CHAPTER 78 Surgical Management of Tremor
Spiegel and Wycis1–3 pioneered stereotactic surgery for the treatment of Parkinson’s disease (PD) by introducing coagulation of the globus pallidus—the stereotactic pallidotomy. Hassler4,5 subsequently introduced ablation of the ventral oral nucleus of the thalamus (consisting of the anterior [VOA] and posterior [VOP] subnuclei)—the terminus for pallidal afferents to the thalamus—to treat movement disorders. Microelectrode recordings later demonstrated that the area posterior to VOP—that is, the terminus of cerebellar afferents (the ventral intermediate [VIM] nucleus)—has rhythmic bursting activity close to the frequency of tremor.6 The VIM nucleus then became the lesioning target of choice for the treatment of all types of tremor.
DBS was first applied in the hypothalamus for the treatment of chronic pain by Pool and colleagues7 in the 1950s. Thereafter, DBS was applied primarily to the somatosensory thalamus and the paraventricular gray matter for the treatment of chronic pain.8,9 Surgeons noted, however, that the stimulation used to localize thalamic targets to be lesioned for pain relief also improved the symptoms of movement disorders.5,10 In addition, thalamic stimulation for the relief of chronic pain improved movement disorders associated with thalamic pain syndrome.11 In the 1980s and 1990s, stimulation was reported as an effective treatment for tremor12,13 and as a specific therapeutic modality for PD.14,15
The introduction of levodopa in the early 1970s led to a dramatic decrease in the number of stereotactic surgeries for movement disorders.16 Two decades later, side effects of drug treatment, including dyskinesias and fluctuations, had complicated the long-term care of PD patients and could become extremely disabling on their own.17,18 There was renewed interest in lesional surgical approaches to movement disorders after neurophysiologic studies in the realistic 1-methyl, 4-phenyl, 1,2,3,6-tetrahydropyridine (MPTP) model of PD demonstrated hyperactivity in the subthalamic nucleus (STN) and internal globus pallidus (GPi) that could be reversed by lesioning.19 This was demonstrated clinically in humans with advanced PD via a posteroventral GPi pallidotomy.20 Finally, the introduction of stimulation techniques led to a decreased role for ablative approaches in the treatment of movement disorders.13,21,22
Pathophysiology: Tremor Generation
Parkinson’s Tremor
In the case of Parkinson’s tremor, the most current hypothesis is that a thalamic oscillator is activated by hyperpolarization of the thalamocortical cells in ventral lateral (pars oralis)/VOP owing to increased inhibitory input from the pallidum in PD.19,23–27 When inhibited, these cells generate a somatic calcium spike that produces an action potential burst (low-threshold spike burst) followed by inhibition, leading to recurrent oscillations.23,28,29 However, such bursts are rarely found in recordings from the VIM nucleus and VOP of parkinsonian patients undergoing thalamotomy.30,31
Another proposed central generator is the GPi—specifically, neurons within the GPi that project to and inhibit VOP neurons. This hypothesis is inconsistent with higher rates of tremor-related activity in the cerebellar relay nucleus (VIM nucleus, 25%) compared with the pallidal relay in patients with PD (VOP, 21%).32,33 Furthermore, GPi activity at tremor frequency is rarely correlated with tremor in PD patients34,35 or in monkey models of parkinsonism.36 Based on these reports, the activity of cells in the thalamus and pallidum is not consistent with a pallidal generator for parkinsonian tremor.
