Intrathalamic non-propagating generators of high-frequency (1000 Hz) somatosensory evoked potential (SEP) bursts recorded subcortically in man

https://doi.org/10.1016/S1388-2457(02)00119-0Get rights and content

Abstract

Objectives: Recently, bursts of high-frequency (1000 Hz) median nerve somatosensory evoked potential (SEP) wavelets were recorded subcortically near and inside the thalamus from deep brain electrodes implanted for tremor therapy. This study aimed to clarify whether these subcortical SEP bursts reflect evoked axonal volleys running in the thalamocortical radiation or a locally restricted intrathalamic response.

Methods: During deep brain electrode implantation, median nerve SEP were recorded in 7 patients sequentially along the subcortical stereotactic trajectory at sites +20 and +10 mm above the respective target nucleus (ventral intermediate thalamus or nucleus subthalamicus). Low- and high-frequency SEP components (corner frequency 430 Hz) were analyzed separately with respect to peak latency and amplitude as they changed along the recording trajectory.

Results: Individual wavelets of the subcortical 1000 Hz SEP burst showed fixed peak latencies independent from the depth of the electrode penetration; they increased markedly in amplitude with decreasing distance to the thalamus. In contrast, the amplitude gradient between the two recording sites was shallower for the low-frequency SEP component, which peaked earlier at the lower recording site.

Conclusions: Subcortically recorded 1000 Hz SEP wavelet bursts predominantly reflect locally restricted near-field activity, presumably generated in the somatosensory relay nucleus. In contrast, the variable peak latency of the subcortical low-frequency component could reflect postsynaptic potentials sequentially evoked during passage of the lemniscal afferences curving through the thalamus and contributions from the thalamocortical radiation.

Introduction

Neurosurgical interventions allow for intracerebral near-field recordings of somatosensory evoked potentials (SEP) from the deep lemniscal and thalamocortical system. Such studies on subcortically recorded SEP helped to establish concepts on the generators of the earliest SEP components which can be picked up in part as far-field micropotentials on the scalp (Abbruzzese et al., 1978, Allison and Hume, 1981, Suzuki and Mayanagi, 1984, Tsuji et al., 1984, Albe-Fessard et al., 1986, Katayama and Tsubokawa, 1987, Morioka et al., 1989, Yamashiro et al., 1989, Vanderzant et al., 1991): corresponding to the scalp responses peaking at about 9, 11, 13 and 14 ms after median nerve stimulation, a successive activation of the lemniscal system at the spinal entry zone of the peripheral nerve, the dorsal column, the cuneate nucleus and the medial lemniscus was demonstrated (Allison and Hume, 1981, Suzuki and Mayanagi, 1984). For the component peaking at about 16 ms poststimulus (P16), which exhibits a duration of several milliseconds in subcortical recordings, a local postsynaptic generation in the somatosensory thalamic relay nucleus was proposed (Mauguiere et al., 1983, Hashimoto, 1984, Tsuji et al., 1984, Vanderzant et al., 1991). Recent subcortical recordings showed that this slow component was superimposed by a burst of high-frequency (about 1000 Hz) low-amplitude wavelets, extending from 14 to 22 ms post-stimulus (Klostermann et al., 1999).

When SEP are recorded non-invasively at the scalp (Emerson et al., 1988, Yamada et al., 1988), electroencephalogram (EEG) mapping analyses of burst SEP (spectral energy maximum around 600 Hz) suggested early (around 16 ms) deep generators, presumably being located in the near-thalamic thalamocortical radiation, as well as later (around 20 ms) superficial cortical generators (Gobbele et al., 1998; Gobbele et al., 1999). The radiation fiber model is consistent with the spike-like character of the burst wavelets, pointing to fast axonal generators rather than to slower postsynaptic responses (cf. Allison and Hume, 1981, Katayama and Tsubokawa, 1987). The more superficial burst components are superimposed on the primary cortical N20 component of median nerve SEP and could be interpreted at least partially as the presynaptic arrival of thalamic activity at the primary somatosensory (S-I) cortex. Notably, additional burst components generated intracortically are likely to contribute (Curio, 2000, Curio et al., 1994, Curio et al., 1997), with a recent emphasis on cortical interneurons discharging synchroneously at 600 Hz (Hashimoto et al., 1996, Swadlow et al., 1998, Jones et al., 2000).

A still unresolved, critical issue is the difference in intraburst frequencies recorded subcortically (∼1000 Hz) and at the scalp (∼600 Hz). Specifically, with regard to the 1000 Hz burst components which can be recorded at the thalamus in patients the question needs to be addressed how to discriminate intrathalamic stationary sources from generators propagating in suprathalamic white matter. To this end, SEP were recorded here at different brain depths when stereotactically advancing the electrode towards the respective thalamic targets, in order to assess possible latency and amplitude shifts along the recording trajectory.

Section snippets

Methods

In 7 patients with movement disorders (6 Parkinson disease, one essential tremor; 6 male, one female; 67.4±7.8 years, range: 58–78 years), SEPs were recorded intraoperatively from the electrode implant for deep brain stimulation. None of the patients suffered from a somatosensory impairment; all patients gave informed consent under a study protocol approved by the local Ethics Committee. When driving the electrode from a frontal burrhole towards the thalamic target site, recordings were

Results

In all patients between 4 and 7 burst peaks (mean±SD=5.4±1.2), were identified reliably at both subcortical recording sites, occurring between 13.7 and 24.7 ms post-stimulus. When the successive burst peaks of each subject were paired according to their intra-burst positions between recordings from the two electrode sites, no significant peak latency shifts were observed (Fig. 1). As the high-frequency burst duration differed between subjects, peaks 1 to 4 could be compared between both

Discussion

The present study provided two main results. First, the peak latencies of high-frequency (1000 Hz) wavelets in subcortical SEP bursts were found stable across two recording sites spaced 10 mm apart in depth, contrasting with a latency increase of subcortical low-frequency SEP between lower and more superficial recording sites. Second, a marked amplitude decrease of the 1000 Hz wavelet burst was observed with increasing distance from the thalamus, unlike the shallower amplitude gradient of the

Acknowledgements

Supported by DFG grants KI 1276/1-1 and Ma 1782/1-4.

References (35)

Cited by (0)

View full text