, J. Clin. Neurophysiol. 17 (4) (2000) 377–396. [8] J.E. Desmedt, G. Cheron, Central somatosensory conduction in man: neuronal generators and interpeak latencies of the far-field components recorded from neck and right or left scalp and earlobes, Electroencephalogr. Clin. Neurophysiol. 50 (1980) 382–403. [9] J.E. Desmedt, T.H. Nguyen, M. Bourguet, Bit-mapped color imaging of human evoked potentials with reference to the N20, P22, P27 and N30 somatosensory responses, Electroencephalogr. Clin. Neurophysiol. 68 (1987) 1–19. [10] J.E. Desmedt, I. Ozaki, SEPs to finger joint input lack the N20–P20 response that is evoked by tactile inputs: contrast between cortical generators in areas 3b and 2 in humans, Electroencephalogr. Clin. Neurophysiol. 80 (1991) 513–521. [11] T. Emori, T. Yamada, Y. Seki, A. Yasuhara, K. Ando, Y. Honda, A.A. Leis, P. Vachatimanont, Recovery functions of fast frequency potentials in the initial negative wave of median SEP, Electroencephalogr. Clin. Neurophysiol. 78 (1991) 116–123. [12] H. Franssen, D.F. Stegemann, J. Moleman, R.P. Schoobar, Dipole modelling of median nerve SEPs in normal subjects and patients
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20] [21]
with small subcortical infarcts, Electroencephalogr. Clin. Neurophysiol. 84 (1992) 40–47. ` Unmasking of cortical L. Garcia-Larrea, H. Bastuji, F. Mauguiere, SEP components by changes in stimulus rates: a topographic study, Electroencephalogr. Clin. Neurophysiol. 84 (1992) 71–83. I. Hashimoto, T. Mashiko, T. Imada, Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex, Electroencephalogr. Clin. Neurophysiol. 100 (1996) 189–203. J. Haueisen, T. Heuer, H. Nowak, J. Liepert, C. Weiller, Y. Okada, G. Curio, The influence of lorazepam on somatosensory-evoked fast frequency (600 Hz) activity in MEG, Brain Res. 874 (2000) 10–14. J. Haueisen, B. Schack, T. Meier, G. Curio, Y. Okada, Multiplicity in the high-frequency signals during the short-latency somatosensory evoked cortical activity in humans, Clin. Neurophysiol. 112 (2001) 1316–1325. F. Klostermann, G. Nolte, G. Curio, Independent short-term variability of spike-like (600 Hz) and postsynaptic (N20) cerebral SEP components, Neuroreport 12 (2001) 349–352. ¨ H. Luders, R.P. Lesser, J. Hahn, D.S. Dinner, G. Klem, Cortical somatosensory evoked potentials in response to hand stimulation, J. Neurosurg. 58 (1983) 885–894. B.M. Mackert, S. Weisenbach, G. Nolte, G. Curio, Rapid recovery (20 ms) of human 600 Hz electroencephalographic wavelets after double stimulation of sensory nerves, Neurosci. Lett. 286 (2000) 83–86. R.M. McKernan, P.J. Whiting, Which GABA-receptor subtypes really occur in the brain, Trends Neurosci. 19 (1996) 139–143. D. Restuccia, M. Valeriani, C. Barba, D. Le Pera, P. Tonali, F.
[22] [23]
[24]
[25] [26]
[27]
[28]
[29]
11
` Mauguiere, Different contribution of joint and cutaneous inputs to early scalp somatosensory evoked potentials, Muscle Nerve 22 (1999) 910–919. M. Scherg, J. Vajar, T.W. Picton, A source analysis of the human auditory potentials, J. Cogn. Neurosci. 1 (1989) 336–354. M. Scherg, Fundamentals of dipole source potential analysis, in: F. Grandoni, M. Hoke, G.L. Romani (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials, Advances in Audiology, Vol. 6, Karger, Basel, 1990, pp. 40–69. T.B. Sloan, M.L. Fugina, J.R. Toleikis, Effects of midazolam on median nerve somatosensory evoked potentials, Br. J. Anaesth. 64 (5) (1990) 590–593. T.B. Sloan, Anesthetic effects on electrophysiologic recordings, J. Clin. Neurophysiol. 15 (1998) 217–226. A. Todorova, Effects of diazepam and the specific benzodiazepine antagonist flumanezil on somatosensory evoked potentials in rat, Arch. Int. Pharmacodyn. Ther. 321 (1993) 14–29. M. Valeriani, D. Restuccia, V. Di Lazzaro, D. Le Pera, P. Tonali, The pathophysiology of giant SEPs in cortical myoclonus: a scalp topography and dipolar source modeling study, Electroencephalogr. Clin. Neurophysiol. 104 (1997) 122–131. M. Valeriani, D. Restuccia, V. Di Lazzaro, D. Le Pera, C. Barba, P. ` Tonali, F. Mauguiere, Dipolar sources of the early scalp somatosensory evoked potentials to upper limb stimulation. Effect of increasing stimulus rates, Exp. Brain Res. 120 (1998) 306–315. T. Yamada, S. Kameyama, Y. Fuchigami, Y. Nakazumi, Q.S. Dickins, J. Kimura, Changes of short latency somatosensory evoked potentials in sleep, Electroencephalogr. Clin. Neurophysiol. 70 (1988) 126–136.
