- Email: [email protected]
Comparison of somatosensory evoked high-frequency oscillations after posterior tibial and median nerve stimulation
Clinical Neurophysiology 110 (1999) 1948±1952 www.elsevier.com/locate/clinph
Comparison of somatosensory evoked high-frequency oscillations after posterior tibial and median nerve stimulation Shuji Nakano a, b,*, Isao Hashimoto b, c a
Health Care Division, Yanagi-cho Works, Toshiba Corporation, 70 Yanagi-cho, Kawasaki 210-8501, Japan b Department of Psychophysiology, Tokyo Institute of Psychiatry, Tokyo 156-8585, Japan c Department of Integrative Physiology, National Institute for Physiological Sciences, Okazaki 444-8585, Japan Accepted 26 May 1999
Abstract Objective: We compared the high-frequency oscillations (HFOs) evoked by posterior tibial nerve (PTN) and median nerve (MN) stimulation. Methods: Somatosensory evoked potentials (SEPs) were recorded with a ®lter set at 10±2000 Hz to right PTN and to right MN stimulation in 10 healthy subjects. The HFOs were obtained by digitally ®ltering the wide-band SEPs with a band-pass of 300±900 Hz. Results: HFOs were recorded in 8 of the 10 subjects for PTN, and in all subjects for MN stimulation. The HFOs after both PTN and MN stimulation started approximately at or after the onset of the primary cortical response (P37 and N20) and ended around the middle of the second slope. HFO amplitudes and area after PTN stimulation were signi®cantly smaller than those after MN stimulation. HFO duration after PTN stimulation was markedly longer than that after MN stimulation. However, HFO interpeak latencies did not differ between the two nerves. Conclusions: The present ®ndings suggest that the HFOs after PTN and MN stimulation re¯ect a neural mechanism common to the hand and foot somatosensory cortex. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: High-frequency oscillations; Posterior tibial nerve stimulation; Median nerve stimulation; Primary somatosensory responses (P37 and N20)
1. Introduction High-frequency oscillations (HFOs) in the range of 300± 900 Hz have been shown to occur simultaneously with the primary response (N20) of the somatosensory cortex after median nerve (MN) stimulation recorded electrically (Cracco and Cracco, 1976; Eisen et al., 1984; Yamada et al., 1988; Emori et al., 1991; Sonoo et al., 1997; Gobbele et al., 1998; Ozaki et al., 1998) and magnetically (Curio et al., 1994; Hashimoto et al., 1996, 1999). The HFOs reportedly occur primarily in response to MN stimulation. Only a few studies have reported a response to stimulation of other somatosensory nerves; ulnar nerve (Curio et al., 1997), common peroneal nerve (Eisen et al., 1984) and posterior tibial nerve (PTN) (Nakano and Hashimoto, 1998; Sakuma and Hashimoto, 1999; Sakuma et al., 1999). Eisen et al. (1984) reported that the interpeak latencies of HFOs differed between MN and common peroneal nerve. There have been no other reports, however, compar* Corresponding author. Address a. Tel.: 181-44-548-5568; fax: 18144-520-5801.
ing the interpeak latencies and other parameters of HFOs between lower and upper limb nerve stimulation. In this study, we examined somatosensory evoked HFOs following PTN and MN stimulation in the same subjects, to clarify whether there is a difference in the HFOs between lower and upper limb nerve stimulation. 2. Materials and methods We studied 10 healthy male subjects, aged between 22 and 32 years (26:3 ^ 2:8; mean ^ SD). The subjects lay supine on a bed and were instructed to stay awake with their eyes open. The electrical stimuli of 0.2 ms duration were delivered to the right PTN at the ankle (cathode proximal) and to the right MN at the wrist at a regular interval with a repetition rate of 4 Hz. The stimulus intensity was 3 times the sensory threshold. The original SEPs were obtained using a band-pass ®lter of 10±2000 Hz. Recording electrodes were placed on Fz, C3 0 (2 cm posterior to C3) and Cz 0 (2 cm posterior to Cz) (international 10±20 system). The right ear served as a reference. Electrode impedances were less than 5 kV. Data were stored on ¯oppy disks for later
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S13 88-2457(99)0014 5-5
CLINPH 98119
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
Fig. 1. Original wide-band (10±2000 Hz), low-pass (10±300 Hz) and highpass ®ltered (300±900 Hz) somatosensory evoked potentials from Cz 0 -Fz lead following posterior tibial nerve stimulation. The highpass trace shows a short burst of high-frequency oscillations (HFOs) around P37, which can be recognized in the wide-band record as notches or even in¯ections overlying the P37 response. The HFOs started approximately at or after the onset of the primary cortical response (P37) and ended in the middle of the second slope, showing several peaks.
