Deep Brain Stimulation
Deep Brain Stimulation
The current understanding of PD pathophysiology centers around abnormal β band oscillations (13–30 Hz) in the basal ganglia–cortical loop. These pathological oscillations are suppressed by movement, dopaminergic medications, and DBS and are believed to be closely related to the bradykinesia characteristic of PD.
The antikinetic nature of β oscillations has led to investigations of how they affect the relationship between the STN and primary motor cortex. An animal model of the therapeutic effects of DBS using optogenetics technology has further supported the hypothesis that high-frequency stimulation affects this relationship. Importantly, high-frequency stimulation to primary motor (M1) afferents in the STN decreased bradykinesia, while stimulation in the β range exacerbated symptomatology. However, the mechanism by which β synchrony interferes with voluntary movement continues to be an area of intense study.
Local field potential recordings of M1 in patients undergoing DBS for PD suggest increased phase-amplitude coupling of M1 β-phase (13–30 Hz) and γ-amplitude (50–200 Hz) in PD patients. Moreover, phase-amplitude coupling between M1 and STN revealed M1 LFP γ-power peaks occurring at a specific phase of the STN β rhythm in PD at a much higher magnitude than that of the STN β–M1 β coherence. This M1 β phase-coupled M1 broadband γ activity actually precedes STN β troughs, suggesting the existence of a feedback loop between the structures. It appears that pathological M1 broadband γ activity may be an important driver in maintaining aberrant STN oscillations. In turn, excessively synchronized STN and GPi β oscillations reinforce the pathological cortical β-phase and broadband γ-amplitude coupling. Another publication by the same group showed that epochs of M1 phase-amplitude coupling predicted STN spikes. This theory contrasts with older literature emphasizing the importance of intrastriatal β-synchrony as the driver of pathological oscillations.
Oscillatory activity in the motor cortex is now also being studied with magnetoencephalography as a possible biomarker for PD. The planning, execution, and termination of movement are known to be associated with consistent within-subject patterns of M1, primary sensory, and supplementary motor area oscillatory activity. Movement is preceded by a strong β desynchronization, beginning 600 msec prior to movement and lasting roughly 400 msec after the onset of movement. After this initial desynchronization, there is a strong β resynchronization called the postmovement β rebound that begins 500–800 msec after initiation of movement and lasts for 1000 msec. A brief period (100–200 msec) of increased γ band activity is also associated with movement onset. Beta desynchronization is believed to be associated with movement selection, and therefore excess β synchrony may underlie difficulty with movement initiation. In addition to excess β, PD patients were found to have diminished γ response amplitude and peak frequency.
Taken together, these data fit into the model proposed by Shimamoto and colleagues in which excess motor cortical β synchrony, manifesting clinically as hypokinesia, is a result of strong pathological β oscillations passed from the basal ganglia. This increased cortical β synchronization, in turn, leads to reinforcement of the basal ganglia β oscillations through pathological M1 β-phase γ-amplitude coupling (Fig. 1). This aberrant coupling decreases the cortex's capacity for activation-related γ activity, leading to difficulty initiating movement. Subthalamic nucleus DBS may have its effect on β oscillations and therefore movement initiation by altering the timing of M1 firing via orthodromic stimulation of afferents, limiting aberrant phase-amplitude coupling.
(Enlarge Image)
Figure 1.
Pathological phase-amplitude coupling in PD creates a self-reinforcing loop. 1: Motor cortex (M1) β-phase oscillations drive M1 γ-amplitude changes reflected by intracortical β-phase γ-amplitude coupling. 2: Changing M1 γ amplitude drives/reinforces STN β-phase oscillations via the glutamatergic hyperdirect pathway. 3: Beta-phase oscillations propagate throughout basal ganglia via glutamatergic STN-to-GPe, STN-to-GPi, and STN-to-substantia nigra pars reticulata neurons. 4: Beta-phase oscillations in the basal ganglia reinforce β-phase oscillations in M1. Reinforced β-phase oscillations in M1 prevent M1 β desynchronization necessary to initiate movement, leading to bradykinesia. M1 β-phase–M1 γ-amplitude coupling may also prevent the normal increase in γ band activity associated with initiation of movement.
The GPi remains a common target for stimulation, although the mechanism of action of GPi DBS is still debated. Cleary and colleagues found that therapeutic GPi stimulation reduced mean firing rate and increased firing regularity of local neurons during electrical stimulation, importantly decreasing burst firing for a short period of time after firing. Because stimulation of both the GPi and STN increase the regularity of thalamic neuronal firing, as well as create complex "entrained" firing patterns in local GPi neurons, it is likely that stimulation of the two regions has a similar mechanism of action. Alternative models of GPi stimulation suggest therapeutic benefit derives from stimulation of adjacent axonal projections, such as the medial medullary lamina (bradykinesia) and the internal capsule (rigidity).