Research report
Dissociated changes of somatosensory evoked low-frequency scalp responses and 600 Hz bursts after single-dose administration of lorazepam Domenico Restuccia a , *, Massimiliano Valeriani a,b , Eugenio Grassi a , Salvatore Mazza a , Pietro Tonali a a
Department of Neurology, Catholic University, Policlinico A. Gemelli, L.go A. Gemelli 8, 00168 Rome, Italy b Division of Neurology, Ospedale Pediatrico Bambino Gesu` , IRCCS, Rome, Italy Accepted 2 April 2002
Abstract Electrical stimulation of upper limb nerves allows one to record two types of scalp responses, that is conventional low-frequency somatosensory evoked potentials (SEPs), and bursts of high-frequency (about 600 Hz) wavelets. To further clarify the functional meaning of both types of responses, we investigated whether changes of GABAergic drive could cause significant modifications of conventional as well as high-frequency SEPs. We recorded median nerve SEPs from six healthy volunteers before and after a single oral administration of lorazepam. In order to explain scalp SEP distribution before and after lorazepam administration, we performed the brain electrical source analysis of raw data. After lorazepam administration, conventional scalp SEPs showed a significant amplitude decrease of all cortical components including the primary N20 / P20 response, while the subcortical P14 response remained substantially unchanged. Similarly, dipolar analysis showed a significant strength decrease of all cortical dipoles, whereas the strength of both subcortical dipoles (possibly located at the level of the brainstem and thalamus, respectively) remained unchanged. By contrast, no significant changes of high-frequency SEPs were induced by drug intake. Therefore, our findings suggest that the inhibitory effect induced by lorazepam mainly affects intracortical circuitry. Tonic increase of the inhibitory drive, possibly mediated by GABAA receptors, can account for the reduced activity of first order deep spiny neurons generating the primary N20 / P20. Conversely, intrinsic firing properties of the cell population generating high-frequency SEP responses are unaffected by the increase of GABAergic drive. This finding lends further substance to the hypothesis that conventional and high-frequency SEPs are generated by different cell populations. 2002 Elsevier Science B.V. All rights reserved. Theme: Sensory systems Topic: Somatosensory cortex and thalamocortical relationships Keywords: Somatosensory evoked potential; Generator source; GABA; Dipolar analysis
1. Introduction Previous studies on low-frequency scalp responses to the electrical stimulation of peripheral nerves (somatosensory evoked potentials, SEPs) provided some definitive answer about their generation. It seems definitively ascertained that the frontal positive response at 20 ms of latency (P20), together with the negative parietal potential at 20 ms of latency (N20), reflect the opposite activities of an equiva*Corresponding author. Fax: 139-06-3550-1909. E-mail address: [email protected] (D. Restuccia).
lent dipole in the 3b somatosensory area [1,2,8,10]. The depolarization of pyramidal cells in deep layers, subsequent to the summation of simultaneous EPSPs, is the possible mechanism underlying the activation of this dipole [1,18]. More recently, the use of narrow bandpass filtering allowed to detect, within the same latency range of the first cortical low-frequency SEP, high-frequency (600– 700 Hz) bursts of repetitive wavelets that are more evident in the fronto–parietal region contralateral to the stimulated side [7]. Although a significant part of high-frequency responses is likely to be generated in the primary somatosensory area [6], they probably reflect the activity of a cell
0006-8993 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 02 )02818-4
2
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
population different from that generating the primary N20 response, since they are dramatically reduced during sleep, while the N20 SEP remains unaffected [29] or slightly enhanced [14]. Earlier studies hypothesized that the successive peaks of high-frequency bursts are generated by the activation of a polysynaptic cortical network; in fact, during double shock stimulation, later peaks show a slower recovery than earlier ones [11]. More recently, this different recovery for earlier and later peaks of the burst has been explained by suggesting that the whole succession of wavelets is generated by specialized cell populations, such as inhibitory interneurons or chattering cells, capable to discharge high-frequency bursts of fast sodium spikes [19]. A useful approach to study the functional characteristics of conventional as well as high-frequency SEPs is to test their respective changes after pharmacological manipulations which modifies the intracortical inhibitory drive. Benzodiazepines are known to influence both latency and amplitude of conventional SEPs by enhancing GABAergic transmission [24–26]. However, earlier studies have been performed by using a limited number of scalp leads. Moreover, the influence of benzodiazepines on high-frequency bursts has been thus far described in only one pilot study [15]. To determine whether physiologic scalp responses are influenced by inhibitory circuitry in somatosensory cortex, we recorded SEPs from six healthy volunteers before and after a single oral administration of lorazepam. Conventional as well as high-frequency SEPs have been obtained after simulation of the right median nerve. Lastly, in order to increase both temporal and spatial resolution of SEPs, we further analyzed SEP data by a spatial analysis not dependent on spatial overlapping between multiple source activities of field distributions, i.e., the brain electrical source analysis (BESA), which is useful for separating the activities of neighboring cerebral structures [4,12,22,23].
