Deep Brain Stimulation for Parkinson Disease

Brief Historical Overview

Descriptions of PD symptoms and treatment with the dopaminergic extract “cowage,” from the Mucuna pruriens seed, appear in the Ayurveda, an ancient medical text of the Indian subcontinent from 1000 bce . ^, The Nei Jing, a 2500-year-old medical text from China, echoes similar ideas, yet the first western description of the disorder emerged only two centuries ago when a Londoner, James Parkinson, described six patients with “paralysis agitans.” Sixty years later, Jean-Martin Charcot refined the earlier description and coined the term “maladie de Parkinson.” Anticholinergic treatment was used as early as the mid-19th century. Arvid Carlsson and George Cotzias introduced oral levodopa/carbidopa as the “gold standard” of medical therapy in 1968.

Surgical treatment of PD began in the 1940s with resection of premotor and motor cortices in hopes of alleviating parkinsonian tremor. The resultant improvement had to be weighed against significant iatrogenic motor deficits, while no effect was seen on either rigidity or bradykinesia. The move toward basal ganglia and thalamic targets addressed these shortcomings. Interruption of the pallidofugal fibers exiting the GPi by Spiegel and Wycis ^, and by Meyers improved both tremor and rigidity. In 1952, Irving Cooper’s fortuitous sacrifice of the anterior choroidal artery in a 39-year-old man incapacitated by tremor and bradykinesia alleviated the patient’s symptoms without motor or sensory deficits.

Lesioning work on the basal ganglia and thalamus largely ceased in the 1970s because of the dramatic immediate effects of oral levodopa/carbidopa and the significant risks of surgery. By the 1990s, however, recognition of the long-term side effects of the medication—dyskinesias and motor fluctuations—brought about a renewed interest in surgical solutions. ^,

At the same time, the integration of modern imaging techniques such as CT in the late 1970s and MRI with frame-based stereotaxy in the 1980s improved the safety and accuracy of surgery at deep brain targets. Attempts to use DBS in thalamic, basal ganglia, and cerebellar regions for movement disorders were made in the early 1980s. ^,

In 1987, Benabid and colleagues showed that high-frequency stimulation could mimic a lesion in a controllable, reversible manner. Yet stimulation of the ventralis intermedius thalamic nucleus target again aided only the tremor symptomatology, leaving rigidity and bradykinesia untreated. Series of safe and effective pallidotomies from the early 1990s resurrected the concept of GPi lesioning for those symptoms, yet were technically difficult procedures limited to one hemisphere. GPi-DBS was introduced in 1994 as a safer reversible alternative with an ability to implant and modulate both hemispheres for bilateral and axial symptomatology.

In 1990, Bergman and associates showed in a nonhuman primate model of PD that the induction of parkinsonism is associated with excessive and abnormally patterned discharge in the STN, and that ablation of the nucleus alleviated all parkinsonian motor signs. Based on this work, Limousin and coworkers implanted the first chronic subthalamic stimulator for PD in the early 1990s and subsequently documented alleviation of all cardinal motor signs of PD in a case series in 1998.

Anatomy and Physiology of Targets

Both the STN and GPi are components of the basal ganglia, a collection of subcortical nuclei involved in scaling and focusing of movement, as well as motor learning. The basal ganglia also include the striatum (caudate and putamen), globus pallidus externa (GPe), and substantia nigra, subdivided into the substantia nigra pars reticularis and substantia nigra pars compacta. DBS for treatment of PD motor symptomatology is based on the “segregated circuit hypothesis.” The numerous functions of the basal ganglia within the cortex–basal ganglia–thalamus loop (motor, oculomotor, associative, and limbic) run in parallel and occupy anatomically distinct areas of the nuclei. It is therefore in principle possible to target the motor areas without compromising nonmotor functions.

The STN, shaped like a small, thick biconvex lens, is located medial to the internal capsule, lateral to the red nucleus, superior to the substantia nigra, and inferior to the thalamus. The GPi is located medial and inferior to the medial medullary lamina and GPe. The GPi overlies the choroidal fissure and the optic tract. Medially, it is limited by the genu and the lateral aspect of the posterior limb of the internal capsule.

In the basal ganglia–thalamocortical circuit ( Fig. 111.1 ), the major input structures to the basal ganglia are the striatum and STN, and the GPi serves as the main output. The input and the output nuclei are connected by a direct massive GABAergic striatopallidal pathway, and by an indirect route via the GPe and STN before arriving at the output nucleus, the GPi. The balance between the activating direct and the inhibitory indirect pathways controls movement. In the “rate model” developed originally by Albin, Young, and DeLong, the parkinsonian state is modeled as a hyperexcitation of the STN causing an imbalance in favor of the indirect pathway and excessive GPi excitation. ^, Targeting the STN or GPi with lesioning or stimulation was thought to correct excessive and abnormally patterned basal ganglia output. Some elements of the rate model have been confirmed experimentally. Elevated GPi and STN discharge rates in PD, compared with nonparkinsonian conditions, were seen by Starr and colleagues, Schrock and associates, and Steigerwald and colleagues. Optogenetic confirmation in rodents of the prokinetic and antikinetic functions of the direct and indirect pathways, respectively, was shown by Kravitz and coworkers in 2010. The model does not, however, explain the benefit of pallidal stimulation in hyperkinetic disorders. Its initial assumption that STN-DBS suppressed downstream basal ganglia activity has also been called into question. ^, The rate model remains important because of its heuristic value, as newer theories aim to rectify its shortcomings.

The most influential contemporary model of brain network dysfunction in PD is the “oscillatory synchrony model,” which posits that akinesia and bradykinesia in PD are due to excessive neuronal synchronization in specific frequency bands, especially the beta band. The beta band (13–35 Hz) is present in primary motor (M1) and sensory cortices during postural maintenance but is superseded by the high-frequency band (76–100 Hz) with planning and execution of movement. The basal ganglia show an increase in neuronal synchronization in the beta activity in PD, in the form of abnormally long “bursts” of beta activity. DBS at both the STN and GPi may exert its therapeutic effect by desynchronization of neuronal activity at specific frequencies.

Patient Selection

The most important part of DBS for PD is patient selection ( Table 111.1 ). The key to proper patient selection is a combined approach that includes evaluation by specialists in movement disorders, neurology, neurosurgery, and neuropsychology. An interdisciplinary clinic in which these specialties provide integrated care streamlines the process of patient selection and work-up.

Target Selection Factors