3.1 Overview of Microelectrode Recording
Since its introduction by Albe-Fessard and Guiot  in the early 1960s, microelectrode recording (MER) has been performed in the human thalamus during surgery for parkinsonism and other movement disorders [23-33]. There is now a growing literature on microelectrode recording in the human GPi [34-41] and the human STN . The most common technique is recording of single-unit, extracellular action potentials using high impedence (0.1-1.0 Mohm at 1000 Hz) tungsten or platinum-iridium microelectrodes [28,36,41,43].
The utility of microelectrode recording for target localization is based on several principles [28,36,41,43]. Transitions between gray and white matter may be identified, as extracellularly recorded action potentials in gray and white matter have distinguishable waveforms. Different basal ganglia nuclei have characteristic patterns of spontaneous discharge, which are relatively easy to identify. Some of these patterns are shown in Figure 1. Motor subterritories of a region can be distinguished from nonmotor regions by finding neurons whose discharge frequencies are modulated by movement. Localization within a motor region can be accomplished by mapping the receptive field of a movement-sensitive cell during motor examination of the patient, then comparing the cell's receptive field with the known so-matotopic organization of the nucleus [36,41]. Microstimulation, or passing current through the microelectrode, can evoke motor and sensory phenomena and thus localize motor and sensory pathways. Finally, the spatial resolution of microelectrode techniques is high; structural boundaries may be identified with submillimetric precision [28,36,41,43].
"Semimicroelectrode" recording refers to the use of slightly larger-tipped, lower impedance electrodes that record from a small group of cells but are not fine enough to resolve individual action potentials. This technique may permit accurate identification of nuclear borders and movement-responsive cell regions , but does not allow detailed resolution of the discharge rate and pattern, which is especially useful for globus pallidus localization [36,41].
Hardware for MER normally provides the following functions: amplification, filtering, visual display and audio monitoring of the microelectrode signal,
microstimulation, and impedance monitoring. Optional additional functions include data recording and storage, and on-line or off-line data analysis. Systems approved by the Food and Drug Administration (FDA) for micro-electrode recording in the human are commercially available . Alternatively, it is possible to put together one's own recording rack from individual components. This is less expensive than purchasing a commercial system, but is much more labor intensive .
In addition to good electronics, single-unit recordings require high quality microelectrodes and a low-noise environment. We mainly use platinum-coated tungsten microelectrodes, with 15 to 25 micron tip diameter and impedance 0.2 to 0.6 mft at 1 KHz . Similar microelectrodes for extracellular single unit recording are commercially available from several sources, including Frederick Haer (Brunswick, Maine) and Microprobe, Inc. (Gaithersburg, Maryland). To reduce electrical noise at line frequency, it is important to place the headstage of the preamplifier close to the microelec-trode itself, which usually involves mounting it on the stereotactic arc . Alternatively, an actively shielded cable between electrode and amplifier may be used . The best ground is usually the microelectrode guide tube or the stereotactic arc. Finally, it may be necessary to unplug electrical devices near the MER apparatus, such as an electrically powered operating room table or coagulation units.
On a frontal approach to the pallidum, the microelectrode usually encounters striatum, then the external globus pallidus (GPe), prior to GPi. The majority of striatal neurons have very low (0-10 Hz) spontaneous discharge rates (Fig. 1). Neuronal discharge patterns in GPe and GPi have characteristic discharge patterns that are specific to the disease state. Most published physiological data are for Parkinson's disease (PD) (Fig. 1) [34-41,43], but some data are available for dystonia and hemiballismus [46-48]. In PD, GPe neurons have spontaneous discharge rates of 30 to 60 Hz and typically discharge in "bursting" or "pausing" patterns. Globus pallidus internal neurons in PD are faster than GPe cells, with spontaneous discharge rates of 60 to 100 Hz. In dystonia and hemiballismus, GPi discharge rates may be much lower [46-48]. Cells with very regular discharge at 20 to 40 Hz, so-called "border" cells, are typically found in the white matter laminae surrounding GPe and GPi . The optic tract (OT) can be identified by light-evoked fiber activity below the inferior margin of GPi [36,41].