2. Materials and methods
2.1. Subjects Six healthy volunteers were included in the study (two men, four women; age 21–27 years, mean 23.5). None of the subjects had a history of neurological illness. Approval was obtained from the local ethical committee. All subjects gave their informed consent. Studies were performed according to the declaration of Helsinki. Since effective lorazepam concentrations are obtained within 30–60 min and maintained for 4–6 h [3], each subject underwent SEP recording before and 150 min after the administration of 2 mg lorazepam. Since sleep per se induces marked changes of high-frequency SEPs [14,29], subjects were instructed to stay awake.
2.2. SEP recording For SEP recording, subjects lay on a couch in a warm and semidarkened room. Each recording session included the stimulation of the right median nerve at wrist (cathode proximal, stimulus intensity just above the motor threshold, frequency 1.6 Hz, duration 0.2 ms). Disk recording electrodes (impedance below 5 KV) were placed at 19 locations of the 10–20 system (excluding Fpz and Oz). The reference electrode was at the lobe of the ear ipsilateral to the stimulated side, and the ground at Fpz. The analysis time was 62.5 ms, with a bin width of 125 ms. The amplifier bandpass was 1–3000 Hz (12 dB roll off). An automatic artifact-rejection system excluded from the average all runs containing transients exceeding 665 mV at any recording channel. In order to ensure baseline stabilization, SEPs were digitally filtered off-line by means of a digital filter with a bandpass of 20–2000 Hz. Two averages of 1500 trials each were obtained; they were superimposed to verify their reproducibility and then further averaged. Frozen maps showing the distribution of the responses over the scalp were obtained by linear interpolation from the four nearest electrodes. To selectively study high-frequency responses, a further digital filtering was performed using a narrow bandpass (500–700 Hz, 24 dB roll off).
2.3. Data analysis Wide bandpass SEPs were identified on the basis of latency, polarity and scalp distribution. Amplitudes and peak latencies were measured on the average of the two runs. Amplitudes were measured from the baseline. We evaluated the principal scalp components, labeled as in earlier reports [9,13,21,27,28], i.e., parietal N20 wave, frontal P20, central P22, frontal N24, parietal P24, and fronto–central N30 waves. To assess the distribution of responses, their amplitudes were submitted to Friedman’s test, considering scalp locations as source of variability. When statistical significance was reached, a post-hoc analysis using the Wilcoxon test was performed. To assess amplitude changes before and after lorazepam administration, response amplitudes were compared by means of Wilcoxon tests. When scalp distribution was unchanged (in all cases: see Results section), comparison was performed at the recording location where the response was maximal. High frequency bursts were evaluated over the F3, C3 and P3 locations, by measuring the latency of any reproducible wavelet; responses were labeled according to their polarity and latency. We measured: (1) the peak-topeak amplitudes of the positive frontal and negative central and parietal component whose latency was coincident to those of the wide bandpass N20 response in the same
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
subject; this response is labeled in the tables as N1 (P3 recordings) and P1 (frontal recordings); (2) the latency of reproducible wavelets following the N1 / P1; (3) the duration of the whole burst (from the N1 / P1 to the last reproducible wavelet). Latencies of any wavelet replicated in all subject in the 20–25 ms latency range, amplitudes of the wavelets occurring at about 20 ms at F3, P3 and C3 recording locations, and burst durations were compared before and after lorazepam administration by means of paired t-tests (corresponding latency values) and Wilcoxon tests (amplitude values).
2.4. Brain electric source analysis Detailed description of BESA is reported elsewhere [23]. The BESA program calculates potential distributions over the scalp from preset voltage dipoles within a threeshell model of the head. It also evaluates the fit between the recorded and calculated field distributions. The percentage of data that cannot be explained by the calculated field distribution is expressed as residual variance (RV). Of course, the lower the RV the better the dipolar model and, in an ideal case, the RV should only be due to the recorded noise. In general, RV values lower than 10% are considered acceptable, particularly when obtained from individual recordings. However, it should also be clear that even zero RV does not prove the model to be correct, on account of the infinite number of solutions to the ‘inverse problem’ of deriving intracranial sources from the extracranial potential field. BESA uses a spherical three-shell model with an 85-mm radius and assumes that the brain surface is at 70 mm from the center of the sphere. The
3
spatial position of each dipole is described on the basis of three axes: (1) the line through T3 and T4 (x-axis); (2) the line through Fpz and Oz ( y-axis); (3) the line through Cz (z-axis). The three axes have their intersection point at the center of the sphere. The spatial orientation of the dipoles is described by two angles: (1) f is the angle in the x–y plane measured counterclockwise from the nearest x-axis; (2) u is the vertical angle that is measured from the z-axis and is positive for the right hemisphere. The strength is given in ‘mVeff’, 1 mVeff being the strength of a horizontal dipole, located at y550 mm, which produces a voltage difference of 0.5 mV between C3 and C4. Strengths of any dipole, before and after lorazepam administration, were compared by means of paired t-tests.