off-line analysis. Off-line electronic subtraction of two waves with a right ear reference gave bipolar montages: C3 0 to Fz and Cz 0 to Fz. An epoch of 50 ms duration, with a delay of 10 ms after right PTN stimulation, was digitized at a 10 kHz sampling rate. An epoch of 30 ms duration, with a delay of 6 ms after right MN stimulation, was digitized at a 17 kHz sampling rate. Sweeps contaminated with large artifacts were rejected automatically. Two thousand responses were averaged for PTN SEPs, while 1000 responses were averaged for MN SEPs. Two averages were superimposed to con®rm the reproducibility of the responses. For separation and isolation of the HFOs from underlying P37 and N20 primary cortical responses, the wide-band (10± 2000 Hz) recorded responses were digitally ®ltered through a band-pass of 300±900 Hz and 10±300 Hz, using a Neuropack Sigma computer system (Nihon Kohden). The HFOs could be clearly distinguished from cortical P37 and N20 potentials using cephalic bipolar leads after band-pass ®ltering. Therefore, interpeak latencies, amplitudes, number of negative peaks, duration and area of the HFOs were measured using the responses obtained from a Cz 0 -Fz lead for PTN stimulation, and a C3 0 -Fz lead for MN stimulation. For identi®cation of the HFOs, the wavelets after the onset of primary cortical response (P37 and N20) with an amplitude of twice or more that of the background noise level were considered as the signal. The noise level was measured between 15 and 25 ms after the stimulus for PTN, and between 9 and 15 ms after the stimulus for MN stimulation, respectively. Interpeak latencies of the HFOs were averaged with the values measured from two successive negative peaks and two successive positive peaks. The amplitudes of the HFOs were averaged with the values measured as the vertical distances between two successive peaks. The duration of the HFOs was measured between the
1949
onset and endpoint. The area (nV´ms) was measured as the total area of positive and negative de¯ections from the baseline. Primary cortical response (P37 and N20) were bandpass ®ltered (10±300 Hz). The amplitude of the primary cortical response was measured as the vertical distance between the onset and peak of the primary cortical response. The duration of the primary cortical response was measured as the time between its onset and the second peak (N45 and P27) for PTN and MN stimulation, respectively. A Fast Fourier transform (FFT) analysis of the wide-band records was performed using a Hanning window between 10 and 60 ms post-stimulus after PTN stimulation and between 6 and 36 ms post-stimulus after MN stimulation in a typical normal subject, to compare with the frequency estimated with the interpeak latencies. Differences in the above parameters were statistically analyzed by paired t test, and P , 0:05 was considered to be signi®cant.
3. Results Figs.1 and 2 show the original wide-band (10±2000 Hz), low-pass (10±300 Hz) and high-pass (300±900 Hz) ®ltered SEPs following PTN and MN stimulation. The high-pass trace shows a short burst of HFOs around P37 and N20 which can be recognized already in the wide-band records as small in¯ections superimposed on the P37 and N20 responses. After ®ltering the original wide-band SEPs with a band-pass of 300±900 Hz, these HFOs were isolated and could be discerned from the background noise in 8 of the 10 subjects for PTN stimulation, and in all 10 subjects for MN stimulation. The HFOs started approximately at or after the onset of the primary cortical response (P37 and N20) and ended between the middle and end of the second
Fig. 2. Original wide-band (10±2000 Hz), low-pass (10±300 Hz) and highpass ®ltered (300±900 Hz) somatosensory evoked potentials from C3 0 -Fz lead following median nerve stimulation in the same subject of Fig. 1. The highpass trace shows a short burst of high- frequency oscillations (HFOs) around N20, which can be recognized in the wide-band record as notches or even in¯ections overlying the N20 response. The HFOs started approximately at or after the onset of the primary cortical response (N20) and ended at the end of the second slope, showing several peaks.
1950
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
slope, showing several peaks. HFOs after PTN stimulation end in the middle of the second slope, while HFOs after MN stimulation end around the end of the second slope. Table 1 shows the comparison of HFO and primary cortical response parameters between PTN and MN stimulation in 8 subjects in whom HFOs could be recorded to both PTN and MN stimulation. HFO amplitudes and area after PTN stimulation were signi®cantly smaller than those after MN stimulation. HFO duration after PTN stimulation was significantly longer compared with that after MN stimulation. However, interpeak latencies and the number of negative peaks did not differ signi®cantly. Average interpeak latencies were 1:59 ^ 0:19 ms and 1:57 ^ 0:28 ms for PTN and MN stimulation, indicating the HFO frequencies were approximately 630 Hz for stimulation of both nerves. Fig. 3A,B shows an FFT analysis of the wide-band records after PTN and MN stimulation in a typical normal subject. The main signal energy was distributed broadly between 20 and 300 Hz with a peak around 40±60 Hz, and a weaker signal energy was found between 600±800 Hz with a peak around 680 Hz for both nerves. Average interpeak latencies in this subject were 1.44 and 1.46 ms for PTN and MN stimulation, indicating the HFO frequencies were approximately 690 Hz for both nerve stimulation. The amplitude of the primary cortical response and the HFOs/primary cortical response amplitude ratio were signi®cantly smaller for PTN stimulation than those after MN stimulation. The duration of primary cortical response (P37) after PTN stimulation was signi®cantly longer than that after MN stimulation. 4. Discussion We observed that HFOs after PTN stimulation were superimposed on primary cortical response (P37), and started approximately at or after the onset of the P37 and ended in the middle of the second peak (N45). It has been established that P37 is the ®rst cortical response and is generated by EPSPs in pyramidal cells of area 3b of the foot sensory area (Cruse et al., 1982; Kakigi et al., 1995). HFOs after MN stimulation were also observed in the primary cortical response (N20). HFOs after median nerve