Deep brain stimulation is a well-accepted approach to managing PD in patients with inadequate control of symptoms or with significant side effects from levodopa. Class 1 evidence supports the use of STN DBS when compared with best medical therapy, and in trials comparing the stimulation-on state versus the stimulation-off state. However, several aspects of this accepted standard are in flux. Stimulation of the GPi has achieved wide acceptance after it was found to cause less decline in visuomotor function and decreased depression while maintaining equivalent primary outcome compared with STN stimulation, although the latter allowed greater reduction in medication dose.
In addition to the STN and GPi, several other nuclei are accepted or under investigation for stimulation. The nucleus ventralis intermedius (VIM) of the thalamus is a standard target for alleviating tremor in PD. The pedunculopontine tegmental nucleus is a target for gait disorder and sleep modulation, sometimes in tandem with stimulation of other nuclei. Other targets in early stages of exploration include the posterior subthalamic area, caudal zona incerta, prelemniscal radiation, thalamic centromedian-parafascicular complex, and cerebral cortex. As the currently approved targets only address motor symptoms of PD, more work is needed to identify the appropriateness of DBS for nonmotor PD symptoms.
The cognitive or nonmotor effects of PD are not as well defined as the motor effects. Motor effects are more commonly associated with presentation and disease burden, as they occur early in the course of the disease when the patient is in the most active and productive years of life. Cognitive decline is observed in advanced PD, a time during which DBS has historically been offered to the patient. However, the deleterious effect of compounding the natural progression of cognitive changes with the effects of DBS may outweigh DBS-derived motor improvement.
Initial long-term studies suggested an absence of significant change in cognition 5 years after STN DBS, suggesting the promise of the technology's neuroprotective effects. However, other early studies comparing STN and GPi DBS targets reveal increased adverse cognitive and behavioral effects after STN DBS. Speculation as to the potential cause of cognitive decline in early versus more recent studies may stem from the close anatomical apposition of motor, associative, and limbic pathways in the STN. As targeting techniques have improved, side effects of stimulation of these nonmotor pathways may have decreased. Definitive conclusions may also have been elusive due to small sample size and the study design. Woods and colleagues evaluated 30 studies investigating cognitive changes after DBS and identified only 2 that had sufficient statistical power on which to base conclusions. Another meta-analysis found STN DBS to be relatively safe from a cognitive standpoint, except for a measurable decline in verbal fluency.
Recent investigations in the US have corroborated the persistent decline in verbal fluency in the STN cohort, as well as worsened dementia rating scores. However, a European randomized controlled study evaluating the effects of STN versus GPi DBS in 128 patients with PD found no significant difference in cognitive side effects (a composite of multiple factors such as depression, anxiety, psychosis) in either group. In fact, the authors recommended STN DBS due to superior overall outcomes of secondary investigative endpoints.
Although DBS has traditionally been reserved for PD patients with intractable symptoms, dyskinesias, or severe levodopa side effects, a recent study in patients with early motor symptoms of PD showed promising results. This randomized prospective trial compared DBS combined with medication against medication alone in patients with early motor signs of PD (average duration of disease of 7.5 years). The primary outcome, quality of life (assessed using the Parkinson Disease Questionnaire-39), improved by 7.8 points in patients receiving a combination of DBS and medication, compared with a decrease of 0.2 points in patients receiving medication only. Patients who underwent surgery also experienced improved secondary outcomes, including decreased motor disability, improvement in performing activities of daily living, and fewer levodopa side effects. There was also an average of 1.9 hours/day increase in time with good movement and no dyskinesia, along with an average of 1.8 hours/day decrease in poor mobility time. Although patients in the stimulation group had slightly higher rates of mild adverse events, the authors argued that neurostimulation can and should be used to optimize treatment early in PD, before significant disabling motor and cognitive symptoms arise. It is also likely that performing surgery in patients who are younger and likely healthier will afford better surgical outcomes and a decreased risk of operative morbidity and death.
Other future directions of DBS for PD include tailoring the selection of nuclei to the individual's exact symptomatology, although target selection remains an area of debate. Different modes of stimulation are also being attempted, including constant stimulation and interleaved stimulation.