3. Results
3.1. Wide bandpass SEPs 3.1.1. Before lorazepam In all our subjects, we could identify N20, P20, P22, P24 and N30 SEP components in parietal, central and frontal traces (see Table 1 and Fig. 1). After median nerve stimulation, the N30 response was well identifiable over central and frontal locations. It was preceded by a negative N24 frontal wave which appeared as a shoulder on the rising phase of the N30 potential in five out of six subjects (see Fig. 1). Concerning scalp distribution, parietal N20 and P24 responses had their maximal amplitude at P3 leads, while a P22 central response was maximal at C3 locations. The N24 frontal wave was well evident at all frontal locations. Its mean amplitude was higher at F3 and
Table 1 Sub.
Age
SEPs right median nerve (before lorazepam) Latencies (ms)
1 2 3 4 5 6
23 27 21 21 24 25
Amplitudes (mV)
P14
N20
P22
P24
N24
N30
P14
N20
P22
P24
N24
N30
14.77 13.55 14.4 14.4 14.4 15.26
19.17 17.58 19.04 18.55 19.04 20.26
21.73 19.87 21.24 21 21.97 22.83
25.88 24.66 30.27 24.9 25.15 25
25.02 26.25 26.12 25.15 24.17 –
28.93 30.4 29.79 29.3 30.52 30.52
0.46 0.37 0.54 1.46 0.81 0.56
1.27 0.77 1.51 3.2 2.81 2.85
0.73 1.1 1.22 2.47 0.97 0.25
1.1 1.12 1.98 3.25 1.05 1.39
1.51 0.81 0.61 2.88 1.54 –
1.42 0.59 2.37 4.54 1.03 1.54
SEPs right median nerve (after lorazepam) Latencies (ms)
1 2 3 4 5 6
23 27 21 21 24 25
Amplitudes (mV)
P14
N20
P22
P24
N24
N30
P14
N20
P22
P24
N24
N30
14.77 13.18 14.16 14.4 14.4 15.63
18.8 17.21 18.55 18.31 19.29 19.9
20.93 19.9 21 20.5 20.75 22.71
25.39 24.66 25.39 24.66 26.37 25
26 25.88 25.63 25.15 25.63 –
28.08 30 29.76 30 29.32 30
0.44 0.37 0.78 1.17 0.98 0.44
1.25 0.63 1.5 2.66 1.68 2.2
0.63 0.73 0.71 1.64 0.73 0.12
1.03 1.03 0.12 2.69 0.1 0.68
1.1 0.56 0.49 2.29 1.17 –
0.98 0.56 1.22 3.37 0.68 1
N24 amplitudes are measured at F3 locations; N30 amplitudes are measured at Fz locations
4
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
Fig. 1. Low-frequency SEPs to right median nerve stimulation in subject Nos. 2 (A), 4 (B), and 6 (C). SEPs obtained before (thin traces) and after (thick traces) drug administration are superimposed. P14 subcortical component does not show any difference before and after drug intake, while all cortical responses, including the primary N20 / P20 component, show a slight but significant amplitude reduction. SEP latencies are substantially unchanged after lorazepam administration.
Fz locations, but there were no significant differences among the frontal electrodes (Friedmann test, Z50.4, P. 0.05). The N30 mean amplitude was higher at Fz, F3 and
Cz locations after median nerve stimulation, but there were no significant differences among the frontal electrodes (Friedmann test Z57.3, P.0.05).
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
3.1.2. After lorazepam By comparing SEP amplitudes before and after lorazepam administration, we found a significant difference, caused by a clear-cut amplitude reduction after lorazepam intake, concerning all cortical components (N20, Wilcoxon test, Z52.201, P,0.05; P20, Wilcoxon test, Z52.201, P,0.05; P22, Wilcoxon test, Z52.201, P,0.05; N24, Wilcoxon test, Z52.023, P,0.05; P24, Wilcoxon test, Z52.201, P,0.05; N30 Wilcoxon test, Z52.201, P,0.05; see Table 1 and Fig. 1). Concerning scalp SEP distribution, it remained substantially unchanged after lorazepam administration: the mean amplitude of both N24 and N30 was higher at F3 and Fz locations, but there were no significant differences among the frontal electrodes (Friedmann test, P.0.05). 3.2. High frequency SEPs 3.2.1. Before lorazepam After the N1 and P1 components, F3, P3 and C3 traces showed a number (ranging from 4 to 8) of high frequency (about 600 Hz) wavelets. Table 2 shows the mean values of the three parameters we considered. 3.2.2. After lorazepam Latencies of all wavelets remained substantially unchanged after lorazepam administration (paired t-tests, P. 0.05). Moreover, no significant changes were observed for
5
both amplitude (Wilcoxon tests, P.0.05) and duration (paired t-test, P.0.05) values (see Table 2 and Fig. 2).
