Robert S. Turner - US grants

Several studies indicate that the frontal cortical and associated basal ganglia circuitry may be involved in relatively abstract aspects of motor control rather than mechanics of movement. This proposal is based on the general hypothesis that frontal cortical and basal ganglia circuitry are specifically involved in the learning of motor sequences. The studies will employ a serial reaction time task where one portion of a sequence of movements is kept constant while the others vary randomly. Over several trials, humans demonstrate learning for the constant portion of the sequence without explicit awareness of the learning. Three specific aims are outlined. The first will determine if there are neurons in motor and mesial premotor areas that discharge for specific sequences regardless of the limb used to perform the task. The hypothesis is that there will be neurons with an abstract representation of motor sequences that do not depend upon the motor apparatus used to execute the task. The second specific aim will determine if monkeys with bilateral lesions of the GPi are specifically impaired in learning new motor sequences. It is hypothesized that pallidotomy will disrupt learning of new motor sequences but will not disrupt performance of previously learned sequences. The third specific aim will be to determine if pallidotomy will affect neural correlates of sequence learning in primary motor cortex and dorsal and mesial premotor cortex. It is hypothesized that pallidotomy will slow or eliminate the development of sequence-related neural discharge, whereas discharge related to overlearned motor sequences will remain unchanged.

DESCRIPTION (provided by applicant): Deep brain stimulation (DBS) of either the internal segment of the globus pallidus (GPO or the subthalamic nucleus (STN) is an effective treatment for most if not all symptoms of Parkinson's disease (PD). Several aspects of the reduction of symptoms with DBS provide tantalizing hints that different symptoms may be mediated by distinct pathways and/or physiological processes involving the motor and premotor cortices. The goals of this project are to use a non-human primate model of PD to gain a better understanding of the cortical mechanisms by which DBS produces clinical benefit, as well as to determine if different symptoms have different neuroanatomic/physiologic substrates. Animals will perform tasks that measure symptom-relevant behavioral parameters: movement selection/initiation/sequencing (akinesia), movement kinematics (bradykinesia), and rigidity. Neuronal activity at multiple locations in the four principal motor cortices [in different animals, primary motor (M1), ventral premotor (PMv), dorsal premotor (PMd), or mesial premotor (SMA)] will be monitored using a multielectrode array. Single cell activity will be assessed for changes in resting firing rate, task-related activity, and cell-to-cell interactions (synchronized firing) in response to DBS in GPI or STN before and after animals are rendered parkinsonian by intracarotid infusion of MPTP. The predictions are that: DBS-related changes in resting discharge will not be correlated with specific changes in symptoms. Increased activity and synchrony in SMA will be associated with reduced akinesia. Increases of the same in M1 will accompany reduced bradykinesia. Reductions in rigidity will be linked with a drop in M1 responses to passive movement and increased directional specificity in movement related activity. In addition, DBS may reduce abnormally-increased activity in PMv and PMd These hypotheses will be tested in three specific aims: Specific aim I will study the interacting effects of DBS and the type of motor task being performed. Specific aims 2 and 3 will identify cortical activities that change in concert with the time course (SA 2) and parametric relations (SA 3, DBS location, frequency, and strength) of symptom reduction with DBS. The results of these experiments will improve understanding of both the neuronal basis of different symptoms of PD and the mechanisms of action of DBS. Ultimately, these studies will advance a more complete pathophysiologic model of PD by incorporating the full array of parkinsonian symptoms.