DBS in Parkinson Disease
Mechanistic Understanding
The current understanding of PD pathophysiology centers around abnormal β band oscillations (13–30 Hz) in the basal ganglia–cortical loop. These pathological oscillations are suppressed by movement, dopaminergic medications, and DBS and are believed to be closely related to the bradykinesia characteristic of PD.
The antikinetic nature of β oscillations has led to investigations of how they affect the relationship between the STN and primary motor cortex. An animal model of the therapeutic effects of DBS using optogenetics technology has further supported the hypothesis that high-frequency stimulation affects this relationship. Importantly, high-frequency stimulation to primary motor (M1) afferents in the STN decreased bradykinesia, while stimulation in the β range exacerbated symptomatology. However, the mechanism by which β synchrony interferes with voluntary movement continues to be an area of intense study.
Local field potential recordings of M1 in patients undergoing DBS for PD suggest increased phase-amplitude coupling of M1 β-phase (13–30 Hz) and γ-amplitude (50–200 Hz) in PD patients. Moreover, phase-amplitude coupling between M1 and STN revealed M1 LFP γ-power peaks occurring at a specific phase of the STN β rhythm in PD at a much higher magnitude than that of the STN β–M1 β coherence. This M1 β phase-coupled M1 broadband γ activity actually precedes STN β troughs, suggesting the existence of a feedback loop between the structures. It appears that pathological M1 broadband γ activity may be an important driver in maintaining aberrant STN oscillations. In turn, excessively synchronized STN and GPi β oscillations reinforce the pathological cortical β-phase and broadband γ-amplitude coupling. Another publication by the same group showed that epochs of M1 phase-amplitude coupling predicted STN spikes. This theory contrasts with older literature emphasizing the importance of intrastriatal β-synchrony as the driver of pathological oscillations.
Oscillatory activity in the motor cortex is now also being studied with magnetoencephalography as a possible biomarker for PD. The planning, execution, and termination of movement are known to be associated with consistent within-subject patterns of M1, primary sensory, and supplementary motor area oscillatory activity. Movement is preceded by a strong β desynchronization, beginning 600 msec prior to movement and lasting roughly 400 msec after the onset of movement. After this initial desynchronization, there is a strong β resynchronization called the postmovement β rebound that begins 500–800 msec after initiation of movement and lasts for 1000 msec. A brief period (100–200 msec) of increased γ band activity is also associated with movement onset. Beta desynchronization is believed to be associated with movement selection, and therefore excess β synchrony may underlie difficulty with movement initiation. In addition to excess β, PD patients were found to have diminished γ response amplitude and peak frequency.
Taken together, these data fit into the model proposed by Shimamoto and colleagues in which excess motor cortical β synchrony, manifesting clinically as hypokinesia, is a result of strong pathological β oscillations passed from the basal ganglia. This increased cortical β synchronization, in turn, leads to reinforcement of the basal ganglia β oscillations through pathological M1 β-phase γ-amplitude coupling (Fig. 1). This aberrant coupling decreases the cortex's capacity for activation-related γ activity, leading to difficulty initiating movement. Subthalamic nucleus DBS may have its effect on β oscillations and therefore movement initiation by altering the timing of M1 firing via orthodromic stimulation of afferents, limiting aberrant phase-amplitude coupling.
(Enlarge Image)
Figure 1.
Pathological phase-amplitude coupling in PD creates a self-reinforcing loop. 1: Motor cortex (M1) β-phase oscillations drive M1 γ-amplitude changes reflected by intracortical β-phase γ-amplitude coupling. 2: Changing M1 γ amplitude drives/reinforces STN β-phase oscillations via the glutamatergic hyperdirect pathway. 3: Beta-phase oscillations propagate throughout basal ganglia via glutamatergic STN-to-GPe, STN-to-GPi, and STN-to-substantia nigra pars reticulata neurons. 4: Beta-phase oscillations in the basal ganglia reinforce β-phase oscillations in M1. Reinforced β-phase oscillations in M1 prevent M1 β desynchronization necessary to initiate movement, leading to bradykinesia. M1 β-phase–M1 γ-amplitude coupling may also prevent the normal increase in γ band activity associated with initiation of movement.
The GPi remains a common target for stimulation, although the mechanism of action of GPi DBS is still debated. Cleary and colleagues found that therapeutic GPi stimulation reduced mean firing rate and increased firing regularity of local neurons during electrical stimulation, importantly decreasing burst firing for a short period of time after firing. Because stimulation of both the GPi and STN increase the regularity of thalamic neuronal firing, as well as create complex "entrained" firing patterns in local GPi neurons, it is likely that stimulation of the two regions has a similar mechanism of action. Alternative models of GPi stimulation suggest therapeutic benefit derives from stimulation of adjacent axonal projections, such as the medial medullary lamina (bradykinesia) and the internal capsule (rigidity).