3.3. Dipolar analysis: wide bandpass SEPs To build dipolar models of wide-range SEPs in our subjects, we employed a ‘sequential strategy’, as described in detail elsewhere [27,28]. We divided the analysis time (from the subcortical P14 to the N30 response) into two intervals, choosing the peak of the N20 response as the division point. In the earlier interval, which was analyzed first, two subcortical and two cortical dipolar sources were activated. Two subcortical sources were believed necessary according to other studies [5], showing that subcortical activity can be explained by two different sources possibly located in the brainstem and in the thalamus, respectively. Moreover, previous results suggested that two different cortical generators contribute to the SEP topography in the 20 ms latency range [27]. When we added to the analysis the interval after the N20 peak latency, another dipole was needed to explain the scalp SEP topography. This fivedipole model explained well the SEP distribution in individual traces (RV values ranging from 1 to 6.22%; see Table 3 and Figs. 3, 4 and 5). The first dipole (No. 1), whose peaking activity had the same latency as the P14, was placed at the base of the skull, while the second dipole (No. 2) was placed near the center of the head peaking at 16 ms. The other three
Table 2 High frequency bursts: values before lorazepam administration P3 N1 Latency (ms) mean 18.82 SD 0.66
F3
C3
N2
N3
N4
P1
P2
P3
P4
N1
N2
N3
20.62 1.02
22.22 1.37
22.92 0.89
18.94 0.84
20.42 0.72
22.42 0.98
23.38 0.7
19.26 0.52
20.96 0.65
22.64 0.71
Amplitude (mV) mean 0.039 SD 0.011
0.046 0.017
Burst duration (ms) mean SD
0.035 0.014
7.04 1.45
6.86 1.28
6.56 0.84
High frequency bursts: values after lorazepam administration P3 N1 Latency (ms) mean 18.8 SD 0.64
F3 N2
N3
N4
P1
P2
P3
P4
N1
N2
N3
20.58 1.15
22.34 1.2
23.28 1.09
18.92 0.75
20.3 0.83
22.2 0.83
23.68 1.02
19.26 0.52
20.84 0.55
22.64 0.66
Amplitude (mV) mean 0.042 SD 0.016 Burst duration (ms) mean SD
C3
0.044 0.019 6.42 1.15
0.037 0.014 6.42 1.15
6.62 0.73
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
6
potentials. Dipole No. 4 showed an inconstant early peak of activity slightly preceding the first peak of dipole No. 3; later, it was activated with opposite polarity at the same latency as the P22 response. The fifth dipole (No. 5) reached a radial orientation and a medial location and showed a late peak of activity at the latency of the fronto–central N30. After lorazepam administration (see Figs. 3, 4, and 5), the strengths of subcortical dipoles remained substantially unchanged (paired t-tests, P.0.05), while the strengths of the remaining three perirolandic dipoles were significantly decreased (paired t-tests, P,0.05). RV values remained substantially unchanged (individual values ranging from 1.35 to 7.89%).
3.4. Dipolar analysis: high frequency SEPs
Fig. 2. SEPs to right median nerve stimulation in subject No. 1. SEPs obtained before (thin traces) and after (thick traces) drug administration are superimposed. Traces from P3, C3 and Fz locations have been digitally filtered using narrow bandpass (500–700 Hz; lower rows). The highest row shows the P3 recording with a wide (20–2000 Hz) bandpass filtering. The vertical line indicates the latency of the N20 / P20 component, possibly coincident with the arrival of the afferent volley to the somatosensory cortex. High frequency bursts are clearly unaffected by the drug intake.
dipoles had perirolandic locations. Dipole No. 3 was oriented tangentially and was activated at the latencies of both the N20 / P20 and, with inverted polarity, P24 / N24
We further built dipolar models of traces of high frequency SEPs after median nerve stimulation, using the same sequential strategy. SEPs were well explained by a four-dipole model, since no significant activation was found in the 30 ms range of latency. The locations of these four dipoles were similar to those demonstrated for the first four dipoles in low-frequency SEPs. This four-dipole model explained well the SEP distribution (RV values ranging from 6.56 to 12.6%; see Table 3 and Fig. 6); RV values remained substantially unchanged after lorazepam administration (individual values ranging from 6.99 to 13.8%). Dipole strengths of high-frequency models remained unchanged after lorazepam administration (paired t-tests, P.0.05; see Figs. 6 and 7).