stimulation started approximately at or after the onset of the N20 and ended around the second peak (P27) in this study. Although the HFO endpoint is a little different between PTN and MN, the fundamental characteristics of the HFO temporal pattern after PTN stimulation were similar to those of HFOs after MN stimulation. Interpeak latencies of HFOs after PTN and MN stimulation did not differ, indicating that the HFO frequency is almost the same following both PTN and MN stimulation. FFT analysis in a subject supports that the HFO frequency was almost the same for both nerves. Duration of primary cortical response following PTN stimulation was longer than that following MN stimulation. This is probably due to temporal dispersion of axonal volleys, which tends to occur with longer conducting distances. The absence of difference in HFO interpeak latencies and signi®cant difference in primary cortical response duration between PTN and MN SEPs may re¯ect a high synchrony of HFOs and temporal dispersion of primary cortical responses, suggesting that the generating mechanisms for the HFOs and primary cortical response are different. Swadlow et al. (1998) reported a sharp, local synchrony among putative feed-forward inhibitory interneurons in a rabbit somatosensory cortex. These results support the speculation of Hashimoto et al. (1996) that HFOs represent a localized activity of the GABAergic inhibitory interneurons of layer 4, which have been shown in animal experiments to respond monosynaptically to thalamo-cortical input with a high-frequency (600±900 Hz) burst of short duration (Swadlow, 1989). Eisen et al. (1984) reported that interpeak latencies (frequencies) were 1:8 ^ 0:25 ms (555 Hz) and 1:3 ^ 0:22 ms (769 Hz) for common peroneal and MN stimulation, suggesting a clear difference between the two nerves. The discrepancy between the present and Eisen's studies remains unclear, and further study is needed. There have been no reports comparing the amplitude, duration and area of HFOs following upper and lower limb stimulation. The HFO amplitude after PTN stimulation was markedly smaller than that after MN stimulation. Low HFO and primary cortical response amplitude following PTN stimulation is probably due to a smaller amount of the afferent volleys within the somatosensory cortex and a
Table 1 Parameters of HFOs and primary cortical response following posterior tibial and median nerve stimulation (n 8) a
HFO interpeak latency (ms) HFO amplitude (nV) HFO number of negative peaks HFO duration (ms) HFO area (nV´ms) Primary cortical response amplitude (mV) HFOs/primary cortical response amp. ratio (%) Primary cortical response duration (ms) a b
Mean ^ SD. P , 0:05.
Posterior tibial nerve
Median nerve
P value
1.59 ^ 0.19 28.7 ^ 10.9 5.8 ^ 2.1 9.96 ^ 2.29 89 ^ 35 1.28 ^ 0.18 1.02 ^ 0.41 16.16 ^ 1.46
1.57 ^ 0.28 147.7 ^ 29.9 4.4 ^ 1.3 7.01 ^ 1.69 353 ^ 112 2.03 ^ 0.76 7.88 ^ 1.90 6.80 ^ 1.98
0.8455 ,0.0001 b 0.0828 0.0028 b 0.0001 b 0.0177 b ,0.0001 b ,0.0001 b
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
1951
Fig. 3. (A) Fast Fourier transform (FFT) analysis of the wide-band records of SEPs after posterior tibial nerve stimulation shows two peaks of signal energy; one around 40±60 Hz and one around 600±800 Hz in a normal subject. The power spectrum is illustrated on a log-log scale in which abscissa is frequency of the signals and ordinate indicates the signal energy in an arbitrary unit. B: FFT analysis of the wide-band records of SEPs after median nerve stimulation in the same subject, shows two peaks of signal energy; one around 40±60 Hz and one around 600±800 Hz.
deeper generator source. A magnetoencephalogram (MEG) study showed that the source of HFOs following PTN and MN stimulation is the primary sensory cortex (area 3b) very close to that of P37m (Sakuma and Hashimoto, 1999; Sakuma et al., 1999) and the N20m peak (Hashimoto et al., 1996, 1999), respectively. As the primary cortical sensory area for the foot is anatomically located on the medial surface of the postcentral gyrus in the interhemispheric ®ssure, the generators of P37 and superimposed HFOs should be deeper than those of the median nerve N20. Furthermore, the HFO/primary cortical response amplitude ratio after PTN stimulation was also signi®cantly smaller than that after MN stimulation. This suggests that
HFO amplitude is not simply related to the amplitude of primary cortical response. A longer HFO duration after PTN stimulation, compared with that after MN stimulation, may be related to the prolonged duration of the primary cortical response with PTN stimulation. A smaller HFO area after PTN stimulation is considered to re¯ect an extremely low HFO amplitude. The number of negative peaks did not differ statistically between the two nerves, although there was a trend for an increased number of negative peaks after PTN stimulation. This is probably due to a wide interindividual variability and a relatively small number of subjects examined in this study. HFOs after MN stimulation were recorded in all subjects.