Another proposed mechanism of parkinsonian tremor is peripheral feedback. This hypothesis proposes that tremor is the oscillation of unstable stretch reflex arcs (long-loop reflex arcs) that may traverse the motor cortex much as tendon tap reflexes traverse the spinal cord.37–39 The increased gain of these reflexes may cause parkinsonian rigidity and tremor, just as increased spinal reflexes cause spasticity and clonus.27 This hypothesis is supported by the finding that thalamic neuronal activity precedes tremor in thalamic neurons with sensory inputs but not in those without.33,40–42 Therefore, sensory cells participating in a reflex or feedback loop might cause tremor. Finally, a transfer function analysis has demonstrated a feedback loop in more than 90% of cells in the VIM nucleus and VOP.43
Essential Tremor
One particular animal model of tremor that reproduces many aspects of human essential tremor is the harmaline-induced tremor.44,45 Recordings in harmaline-treated cats indicate the importance of a cerebellobulbospinal pathway in the maintenance of this tremor.45 However, in humans with essential tremor there is clear thalamic activity that is directly correlated with tremor frequency46 (Fig. 78-1). Specifically, the VIM nucleus (which is the cerebellar relay nucleus) demonstrates the highest proportion of tremor-related neurons in intraoperative recording studies.47 These neurons become more active during voluntary movements, indicating that essential tremor may be facilitated by voluntary motor circuits that enable tremor-related thalamic activity. Additionally, many of these cells respond to proprioceptive input, meaning that sensory feedback from the periphery might also influence this central tremor circuitry.47
Surgical Approach
Targeting by Magnetic Resonance Imaging
Preoperative planning and target localization currently rely on MRI using one of the stereotactic frame-based systems. The MRI sequences used for electrode placement include T1, T2 fast spin echo, three-dimensional gradient echo, and axial inversion recovery images, which are particularly useful for determining target location relative to the midcommissural point of the anterior commissure (AC)–posterior commissure (PC) line or the PC itself.48–50 Target localization by MRI is not error free, and this can have serious clinical implications because of the small size of the target. The STN, for example, has an average dimension of 3 × 6 × 4 mm.51,52
Errors in electrode placement caused by systematic MRI errors have been measured against plain radiographs, which have less distortion. In a recent report on DBS placement accuracy using MRI target acquisition with the Leksell frame, the location of previous electrodes was examined during revision surgeries, and the error was measured using fluoroscopically derived images.53 The latter images of the old electrode position were taken before these electrodes were removed and compared with MRI-derived coordinates taken on the morning of the revision surgery. This study revealed that errors along the superior-inferior axis were generally larger than errors along the left-right and anterior-posterior axes, although all coordinate discrepancies were less than 1 mm. Larger errors were found in another postoperative review of 27 bilateral STN electrode placements.54 The postoperative MRI-derived coordinates of the tip of the electrode differed from the implantation coordinates, on average, by 0.48 mm, 0.69 mm, and 2.9 mm along the left-right, anterior-posterior, and superior-inferior axes, respectively.
Inhomogeneities within the primary gradient field can produce errors in MRI targeting.55 Artifacts can be induced by metal or magnetic susceptibility effects produced at the interface between materials (e.g., air and bone) that have different tendencies to alter the magnetic field in a region. Attempts to decrease errors in the MRI scan caused by these artifacts include software modifications and overlapping (fusion) of the three-dimensional MRI database with a computed tomography database that is not prone to these types of distortions.56 These databases can then be merged with atlas maps of anatomy to estimate the location of nuclei relative to the radiologic images. These atlas maps can be transformed to match either the AC-PC line in isolation or the AC-PC line along with other structures, such as the margins of the third ventricle or the internal capsule.57,58 Another approach is to estimate the target directly from the AC-PC line as determined radiologically. The target for the VIM nucleus used in our clinical practice is 3 mm anterior and 14 mm lateral to the PC and 2 mm above the AC-PC line. Alternatively, the target in the VIM nucleus can be estimated in the lateral plane midway between the internal capsule and the lateral edge of the T2-intense medial dorsal nucleus of the thalamus.59
Intraoperative Localization
Microelectrode recording as an adjunct to MRI-derived anatomic DBS placement plays an important role in improving the accuracy of targeting in stereotactic placement procedures.48,60–62 Useful features of microelectrode recording include differentiation of gray and white matter locations, differentiation of nuclei or subnuclei on the basis of intrinsic neuronal firing properties, localization of white matter tracts with particular responses to stimulation, and real-time correction for intraoperative shifts in implantation sites.49 The neuronal firing patterns detected in the various nuclei and white matter tracts for placement in the STN and VIM nucleus have been extensively reviewed.49,60–64 The microelectrodes typically consist of glass-insulated tungsten or a platinum-iridium alloy with an exposed tip measured in micrometers.48,60 During recording and stimulation, clinicians look for somatosensory-induced effects (e.g., receptive field–evoked responses, movements about joints producing signals, deep dermal receptive fields) as well as stimulation effects on the patient’s tremor or tone as an aid to localization.61–63 Microelectrodes are also used for stimulation during localization passes to map side effects, and some commercially available microelectrodes allow for macrostimulation at currents up to 2 mA. Although microelectrode recording might optimize the implant position, it does not necessarily help in choosing which of the DBS electrode elements should be used for chronic stimulation. In a study of 20 patients with PD undergoing DBS in the STN, there was no correlation between the mapped STN region and the most effective treatment electrode at 3- and 6-month follow-up.65 This may be related to the complexity of the fields produced by stimulation in the brain.66