Table 3 Coordinates of dipoles: Low frequency median nerve SEP model Dipole 1 x Mean SD
0.21 2.86
Dipole 2 z
theta
phi
x
y
z
theta
phi
x
y
z
theta
phi
210.78 16.97
254.83 4.40
2165.00 13.80
227.63 66.85
214.54 10.84
213.40 6.30
211.80 15.16
282.13 76.85
14.35 52.35
237.33 7.15
8.14 4.33
43.60 6.62
241.33 9.50
212.75 6.75
x
y
z
theta
phi
233.33 7.53
223.94 65.01
232.67 9.67
0.43 14.05
Dipole 4 x Mean SD
Dipole 3
y
Dipole 5 y
238.83 4.83
z 2.76 4.70
theta 42.30 8.75
243.17 7.47
phi 23.57 7.44
53.03 6.76
Coordinates of dipoles: High frequency median nerve SEP model Dipole 1 x 20.67 11.66
Mean SD
Dipole 2 y
z
theta
219.60 18.12
242.60 10.29
y
z
225.00 169.13
phi 50.98 59.44
Dipole 4 x Mean SD
239.80 6.02
4.06 1.39
theta 37.02 6.88
247.40 9.48
phi 25.76 1.73
Dipole 3
x
y
213.46 7.24
210.44 8.22
z 11.02 7.83
theta
phi
x
y
z
theta
phi
260.20 22.75
36.20 30.67
237.20 4.09
8.04 7.35
43.96 2.78
241.40 4.22
212.47 11.43
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
7
Fig. 3. Five-dipole spatiotemporal solution for wide bandpass (20–2000 Hz) median nerve SEPs before and after drug intake; patient No. 4. The residual variance is 1 and 1.9 before and after drug intake, respectively. The source potentials of the dipoles are shown on the left (thick traces: before drug intake. Thin traces: after drug intake); analysis time starts from 8 ms. On the right, three views of the head illustrate the location and orientation of the dipoles. The two upper rows show source potential and location of the subcortical dipoles (dipoles 1 and 2). Source potential and location of the tangential perirolandic dipole (dipole 3) are shown in the third row. The fourth and fifth rows show source potentials and locations of the other two perirolandic dipoles. Note that all perirolandic dipoles (Nos. 3, 4, and 5) are notably weaker after lorazepam administration.
4. Discussion The main finding in this paper is represented by the dissociate behavior of conventional and high-frequency SEPs after single oral dose of lorazepam. While conventional SEPs showed a slight but significant amplitude reduction, high-frequency bursts did not showed any measurable modification. The notion that conventional SEPs are modified by benzodiazepine intake is not entirely new. In past years, in order to verify the reliability of SEP monitoring during surgery, several studies tested the action of anesthetic drugs including benzodiazepines on SEP recordings. Although these earlier studies were performed by using the same simplified recording technique usually utilized in operating room, they were able to demonstrate that benzodiazepines mainly affect the amplitude of cortical SEP components, while latency values showed minor changes or were not affected at all [24,25]. Our present findings suggest that lorazepam administration induces slight but significant amplitude reduction not only of the primary N20 / P20, but also of all following cortical
components. The N20 / P20 response probably reflects the initial excitation of deep layers of the 3b somatosensory area [1,2,8,10,18]. Lorazepam, such as other benzodiazepines, probably exerts its action by enhancing the activity of the BZ1 subtype of GABAA receptors [20]. Tonic and sustained enhancement of such an inhibitory drive could therefore reduce the sharp and synchronous summation of EPSPs that generates the N20 / P20 response. Moreover, this tonic decrease of spiny cell excitability can also explain the reduction of physiologic excitability changes which possibly generate later SEP components [1,2]. Data obtained by dipolar analysis of our traces confirm that lorazepam influence the SEP recording at a cortical level, since both subcortical dipoles, possibly located at the level of brainstem and thalamus, respectively, did not show any significant change. By contrast, high-frequency bursts seem unaffected by lorazepam administration. Besides the visual analysis of raw data, also dipolar analysis clearly showed that cortical generators of high frequency bursts are unaffected by lorazepam administration. In fact, high frequency SEPs in
8
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
Fig. 4. Five-dipole spatiotemporal solution for wide bandpass (20–2000 Hz) median nerve SEPs before and after drug intake; patient No. 2. The residual variance is 1.98 and 2.3 before and after drug intake, respectively. Same presentation as in Fig. 3. All perirolandic dipoles (Nos. 3, 4, and 5) are weaker after lorazepam administration.
Fig. 5. The histogram shows the average strength of each dipole in low-frequency median nerve SEP modeling from all six volunteers, before (black columns) and after (ragged columns) lorazepam administration. Bars above each column represent standard deviations. Strengths of all cortical dipoles were significantly reduced after drug administration (asterisk: paired t-test, P,0.05).
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
9
Fig. 6. Four-dipole spatiotemporal solution for high frequency (500–700 Hz) median nerve SEPs before and after drug intake; patient No. 1. The residual variance is 7.3 and 8.2 before and after drug intake, respectively. The source potentials of the dipoles are shown on the left (black traces: before drug intake. Red traces: after drug intake); analysis time starts from 8 ms. On the right, three views of the head illustrate the location and orientation of the dipoles. The two upper rows show source potential and location of the subcortical dipoles (dipoles 1 and 2). The third and fourth rows show source potentials and locations of the perirolandic dipoles (dipoles 3 and 4). The time courses of all dipoles are unaffected by the drug intake.
Fig. 7. The histogram shows the average strength of each dipole in high-frequency median nerve SEP modeling from all six volunteers, before (black columns) and after (ragged columns) lorazepam administration. Bars above each column represent standard deviations. Strengths of all dipoles remained substantially unchanged after drug administration.