1952
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
However HFOs after PTN stimulation were undetected in 2 of the 10 subjects. This may be due to a low signal strength of the HFOs. In addition, the recording of the signal may be dif®cult in some cases with the ®xed Cz 0 -Fz lead because of an increased variability in the dipole orientation following PTN stimulation (Tsuji and Murai, 1987; Kakigi et al., 1995). Although HFO amplitudes, duration and area differed between PTN and MN stimulation, the HFO interpeak latencies and frequencies did not signi®cantly differ. Furthermore, the temporal pro®le of the HFOs in relation to the underlying primary responses was identical for both nerves. We conclude that HFOs following PTN and MN stimulation re¯ect a neural mechanism common to the stimulation of different somatosensory nerves. Thus, HFOs represent a ubiquitous activity in the primary sensory cortex. References Cracco RQ, Cracco JB. Somatosensory evoked potential in man: far-®eld potentials. Electroenceph clin Neurophysiol 1976;41:460±466. Cruse R, Klem G, Lesser RP, Lueders H. Paradoxical lateralization of cortical potentials evoked by stimulation of posterior tibial nerve. Arch Neurol 1982;39:222±225. Curio G, Mackert BM, Burghoff M, Koetitz R, Abraham-Fuchs K, HaÈrer W. Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system. Electroenceph clin Neurophysiol 1994;91:483±487. Curio G, Mackert BM, Burghoff M, Neuman J, Nolte G, Scherg M, Marx P. Somatotopic source arrangement of 600 Hz oscillatory magnetic ®elds at the human primary somatosensory hand cortex. Neurosci Lett 1997;234:131±134. Eisen A, Roberts K, Low M, Hoirch M, Lawrence P. Questions regarding the sequential neural generator theory of the somatosensory evoked potential raised by digital ®ltering. Electroenceph clin Neurophysiol 1984;59:388±395. Emori T, Yamada T, Seki Y, Yasuhara A, Ando K, Honda Y, Leis AA, Vachatimanont P. Recovery functions of fast frequency potentials in the initial negative wave of median SEP. Electroenceph clin Neurophysiol 1991;78:116±123.
Gobbele R, Buchner H, Curio G. High-frequency (600 Hz) SEP activities originating in the subcortical and cortical human somatosensory system. Electroenceph clin Neurophysiol 1998;108:182±189. Hashimoto I, Mashiko T, Imada T. Somatic evoked high-frequency magnetic oscillations re¯ect activity of inhibitory interneurons in the human somatosensory cortex. Electroenceph clin Neurophysiol 1996;100:189±203. Hashimoto I, Kimura T, Fukushima T, Iguchi Y, Saito Y, Terasaki O, Sakuma K. Reciprocal modulation of somatosensory evoked N20m primary response and high-frequency oscillations by interference stimulation. Clin Neurophysiol 1999;110:1445±1451. Kakigi R, Koyama S, Hoshiyama M, Shimojo M, Kitamura Y, Watanabe S. Topography of somatosensory evoked magnetic ®elds following posterior tibial nerve stimulation. Electroenceph clin Neurophysiol 1995;95:127±134. Nakano S, Hashimoto I. Somatosensory evoked high-frequency oscillations after posterior tibial nerve stimulation. In: Hashimoto I, Kakigi R, editors. Recent advances in human neurophysiology, Amsterdam: Elsevier, 1998. pp. 27±31. Ozaki I, Suzuki C, Yaegashi Y, Baba M, Matsunaga M, Hashimoto I. Highfrequency oscillations in early cortical somatosensory evoked potentials. Electroenceph clin Neurophysiol 1998;108:536±542. Sakuma K, Hashimoto I. High-frequency magnetic oscillations evoked by posterior tibial nerve stimulation. NeuroReport 1999;10:227±230. Sakuma K, Sekihara K, Hashimoto I. Neural source estimation from a timefrequency component of somatic evoked high-frequency magnetic oscillations to posterior tibial nerve stimulation. Clin Neurophysiol 1999;110:1584±1587. Sonoo M, Genba-Shimizu K, Mannen T, Shimizu T. Detailed analysis of the latencies of median nerve somatosensory evoked potential components 2: analysis of subcomponents of the P13/14 and N20 potentials. Electroenceph clin Neurophysiol 1997;104:296±311. Swadlow HA. Efferent neurons and suspected interneurons in S-1 vibrissa cortex of the awake rabbit: receptive ®elds and axonal properties. J Neurophysiol 1989;62:288±308. Swadlow HA, Beloozerova IN, Sirota MG. Sharp, local synchrony among putative feed-forward inhibitory interneurons of rabbit somatosensory cortex. J Neurophysiol 1998;79:567±582. Tsuji S, Murai Y. Variability of initial cortical sensory evoked potentials to posterior tibial nerve stimulation. J UOEH 1987;9:287±298. Yamada T, Kameyama S, Fuchigami Y, Nakazumi Y, Dickins QS, Kimura J. Changes of short latency somatosensory evoked potential in sleep. Electroenceph clin Neurophysiol 1988;70:126±136.