DESCRIPTION (provided by applicant): Rational refinement and targeting of therapies for Parkinson's disease (PD) will be facilitated by an ability to analyze relationships between the pathologic features of the disease and symptoms. Neurotoxin models, however, are limited in their ability to break down the disease into its component features. This problem is exacerbated by two complications that are coming to greater recognition: 1) loss of dopamine (DA) is by no means restricted to the striatum in idiopathic PD or in neurotoxin models of the disease;and 2) individual DA neurons innervate multiple brain regions. We propose a monkey model for studying the pathologic underpinnings of parkinsonism that overcomes the hurdles presented by these complications. Reversible intracerebral blockade of DA neurotransmission will afford precise control over the location of DA loss/blockade, its spatial extent and severity, and its timecourse. It is precisely these aspects of DA dysfunction that are difficult if not impossible to control using neurotoxins. The proposed experiments will use DA antagonists to address the central, yet persistent, question whether loss of DA from the striatum alone can generate the cardinal signs of PD. Although akinesia, bradykinesia, tremor and rigidity are commonly attributed to striatal loss of DA, idiopathic PD and neurotoxin-induced parkinsonism are marked by DA loss in other structures as well. The first aim will determine if these signs can be reproduced in primates by transient antagonism of DA receptors (both D1 and D2) in the motor striatum. Convection-enhanced delivery will be used to produce a homogeneous blockade of DA receptors within significant portions of the posterior putamen (site of the most severe DA loss in idiopathic PD). Parkinsonian signs will be measured using tasks designed to manipulate relevant behavioral parameters: movement initiation/sequencing (akinesia), movement kinematics (bradykinesia), tremor, and muscle tone (rigidity). The internal globus pallidus is a critical link in the translating the DA dysfunction of PD into parkinsonian signs. The second aim will determine which abnormalities in the firing of globus pallidus neurons can be attributed directly to the loss of striatal DA and thus contribute to the impairments elicited. This aim will also correlate specific abnormalities in pallidal firing with the appearance of specific parkinsonian symptoms. PUBLIC HEALTH RELEVANCE: Despite the growing prevalence of Parkinson's disease in the population, refinement and accurate anatomical targeting of new treatments is impeded by our rudimentary understanding of the relationship between the pathology of the disease (i.e., which cells die) and its symptoms. The animal model and experiments proposed here will provide important new information about the critical pathologic defects that give rise to parkinsonian signs. Target selection for new cell and gene therapies will be guided by this information.

The fluid, seemingly effortless execution of sequences of movements is a ubiquitous feature of everyday motor skills. Ample evidence for their importance comes from the common human neuropathologies (Parkinson's disease, in particular) in which sequential skills are especially impaired. Long-term motor sequencing skills are formed, most likely, across multiple time scales in associative, premotor, and motor circuits of the brain. Recent evidence suggests that for each of these brain circuits, a sub-cortical loop through the basal ganglia (BG) contributes selectively to reinforcement-driven modulation of thalamo-cortical plasticity. These findings lead to the hypothesis that BG loops play central roles in the acquisition of sequence information, but are less important in the recall or use of already-learned sequences. The specific aims (SAs) of this proposal will test that general hypothesis by using non-human primates: 1) to determine if neurons in the globus pallidus interna (GPi, the primary BG output nucleus for skeletomotor function) preferentially encode sequence information during new learning;and 2) to test whether intact BG circuits are necessary for new sequence learning. Associative loops through the BG may play a greater role in the fast acquisition of flexible goal-directed representations of sequence information while the premotor and motor loops may mediate slow acquisition of habit-like effector-specific representations. We will infer the circuit membership of individual GPi neurons by stimulating different cortical areas and observing the orthodromic inhibitory effects. Animals will perform a discrete sequence production task using novel, familiar and over-trained sequences. SAl will test if neuronal encoding of sequence-specific information in associative, premotor, and motor circuits of GPi reflects the predicted roles of these circuits in learning novel, familiar, and over-trained sequences. SA2 will determine if an interruption of BG output (i.e., GPi inactivation or lesion) selectively impairs training-related improvements in sequence performance. The prediction is that inactivations or lesions in the associative BG circuit will impair novel sequence learning whereas lesions in premotor and motor circuits will block the further refinement and solidification of performance of already-familiar sequences. Results from these experiments will aid in understanding the physiological basis for RELEVANCE (See instructions): TThe proposed work is central to the problem of understanding the mechansims where practice leads to to reorganization of the human motor system in the face of aging, neurodeneration, stroke or brain injury. Understanding these mechansims has an impact on the design of therapies directed at preserving function, developing compensator movements and ultimately, developing novel motor capacity.