10
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
the 20–25 ms range of latency are explained by two perirolandic sources, which remained substantially unmodified after drug intake. Although this part of highfrequency burst is recorded at the same sites and in the same latency range of the N20 / P20 response, previous studies provided several evidences that it is generated by different cell populations. High-frequency bursts disappear during sleep while the N20 increases or remains unchanged [14,29]; moreover, 600 Hz bursts exhibit a significantly higher short-term variability than low-frequency responses, so that it has been suggested that they might correlate to more variable aspects of the somatosensory processing, such as floating attentional shifts [17]. Many classes of specialized cortical cells, such as pyramidal chattering cells and fast-spiking inhibitory interneurons, are capable to discharge high-frequency bursts [7]. Our present data cannot obviously specify what class of cell account for high-frequency SEP generation, but they clearly indicate that 600 Hz burst are insensitive to the increase of intracortical GABAergic drive. This finding is in agreement with another report suggesting that the enhancement of GABAergic transmission subsequent to valproate administration has no effect on high frequency SEPs [17]. On the contrary, slight but significant modifications of high frequency SEPs after lorazepam administration, mainly concerning their latency, have been found in another study [15]. This apparent discrepancy with our data can be explained by considering the higher dose of lorazepam utilized by Haueisen et al. in the above-mentioned report [15]. It is well known that latency shifts of postsynaptic cortical responses are mainly caused by conduction slowing preceding the arrival of the afferent volley to the cortex [24–26]; it is thus conceivable that higher doses of lorazepam might act also at subcortical levels, thus causing slight latency delays of genuine cortical responses. Previous literature provided evidence that both cortical and subcortical generators contribute to scalp-evoked high-frequency SEPs [6,7]. Also in our study, part of scalp-recorded SEPs was explained by subcortical sources. Lastly, a recent magnetoelectroencephalographic study [16] suggests that burst whose latency is coincident to the ascending slope of the primary N20 probably reflects presynaptic activity. Therefore, it is conceivable that subcortical generators of bursts might be affected by administration of lorazepam doses higher than that we used in the present study, thus entailing slight latency shifts of the scalp-recorded response. It remains to be explained why lorazepam administration does not induce the same changes of high-frequency SEPs as those observed during sleep. It should be considered that the finding of decreased high frequency bursts during lowered vigilance may not reflect uniquely a decreased capacity of single cell bursting, but it could be due to increased burst timing jitter leading to partial cancellation with averaging [7]. This hypothesis agrees with the increased latency of high frequency SEPs by high doses of
lorazepam [15] and explains the lack of any high frequency burst modification in our subjects who received lower doses of this drug. Moreover, sleep involves also mechanisms different from the increase of the GABAergic tone, which were not probably activated by lorazepam administration in our subjects, explicitly asked to stay awake. In conclusion, our study confirms that high frequency SEPs occurring in the 20–25 ms latency range and conventional low-frequency cortical SEPs are probably generated by different cell populations. Moreover, it strongly indicates that specialized cell populations generating cortical high frequency SEPs are substantially insensitive to GABAergic inhibition.
References [1] T. Allison, G. Mc Carthy, C.C. Wood, T.M. Darcey, D.D. Spencer, P.D. Williamson, Human cortical potentials evoked by stimulation of the median nerve. I. Cytoarchitectonic areas generating short-latency activity, J. Neurophysiol. 62 (1989) 694–710. [2] T. Allison, G. Mc Carthy, C.C. Wood, S.J. Jones, Potentials evoked in human and monkey cerebral cortex by stimulation of the median nerve, Brain 114 (1991) 2465–2503. [3] E.G. Bradshaw, A.A. Ali, B.A. Mulley, R.M. Rye, Plasma concentrations and clinical effects of lorazepam after oral administration, Br. J. Anaesth. 53 (5) (1981) 517–522. ¨ [4] H. Buchner, L. Adams, A. Muller, I. Ludwig, A. Knepper, A. Thron, K. Niemann, M. Scherg, Somatotopy of human hand somatosensory cortex revealed by dipole source analysis of early somatosensory evoked potentials and 3D-NMR tomography, Electroencephalogr. Clin. Neurophysiol. 