Current Approach to Therapy
Deep brain stimulation is a well-accepted approach to managing PD in patients with inadequate control of symptoms or with significant side effects from levodopa. Class 1 evidence supports the use of STN DBS when compared with best medical therapy, and in trials comparing the stimulation-on state versus the stimulation-off state. However, several aspects of this accepted standard are in flux. Stimulation of the GPi has achieved wide acceptance after it was found to cause less decline in visuomotor function and decreased depression while maintaining equivalent primary outcome compared with STN stimulation, although the latter allowed greater reduction in medication dose.
In addition to the STN and GPi, several other nuclei are accepted or under investigation for stimulation. The nucleus ventralis intermedius (VIM) of the thalamus is a standard target for alleviating tremor in PD. The pedunculopontine tegmental nucleus is a target for gait disorder and sleep modulation, sometimes in tandem with stimulation of other nuclei. Other targets in early stages of exploration include the posterior subthalamic area, caudal zona incerta, prelemniscal radiation, thalamic centromedian-parafascicular complex, and cerebral cortex. As the currently approved targets only address motor symptoms of PD, more work is needed to identify the appropriateness of DBS for nonmotor PD symptoms.
Cognitive Effects of DBS in the PD Population
The cognitive or nonmotor effects of PD are not as well defined as the motor effects. Motor effects are more commonly associated with presentation and disease burden, as they occur early in the course of the disease when the patient is in the most active and productive years of life. Cognitive decline is observed in advanced PD, a time during which DBS has historically been offered to the patient. However, the deleterious effect of compounding the natural progression of cognitive changes with the effects of DBS may outweigh DBS-derived motor improvement.
Initial long-term studies suggested an absence of significant change in cognition 5 years after STN DBS, suggesting the promise of the technology's neuroprotective effects. However, other early studies comparing STN and GPi DBS targets reveal increased adverse cognitive and behavioral effects after STN DBS. Speculation as to the potential cause of cognitive decline in early versus more recent studies may stem from the close anatomical apposition of motor, associative, and limbic pathways in the STN. As targeting techniques have improved, side effects of stimulation of these nonmotor pathways may have decreased. Definitive conclusions may also have been elusive due to small sample size and the study design. Woods and colleagues evaluated 30 studies investigating cognitive changes after DBS and identified only 2 that had sufficient statistical power on which to base conclusions. Another meta-analysis found STN DBS to be relatively safe from a cognitive standpoint, except for a measurable decline in verbal fluency.
Recent investigations in the US have corroborated the persistent decline in verbal fluency in the STN cohort, as well as worsened dementia rating scores. However, a European randomized controlled study evaluating the effects of STN versus GPi DBS in 128 patients with PD found no significant difference in cognitive side effects (a composite of multiple factors such as depression, anxiety, psychosis) in either group. In fact, the authors recommended STN DBS due to superior overall outcomes of secondary investigative endpoints.
Areas of Evolving Practice
Although DBS has traditionally been reserved for PD patients with intractable symptoms, dyskinesias, or severe levodopa side effects, a recent study in patients with early motor symptoms of PD showed promising results. This randomized prospective trial compared DBS combined with medication against medication alone in patients with early motor signs of PD (average duration of disease of 7.5 years). The primary outcome, quality of life (assessed using the Parkinson Disease Questionnaire-39), improved by 7.8 points in patients receiving a combination of DBS and medication, compared with a decrease of 0.2 points in patients receiving medication only. Patients who underwent surgery also experienced improved secondary outcomes, including decreased motor disability, improvement in performing activities of daily living, and fewer levodopa side effects. There was also an average of 1.9 hours/day increase in time with good movement and no dyskinesia, along with an average of 1.8 hours/day decrease in poor mobility time. Although patients in the stimulation group had slightly higher rates of mild adverse events, the authors argued that neurostimulation can and should be used to optimize treatment early in PD, before significant disabling motor and cognitive symptoms arise. It is also likely that performing surgery in patients who are younger and likely healthier will afford better surgical outcomes and a decreased risk of operative morbidity and death.
Other future directions of DBS for PD include tailoring the selection of nuclei to the individual's exact symptomatology, although target selection remains an area of debate. Different modes of stimulation are also being attempted, including constant stimulation and interleaved stimulation.
Source...