In our institution, a map of physiologic results is made to the same scale as a set of transparent atlas maps from the sagittal sections of the Schaltenbrand and Bailey atlas.67 Parasagittal sections for different targets are as follows: 13.5-mm lateral atlas map for the VIM nucleus and 11 mm for the STN.67 The fusion and fitting of the imaging studies to the physiologic and atlas maps can also be accomplished through a number of commercially available surgical navigation systems. Microelectrodes for physiologic monitoring and recording are designed to isolate single action potentials60,68,69 and to withstand microstimulation, which degrades the electrode. Typically these characteristics are achieved by constructing electrodes from a platinum-iridium alloy or from tungsten, producing a tapered tip, and insulating with glass.60,69–73 The electrode impedance is usually greater than 500 kOhm,60,68,74 which is required to isolate single units.60
The assembled electrode is attached to a hydraulic or piezoelectric microdrive and mounted on the stereotactic frame. Some microdrive systems incorporate a coarse drive so that overlying structures can be traversed quickly. The tip is then retracted into a protective cylindrical housing while the whole assembly is advanced to a new depth.61 The microdrive can then be used from this new depth for detailed exploration of deeper structures. Another option is to use the microdrive throughout the trajectory, as in many commercial systems (Alpha Omega Co. USA, Alpharetta, GA). The signal from the microelectrode is amplified and filtered. Multiple neuronal discharges of varying sizes can be seen on a digitized trace and heard with the use of an audio monitor. In addition to recording, microstimulation of subcortical structures can be delivered. We employ biphasic, square-wave pulse trains of 0.1- to 0.3-msec pulses for up to 10 seconds at a frequency of 300 Hz.75 The current used in the stimulation determines the amount of local current spread.76
Semi-microelectode recordings can also be carried out using low-impedance microelectrodes with an impedance of less than 100 kOhm. The semi-microelectrode signal is often amplified against a concentric ring electrode that is mounted concentrically around the microelectrode.70,77,78 Recordings through a semi-microelectrode yield recordings of local field potentials (slow waves) or multiunit activity. Bipolar stimulation through a concentric ring electrode can also be used alone or in combination with recording through a semi-microelectrode.10,79,80
Localization of the Ventral Intermediate Nucleus
Sensory cells responding to sensory stimulation in small, well-defined, receptive fields are found in the ventral caudal (VC) nucleus, posterior to the VIM nucleus.63 There is a well-described mediolateral somatotopy within the VC nucleus, proceeding from representations of oral structures medially to the leg laterally.63,68 In the anterior VC nucleus, and further anteriorly in the VIM nucleus, neuronal firing is related to passive joint movement (deep sensory cells) or to active movement (voluntary cells).42,81 Some neurons respond both to active movement and to sensory stimulation such as joint movement (combined cells).33,42 The activity of deep sensory, combined, and voluntary cells is correlated with electromyographic activity during tremor,32,60,82 and they have a somatotopy parallel to that present in the VC nucleus. Stimulation in the VC nucleus evokes somatic sensations.75 Stimulation in the VIM nucleus may produce brief movements or alter ongoing tremor or dystonia.83
Thalamic semi-microelectrode recordings reveal patterns of neuronal activity parallel to those of microelectrode recordings.70,78,84 Macrostimulation through a low-impedance electrode (often <1000 kOhm) can reliably identify the VIM nucleus, by alterations in the movement disorder,10 and the capsule, by stimulation-evoked tetanic contraction of skeletal muscle at a low threshold.10,79 Stimulation of intralaminar nuclei, medial to the VC nucleus or the VIM nucleus, may evoke the recruiting response—long-latency, high-voltage, negative waves occurring over much of the cortex at the frequency of stimulation (usually <10 Hz).85,86
An analysis of the locations of tremor cells suggests that the optimal target for thalamic stimulation is 2 mm anterior to the VC nucleus and 3 mm above the AC-PC line.87 Alternatively, targets in thalamotomy and thalamic stimulation have been placed anterior to the site at which evoked potentials can be recorded in response to cutaneous stimulation of the fingers.88–90 Lesions have been made in the region where electrical stimulation produces effects on tremor and anterior to the region where electrical stimulation evokes sensations.10 Lesions have also been made in the region where cells respond to somatosensory stimulation of muscle, joint, and tendon and where electrical stimulation effects tremor.91
Localization of the Subthalamic Nucleus
Subthalamic exploration is also made from a coronal bur hole about 30 to 35 mm from the midline. Striatum and the VOA thalamus each have a characteristic pattern of neural activity that can be observed with microelectrode recording during mapping.62 As the electrode approaches the STN from an anterior direction, the dorsal striatal cells encountered are characterized by broad action potentials and a slow firing rate (approximately 1 Hz), often with long silent periods. When the thalamus is entered, the action potentials become narrower and often occur in bursts of the low-threshold spike type. Low-threshold spike bursts are preceded by a silent period of 20 to 100 msec and consist of an initial interspike interval of less than 6 msec, followed by interspike intervals of less than 16 msec.30,92 A relatively acellular gap of 1 to 6 mm is observed below the thalamus and above the STN, depending on the anterior-posterior location of the trajectory. The STN is characterized by multiple spike trains recorded from closely packed neurons, each with a mean rate of about 20 msec. Microstimulation evokes paresthesias posterior to the STN, muscle contractions lateral to the STN, and decreases in tremor or tone within STN.
Stimulation Settings
A multidisciplinary team is required for evaluation and clinical decision making in any patient with a movement disorder who is being considered for surgery.93 This clinical team should include a movement disorder neurologist who assists in the initial evaluation of the patient and later with the programming of stimulation parameters.