Comparison of somatosensory evoked high-frequency oscillations after posterior tibial and median nerve stimulation Shuji Nakano a, b,*, Isao Hashimoto b, c a
Health Care Division, Yanagi-cho Works, Toshiba Corporation, 70 Yanagi-cho, Kawasaki 210-8501, Japan b Department of Psychophysiology, Tokyo Institute of Psychiatry, Tokyo 156-8585, Japan c Department of Integrative Physiology, National Institute for Physiological Sciences, Okazaki 444-8585, Japan Accepted 26 May 1999
Abstract Objective: We compared the high-frequency oscillations (HFOs) evoked by posterior tibial nerve (PTN) and median nerve (MN) stimulation. Methods: Somatosensory evoked potentials (SEPs) were recorded with a ®lter set at 10±2000 Hz to right PTN and to right MN stimulation in 10 healthy subjects. The HFOs were obtained by digitally ®ltering the wide-band SEPs with a band-pass of 300±900 Hz. Results: HFOs were recorded in 8 of the 10 subjects for PTN, and in all subjects for MN stimulation. The HFOs after both PTN and MN stimulation started approximately at or after the onset of the primary cortical response (P37 and N20) and ended around the middle of the second slope. HFO amplitudes and area after PTN stimulation were signi®cantly smaller than those after MN stimulation. HFO duration after PTN stimulation was markedly longer than that after MN stimulation. However, HFO interpeak latencies did not differ between the two nerves. Conclusions: The present ®ndings suggest that the HFOs after PTN and MN stimulation re¯ect a neural mechanism common to the hand and foot somatosensory cortex. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: High-frequency oscillations; Posterior tibial nerve stimulation; Median nerve stimulation; Primary somatosensory responses (P37 and N20)
1. Introduction High-frequency oscillations (HFOs) in the range of 300± 900 Hz have been shown to occur simultaneously with the primary response (N20) of the somatosensory cortex after median nerve (MN) stimulation recorded electrically (Cracco and Cracco, 1976; Eisen et al., 1984; Yamada et al., 1988; Emori et al., 1991; Sonoo et al., 1997; Gobbele et al., 1998; Ozaki et al., 1998) and magnetically (Curio et al., 1994; Hashimoto et al., 1996, 1999). The HFOs reportedly occur primarily in response to MN stimulation. Only a few studies have reported a response to stimulation of other somatosensory nerves; ulnar nerve (Curio et al., 1997), common peroneal nerve (Eisen et al., 1984) and posterior tibial nerve (PTN) (Nakano and Hashimoto, 1998; Sakuma and Hashimoto, 1999; Sakuma et al., 1999). Eisen et al. (1984) reported that the interpeak latencies of HFOs differed between MN and common peroneal nerve. There have been no other reports, however, compar* Corresponding author. Address a. Tel.: 181-44-548-5568; fax: 18144-520-5801.
ing the interpeak latencies and other parameters of HFOs between lower and upper limb nerve stimulation. In this study, we examined somatosensory evoked HFOs following PTN and MN stimulation in the same subjects, to clarify whether there is a difference in the HFOs between lower and upper limb nerve stimulation. 2. Materials and methods We studied 10 healthy male subjects, aged between 22 and 32 years (26:3 ^ 2:8; mean ^ SD). The subjects lay supine on a bed and were instructed to stay awake with their eyes open. The electrical stimuli of 0.2 ms duration were delivered to the right PTN at the ankle (cathode proximal) and to the right MN at the wrist at a regular interval with a repetition rate of 4 Hz. The stimulus intensity was 3 times the sensory threshold. The original SEPs were obtained using a band-pass ®lter of 10±2000 Hz. Recording electrodes were placed on Fz, C3 0 (2 cm posterior to C3) and Cz 0 (2 cm posterior to Cz) (international 10±20 system). The right ear served as a reference. Electrode impedances were less than 5 kV. Data were stored on ¯oppy disks for later
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S13 88-2457(99)0014 5-5
CLINPH 98119
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
Fig. 1. Original wide-band (10±2000 Hz), low-pass (10±300 Hz) and highpass ®ltered (300±900 Hz) somatosensory evoked potentials from Cz 0 -Fz lead following posterior tibial nerve stimulation. The highpass trace shows a short burst of high-frequency oscillations (HFOs) around P37, which can be recognized in the wide-band record as notches or even in¯ections overlying the P37 response. The HFOs started approximately at or after the onset of the primary cortical response (P37) and ended in the middle of the second slope, showing several peaks.