96 (1995) 121–134. [5] H. Buchner, T.D. Waberski, M. Fuchs, H.A. Wishmann, R. Beck¨ mann, A. Rienacker, Origin of P16 median nerve SEP component identified by dipole source analysis. Subthalamic or within the thalamo–cortical radiations?, Exp. Brain Res. 104 (1995) 511–518. [6] G. Curio, B.M. Mackert, M. Burghoff, R. Koetitz, K. Abraham¨ Fuchs, W. Harer, Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system, Electroencephalogr. Clin. Neurophysiol. 91 (1994) 483–487. [7] G. Curio, Linking 600 Hz ‘spikelike’ EEG / MEG wavelets (‘sbursts’) to cellular substrates. Concepts and caveats, J. Clin. Neurophysiol. 17 (4) (2000) 377–396. [8] J.E. Desmedt, G. Cheron, Central somatosensory conduction in man: neuronal generators and interpeak latencies of the far-field components recorded from neck and right or left scalp and earlobes, Electroencephalogr. Clin. Neurophysiol. 50 (1980) 382–403. [9] J.E. Desmedt, T.H. Nguyen, M. Bourguet, Bit-mapped color imaging of human evoked potentials with reference to the N20, P22, P27 and N30 somatosensory responses, Electroencephalogr. Clin. Neurophysiol. 68 (1987) 1–19. [10] J.E. Desmedt, I. Ozaki, SEPs to finger joint input lack the N20–P20 response that is evoked by tactile inputs: contrast between cortical generators in areas 3b and 2 in humans, Electroencephalogr. Clin. Neurophysiol. 80 (1991) 513–521. [11] T. Emori, T. Yamada, Y. Seki, A. Yasuhara, K. Ando, Y. Honda, A.A. Leis, P. Vachatimanont, Recovery functions of fast frequency potentials in the initial negative wave of median SEP, Electroencephalogr. Clin. Neurophysiol. 78 (1991) 116–123. [12] H. Franssen, D.F. Stegemann, J. Moleman, R.P. Schoobar, Dipole modelling of median nerve SEPs in normal subjects and patients
D. Restuccia et al. / Brain Research 946 (2002) 1 – 11
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20] [21]
with small subcortical infarcts, Electroencephalogr. Clin. Neurophysiol. 84 (1992) 40–47. ` Unmasking of cortical L. Garcia-Larrea, H. Bastuji, F. Mauguiere, SEP components by changes in stimulus rates: a topographic study, Electroencephalogr. Clin. Neurophysiol. 84 (1992) 71–83. I. Hashimoto, T. Mashiko, T. Imada, Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex, Electroencephalogr. Clin. Neurophysiol. 100 (1996) 189–203. J. Haueisen, T. Heuer, H. Nowak, J. Liepert, C. Weiller, Y. Okada, G. Curio, The influence of lorazepam on somatosensory-evoked fast frequency (600 Hz) activity in MEG, Brain Res. 874 (2000) 10–14. J. Haueisen, B. Schack, T. Meier, G. Curio, Y. Okada, Multiplicity in the high-frequency signals during the short-latency somatosensory evoked cortical activity in humans, Clin. Neurophysiol. 112 (2001) 1316–1325. F. Klostermann, G. Nolte, G. Curio, Independent short-term variability of spike-like (600 Hz) and postsynaptic (N20) cerebral SEP components, Neuroreport 12 (2001) 349–352. ¨ H. Luders, R.P. Lesser, J. Hahn, D.S. Dinner, G. Klem, Cortical somatosensory evoked potentials in response to hand stimulation, J. Neurosurg. 58 (1983) 885–894. B.M. Mackert, S. Weisenbach, G. Nolte, G. Curio, Rapid recovery (20 ms) of human 600 Hz electroencephalographic wavelets after double stimulation of sensory nerves, Neurosci. Lett. 286 (2000) 83–86. R.M. McKernan, P.J. Whiting, Which GABA-receptor subtypes really occur in the brain, Trends Neurosci. 19 (1996) 139–143. D. Restuccia, M. Valeriani, C. Barba, D. Le Pera, P. Tonali, F.
[22] [23]
[24]
[25] [26]
[27]
[28]
[29]
11
` Mauguiere, Different contribution of joint and cutaneous inputs to early scalp somatosensory evoked potentials, Muscle Nerve 22 (1999) 910–919. M. Scherg, J. Vajar, T.W. Picton, A source analysis of the human auditory potentials, J. Cogn. Neurosci. 1 (1989) 336–354. M. Scherg, Fundamentals of dipole source potential analysis, in: F. Grandoni, M. Hoke, G.L. Romani (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials, Advances in Audiology, Vol. 6, Karger, Basel, 1990, pp. 40–69. T.B. Sloan, M.L. Fugina, J.R. Toleikis, Effects of midazolam on median nerve somatosensory evoked potentials, Br. J. Anaesth. 64 (5) (1990) 590–593. T.B. Sloan, Anesthetic effects on electrophysiologic recordings, J. Clin. Neurophysiol. 15 (1998) 217–226. A. Todorova, Effects of diazepam and the specific benzodiazepine antagonist flumanezil on somatosensory evoked potentials in rat, Arch. Int. Pharmacodyn. Ther. 321 (1993) 14–29. M. Valeriani, D. Restuccia, V. Di Lazzaro, D. Le Pera, P. Tonali, The pathophysiology of giant SEPs in cortical myoclonus: a scalp topography and dipolar source modeling study, Electroencephalogr. Clin. Neurophysiol. 104 (1997) 122–131. M. Valeriani, D. Restuccia, V. Di Lazzaro, D. Le Pera, C. Barba, P. ` Tonali, F. Mauguiere, Dipolar sources of the early scalp somatosensory evoked potentials to upper limb stimulation. Effect of increasing stimulus rates, Exp. Brain Res. 120 (1998) 306–315. T. Yamada, S. Kameyama, Y. Fuchigami, Y. Nakazumi, Q.S. Dickins, J. Kimura, Changes of short latency somatosensory evoked potentials in sleep, Electroencephalogr. Clin. Neurophysiol. 70 (1988) 126–136.