Cells in GPi that are responsive to joint movements usually respond to movement of one or a small number of joints in a restricted region on the contralateral side of the body [36,37,41]. As the motor territory of GPi is somatotopically organized, with leg representations tending to be more dorsal and more medial than arm representations [41,50], the recorded distribution of neuronal receptive fields helps to determine the mediolateral coordinate of an electrode track. In patients with tremor, cells with discharges grouped at tremor frequency are often recorded [35,39]. Micro-
stimulation can be used to localize both the corticospinal tract (CST) and the OT [36,41]. The CST in the internal capsule is identified by evoking muscle contractions (usually of the tongue, face, or hand) at low current thresholds, such as 10 microamperes at 300 Hz, 200 microsecond pulse width. The optic tract is identified when the patient reports visual phenomena (focal scintillating scotomata) at low current thresholds.
3.4 Microelectrode Localization of the Motor Thalamus
During thalamic surgery, MER is especially useful to identify cells that respond to active or passive movements and to identify sensory thalamus either by microstimulation or by recording cells with cutaneous receptive fields. Lenz et al detected 107 movement-related cells out of a total 1012 cells recorded along electrode trajectories that included both motor and sensory thalamus . These cells usually respond to movements of one or a small number of contralateral joints [30,51]. They are further classified as passive or active cells according to whether their discharge frequencies are modulated by passive or active movements . In the subhuman primate, active cells are more likely to be found in the pallidal receiving area (ventralis oralis anterior and posterior, or Voa and Vop, in the Hassler terminology), whereas passive cells are more likely to be found in the cerebellar receiving area (Vim) . This may be true in human PD patients but has not been clearly confirmed .
As the microelectrode descends toward the thalamic target, the caudate nucleus may be encountered first. The next structure traversed, the dorsal thalamus, has relatively low spontaneous activity in the awake patient, but may show occasional bursts. Microelectrode entry into motor thalamus is often heralded by an increase in spontaneous activity, and can be confirmed by the identification of movement-responsive cells and/or cells discharging at tremor frequency. The posterior border of motor thalamus, formed by the lemniscal receiving area of sensory thalamus, contains neurons whose sensory receptive fields are extremely well localized . This thin anterior rim of the sensory ventrocaudal nucleus (Vcae) contains cells responding to deep muscle pressure. The remainder of the ventrocaudal nucleus (Vcpe) responds to extremely light cutaneous stimuli . Recording cutaneous sensory cells provides a precise, reproducible demarcation of the posterior edge of motor thalamus. The mediolateral position of a microelectrode track can be estimated by the somatotopic organization of motor and sensory cells, since face/jaw, arm and leg representations are organized along a medial to lateral axis. This is true of both cerebellar (Vim) and pallidal (Voa/Vop) receiving areas , as well as for the sensory nucleus Vc .
Microstimulation can identify the posterior and lateral borders of Vim. Laterally, low stimulation thresholds for evoked muscle contractions indicate that the electrode is at or has traversed the border between Vim and internal capsule. Posteriorly, low thresholds for evoked paresthesias indicate that the electrode has traversed the border between Vim and Vc .
Single cell discharge characteristics in the human STN have been studied in the parkinsonian state , and are similar to those observed in the parkinsonian monkey . A representative trace is shown in Figure 1. Since cell density in STN is extremely high, background noise is high and individual cells are hard to isolate [42,53]. Single neurons discharge at 20-50 Hz . However, typical recordings are of multiple cells and therefore the apparent discharge frequency may be higher unless great care is taken to isolate individual cells. As the microelectrode passes through the inferior border of STN into SNr, the discharge pattern changes abruptly. Background noise diminishes greatly, single neurons again become easy to isolate, and the discharge rate is high (50-70) (Fig. 1) . Microstimulation up to 100 microamperes may occasionally evoke corticobulbar responses, paresthesia, tremor arrest, or ocular deviation, but these effects are reported to be inconsistent .
3.6 Is MER Necessary?
The role of microelectrode recording during surgery for movement disorders, and particularly for pallidotomy, is actively debated. Some centers argue that it is essential [41,50,54], whereas others are adamant that it is not . Rates of permanent morbidity have been similar, and relatively low, for both groups. Many reports that have documented clear benefits from pallidotomy, as measured by standard rating scales, come from centers that use micro-electrode recording [56-61]. In some recent well-documented series, however, pallidotomy without microelectrode recording yielded similar short-term results (up to 1 year) as centers that perform pallidotomy with microelectrode recording [14,15,62,63]. In one of these series, longer term follow-up produced a diminution in benefit , to a greater extent than the long-term studies performed by the MER groups . Thus, when used for pallidotomy, MER may permit more complete lesioning of the motor territory of the target, which may be associated with more durable benefit.
Was this article helpful?