off-line analysis. Off-line electronic subtraction of two waves with a right ear reference gave bipolar montages: C3 0 to Fz and Cz 0 to Fz. An epoch of 50 ms duration, with a delay of 10 ms after right PTN stimulation, was digitized at a 10 kHz sampling rate. An epoch of 30 ms duration, with a delay of 6 ms after right MN stimulation, was digitized at a 17 kHz sampling rate. Sweeps contaminated with large artifacts were rejected automatically. Two thousand responses were averaged for PTN SEPs, while 1000 responses were averaged for MN SEPs. Two averages were superimposed to con®rm the reproducibility of the responses. For separation and isolation of the HFOs from underlying P37 and N20 primary cortical responses, the wide-band (10± 2000 Hz) recorded responses were digitally ®ltered through a band-pass of 300±900 Hz and 10±300 Hz, using a Neuropack Sigma computer system (Nihon Kohden). The HFOs could be clearly distinguished from cortical P37 and N20 potentials using cephalic bipolar leads after band-pass ®ltering. Therefore, interpeak latencies, amplitudes, number of negative peaks, duration and area of the HFOs were measured using the responses obtained from a Cz 0 -Fz lead for PTN stimulation, and a C3 0 -Fz lead for MN stimulation. For identi®cation of the HFOs, the wavelets after the onset of primary cortical response (P37 and N20) with an amplitude of twice or more that of the background noise level were considered as the signal. The noise level was measured between 15 and 25 ms after the stimulus for PTN, and between 9 and 15 ms after the stimulus for MN stimulation, respectively. Interpeak latencies of the HFOs were averaged with the values measured from two successive negative peaks and two successive positive peaks. The amplitudes of the HFOs were averaged with the values measured as the vertical distances between two successive peaks. The duration of the HFOs was measured between the
1949
onset and endpoint. The area (nV´ms) was measured as the total area of positive and negative de¯ections from the baseline. Primary cortical response (P37 and N20) were bandpass ®ltered (10±300 Hz). The amplitude of the primary cortical response was measured as the vertical distance between the onset and peak of the primary cortical response. The duration of the primary cortical response was measured as the time between its onset and the second peak (N45 and P27) for PTN and MN stimulation, respectively. A Fast Fourier transform (FFT) analysis of the wide-band records was performed using a Hanning window between 10 and 60 ms post-stimulus after PTN stimulation and between 6 and 36 ms post-stimulus after MN stimulation in a typical normal subject, to compare with the frequency estimated with the interpeak latencies. Differences in the above parameters were statistically analyzed by paired t test, and P , 0:05 was considered to be signi®cant.
3. Results Figs.1 and 2 show the original wide-band (10±2000 Hz), low-pass (10±300 Hz) and high-pass (300±900 Hz) ®ltered SEPs following PTN and MN stimulation. The high-pass trace shows a short burst of HFOs around P37 and N20 which can be recognized already in the wide-band records as small in¯ections superimposed on the P37 and N20 responses. After ®ltering the original wide-band SEPs with a band-pass of 300±900 Hz, these HFOs were isolated and could be discerned from the background noise in 8 of the 10 subjects for PTN stimulation, and in all 10 subjects for MN stimulation. The HFOs started approximately at or after the onset of the primary cortical response (P37 and N20) and ended between the middle and end of the second
Fig. 2. Original wide-band (10±2000 Hz), low-pass (10±300 Hz) and highpass ®ltered (300±900 Hz) somatosensory evoked potentials from C3 0 -Fz lead following median nerve stimulation in the same subject of Fig. 1. The highpass trace shows a short burst of high- frequency oscillations (HFOs) around N20, which can be recognized in the wide-band record as notches or even in¯ections overlying the N20 response. The HFOs started approximately at or after the onset of the primary cortical response (N20) and ended at the end of the second slope, showing several peaks.
1950
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
slope, showing several peaks. HFOs after PTN stimulation end in the middle of the second slope, while HFOs after MN stimulation end around the end of the second slope. Table 1 shows the comparison of HFO and primary cortical response parameters between PTN and MN stimulation in 8 subjects in whom HFOs could be recorded to both PTN and MN stimulation. HFO amplitudes and area after PTN stimulation were signi®cantly smaller than those after MN stimulation. HFO duration after PTN stimulation was significantly longer compared with that after MN stimulation. However, interpeak latencies and the number of negative peaks did not differ signi®cantly. Average interpeak latencies were 1:59 ^ 0:19 ms and 1:57 ^ 0:28 ms for PTN and MN stimulation, indicating the HFO frequencies were approximately 630 Hz for stimulation of both nerves. Fig. 3A,B shows an FFT analysis of the wide-band records after PTN and MN stimulation in a typical normal subject. The main signal energy was distributed broadly between 20 and 300 Hz with a peak around 40±60 Hz, and a weaker signal energy was found between 600±800 Hz with a peak around 680 Hz for both nerves. Average interpeak latencies in this subject were 1.44 and 1.46 ms for PTN and MN stimulation, indicating the HFO frequencies were approximately 690 Hz for both nerve stimulation. The amplitude of the primary cortical response and the HFOs/primary cortical response amplitude ratio were signi®cantly smaller for PTN stimulation than those after MN stimulation. The duration of primary cortical response (P37) after PTN stimulation was signi®cantly longer than that after MN stimulation. 4. Discussion We observed that HFOs after PTN stimulation were superimposed on primary cortical response (P37), and started approximately at or after the onset of the P37 and ended in the middle of the second peak (N45). It has been established that P37 is the ®rst cortical response and is generated by EPSPs in pyramidal cells of area 3b of the foot sensory area (Cruse et al., 1982; Kakigi et al., 1995). HFOs after MN stimulation were also observed in the primary cortical response (N20). HFOs after median nerve
stimulation started approximately at or after the onset of the N20 and ended around the second peak (P27) in this study. Although the HFO endpoint is a little different between PTN and MN, the fundamental characteristics of the HFO temporal pattern after PTN stimulation were similar to those of HFOs after MN stimulation. Interpeak latencies of HFOs after PTN and MN stimulation did not differ, indicating that the HFO frequency is almost the same following both PTN and MN stimulation. FFT analysis in a subject supports that the HFO frequency was almost the same for both nerves. Duration of primary cortical response following PTN stimulation was longer than that following MN stimulation. This is probably due to temporal dispersion of axonal volleys, which tends to occur with longer conducting distances. The absence of difference in HFO interpeak latencies and signi®cant difference in primary cortical response duration between PTN and MN SEPs may re¯ect a high synchrony of HFOs and temporal dispersion of primary cortical responses, suggesting that the generating mechanisms for the HFOs and primary cortical response are different. Swadlow et al. (1998) reported a sharp, local synchrony among putative feed-forward inhibitory interneurons in a rabbit somatosensory cortex. These results support the speculation of Hashimoto et al. (1996) that HFOs represent a localized activity of the GABAergic inhibitory interneurons of layer 4, which have been shown in animal experiments to respond monosynaptically to thalamo-cortical input with a high-frequency (600±900 Hz) burst of short duration (Swadlow, 1989). Eisen et al. (1984) reported that interpeak latencies (frequencies) were 1:8 ^ 0:25 ms (555 Hz) and 1:3 ^ 0:22 ms (769 Hz) for common peroneal and MN stimulation, suggesting a clear difference between the two nerves. The discrepancy between the present and Eisen's studies remains unclear, and further study is needed. There have been no reports comparing the amplitude, duration and area of HFOs following upper and lower limb stimulation. The HFO amplitude after PTN stimulation was markedly smaller than that after MN stimulation. Low HFO and primary cortical response amplitude following PTN stimulation is probably due to a smaller amount of the afferent volleys within the somatosensory cortex and a
Table 1 Parameters of HFOs and primary cortical response following posterior tibial and median nerve stimulation (n 8) a
HFO interpeak latency (ms) HFO amplitude (nV) HFO number of negative peaks HFO duration (ms) HFO area (nV´ms) Primary cortical response amplitude (mV) HFOs/primary cortical response amp. ratio (%) Primary cortical response duration (ms) a b
Mean ^ SD. P , 0:05.
Posterior tibial nerve
Median nerve
P value
1.59 ^ 0.19 28.7 ^ 10.9 5.8 ^ 2.1 9.96 ^ 2.29 89 ^ 35 1.28 ^ 0.18 1.02 ^ 0.41 16.16 ^ 1.46
1.57 ^ 0.28 147.7 ^ 29.9 4.4 ^ 1.3 7.01 ^ 1.69 353 ^ 112 2.03 ^ 0.76 7.88 ^ 1.90 6.80 ^ 1.98
0.8455 ,0.0001 b 0.0828 0.0028 b 0.0001 b 0.0177 b ,0.0001 b ,0.0001 b
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
1951
Fig. 3. (A) Fast Fourier transform (FFT) analysis of the wide-band records of SEPs after posterior tibial nerve stimulation shows two peaks of signal energy; one around 40±60 Hz and one around 600±800 Hz in a normal subject. The power spectrum is illustrated on a log-log scale in which abscissa is frequency of the signals and ordinate indicates the signal energy in an arbitrary unit. B: FFT analysis of the wide-band records of SEPs after median nerve stimulation in the same subject, shows two peaks of signal energy; one around 40±60 Hz and one around 600±800 Hz.
deeper generator source. A magnetoencephalogram (MEG) study showed that the source of HFOs following PTN and MN stimulation is the primary sensory cortex (area 3b) very close to that of P37m (Sakuma and Hashimoto, 1999; Sakuma et al., 1999) and the N20m peak (Hashimoto et al., 1996, 1999), respectively. As the primary cortical sensory area for the foot is anatomically located on the medial surface of the postcentral gyrus in the interhemispheric ®ssure, the generators of P37 and superimposed HFOs should be deeper than those of the median nerve N20. Furthermore, the HFO/primary cortical response amplitude ratio after PTN stimulation was also signi®cantly smaller than that after MN stimulation. This suggests that
HFO amplitude is not simply related to the amplitude of primary cortical response. A longer HFO duration after PTN stimulation, compared with that after MN stimulation, may be related to the prolonged duration of the primary cortical response with PTN stimulation. A smaller HFO area after PTN stimulation is considered to re¯ect an extremely low HFO amplitude. The number of negative peaks did not differ statistically between the two nerves, although there was a trend for an increased number of negative peaks after PTN stimulation. This is probably due to a wide interindividual variability and a relatively small number of subjects examined in this study. HFOs after MN stimulation were recorded in all subjects.
1952
S. Nakano, I. Hashimoto / Clinical Neurophysiology 110 (1999) 1948±1952
However HFOs after PTN stimulation were undetected in 2 of the 10 subjects. This may be due to a low signal strength of the HFOs. In addition, the recording of the signal may be dif®cult in some cases with the ®xed Cz 0 -Fz lead because of an increased variability in the dipole orientation following PTN stimulation (Tsuji and Murai, 1987; Kakigi et al., 1995). Although HFO amplitudes, duration and area differed between PTN and MN stimulation, the HFO interpeak latencies and frequencies did not signi®cantly differ. Furthermore, the temporal pro®le of the HFOs in relation to the underlying primary responses was identical for both nerves. We conclude that HFOs following PTN and MN stimulation re¯ect a neural mechanism common to the stimulation of different somatosensory nerves. Thus, HFOs represent a ubiquitous activity in the primary sensory cortex. References Cracco RQ, Cracco JB. Somatosensory evoked potential in man: far-®eld potentials. Electroenceph clin Neurophysiol 1976;41:460±466. Cruse R, Klem G, Lesser RP, Lueders H. Paradoxical lateralization of cortical potentials evoked by stimulation of posterior tibial nerve. Arch Neurol 1982;39:222±225. Curio G, Mackert BM, Burghoff M, Koetitz R, Abraham-Fuchs K, HaÈrer W. Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system. Electroenceph clin Neurophysiol 1994;91:483±487. Curio G, Mackert BM, Burghoff M, Neuman J, Nolte G, Scherg M, Marx P. Somatotopic source arrangement of 600 Hz oscillatory magnetic ®elds at the human primary somatosensory hand cortex. Neurosci Lett 1997;234:131±134. Eisen A, Roberts K, Low M, Hoirch M, Lawrence P. Questions regarding the sequential neural generator theory of the somatosensory evoked potential raised by digital ®ltering. Electroenceph clin Neurophysiol 1984;59:388±395. Emori T, Yamada T, Seki Y, Yasuhara A, Ando K, Honda Y, Leis AA, Vachatimanont P. Recovery functions of fast frequency potentials in the initial negative wave of median SEP. Electroenceph clin Neurophysiol 1991;78:116±123.
Gobbele R, Buchner H, Curio G. High-frequency (600 Hz) SEP activities originating in the subcortical and cortical human somatosensory system. Electroenceph clin Neurophysiol 1998;108:182±189. Hashimoto I, Mashiko T, Imada T. Somatic evoked high-frequency magnetic oscillations re¯ect activity of inhibitory interneurons in the human somatosensory cortex. Electroenceph clin Neurophysiol 1996;100:189±203. Hashimoto I, Kimura T, Fukushima T, Iguchi Y, Saito Y, Terasaki O, Sakuma K. Reciprocal modulation of somatosensory evoked N20m primary response and high-frequency oscillations by interference stimulation. Clin Neurophysiol 1999;110:1445±1451. Kakigi R, Koyama S, Hoshiyama M, Shimojo M, Kitamura Y, Watanabe S. Topography of somatosensory evoked magnetic ®elds following posterior tibial nerve stimulation. Electroenceph clin Neurophysiol 1995;95:127±134. Nakano S, Hashimoto I. Somatosensory evoked high-frequency oscillations after posterior tibial nerve stimulation. In: Hashimoto I, Kakigi R, editors. Recent advances in human neurophysiology, Amsterdam: Elsevier, 1998. pp. 27±31. Ozaki I, Suzuki C, Yaegashi Y, Baba M, Matsunaga M, Hashimoto I. Highfrequency oscillations in early cortical somatosensory evoked potentials. Electroenceph clin Neurophysiol 1998;108:536±542. Sakuma K, Hashimoto I. High-frequency magnetic oscillations evoked by posterior tibial nerve stimulation. NeuroReport 1999;10:227±230. Sakuma K, Sekihara K, Hashimoto I. Neural source estimation from a timefrequency component of somatic evoked high-frequency magnetic oscillations to posterior tibial nerve stimulation. Clin Neurophysiol 1999;110:1584±1587. Sonoo M, Genba-Shimizu K, Mannen T, Shimizu T. Detailed analysis of the latencies of median nerve somatosensory evoked potential components 2: analysis of subcomponents of the P13/14 and N20 potentials. Electroenceph clin Neurophysiol 1997;104:296±311. Swadlow HA. Efferent neurons and suspected interneurons in S-1 vibrissa cortex of the awake rabbit: receptive ®elds and axonal properties. J Neurophysiol 1989;62:288±308. Swadlow HA, Beloozerova IN, Sirota MG. Sharp, local synchrony among putative feed-forward inhibitory interneurons of rabbit somatosensory cortex. J Neurophysiol 1998;79:567±582. Tsuji S, Murai Y. Variability of initial cortical sensory evoked potentials to posterior tibial nerve stimulation. J UOEH 1987;9:287±298. Yamada T, Kameyama S, Fuchigami Y, Nakazumi Y, Dickins QS, Kimura J. Changes of short latency somatosensory evoked potential in sleep. Electroenceph clin Neurophysiol 1988;70:126±136.