DESCRIPTION (provided by applicant): The motor signs of Parkinson's disease (PD) have been linked causally to abnormalities in the spiking activity of neurons in the globus pallidus internus (GPi), a major output nucleus of the basal ganglia (BG). Likewise, deep brain stimulation (DBS) of the subthalamic nucleus (STN) may provide relief from parkinsonian signs by suppressing abnormalities in GPi activity. Efforts to refine DBS and develop new therapeutic targets for PD will be greatly enhanced by elucidation of the specific mechanisms by which abnormal GPi activity and its alteration under DBS impact parkinsonian signs. The experiments in this proposal focus on the idea that communication between GPi and BG-recipient thalamus is a central factor in the pathophysiology of Parkinsonism and its amelioration during DBS. Two hypotheses will be tested: the information hypothesis posits that a loss of independent signaling in the parkinsonian GPi reduces the information-carrying capacity of GPi-recipient neurons in thalamus, while an alternate hypothesis posits that certain firing abnormalities in GPi (e.g., low frequency oscillations or bursts) induce pathologic activity in thalamus, which disrupts function downstream (e.g. in the motor cortices). Each hypothesis predicts that the associated measures of neuronal activity will covary with the severity of parkinsonian signs and their rectification via DBS. These predictions will be tested through an innovative interdisciplinary research plan. Neuronal activity in GPi and GPi-recipient thalamus will be studied using multi-electrode single-unit and local field potential recordings in non-human primates, before and during the slow, progressive induction of parkinsonism, and during sub-therapeutic and therapeutic DBS in the STN. Independent signaling will be quantified as spike correlations, within and between nuclei, and as the specificity of neuronal responses to proprioceptive stimulation of different limbs. Parkinsonian signs will be measured using tasks that assay movement initiation (akinesia), movement kinematics (bradykinesia), and muscle tone (rigidity). Data from these experiments will be analyzed in collaboration with investigators (Rubin and Doiron) who have substantial computational experience, including work on parkinsonian BG dynamics and the propagation of information in neuronal networks. Computational and theoretical methods will tease apart specific ways that changes in GPi output alter thalamic function. Empirical results will be incorporated into Hodgkin-Huxley type neuronal models and mean field and information theoretic analyses to determine how GPi activity influences thalamic correlations and information coding, and to predict downstream effects of changes in thalamic firing properties across normal, parkinsonian and parkinsonian+DBS conditions. Results from these studies will advance our understanding of parkinsonian pathophysiology and test potential therapeutic mechanisms of DBS, suggesting targets for future therapeutic interventions including optimization of DBS. The results may also be relevant to the whole class of clinical disorders that involve BG-thalamic dysfunction as well as to the use of DBS for other neurologic conditions.

This Core will provide investigators with all of the intellectual expertise, skilled technical support and critical resources needed for NHP surgery. A skilled, knowledgeable team will be available and equipped with the technical resources required to provide pre-op, intra-operative and post-op care in facilities specifically designed for monkeys. This Core will have two purpose-designed, state-of-the-art suites for anesthesia, surgery and post-surgical recovery from operations on the brain and spinal cord.

? DESCRIPTION (provided by applicant): The motor thalamus (VL) is densely interconnected with the primary motor cortex (M1) and, at the same time, VL serves as the main gateway by which cerebellum and the basal ganglia influence M1 function. Thus, VL lies at the heart of the vertebrate motor control system. The experiments in this proposal will advance our basic understanding of this essential but little-studied M1-VL circuit. The experiments will contrast two views of VL function: a) as a relay that transmits to M1 the information received from sub-cortical inputs, and b) as a closely interconnected partner with M1 in the dynamic generation of motor commands. Available evidence suggests that the cerebellar-recipient part of VL (VLp) approximates the simple relay view, transmitting cerebellum- derived information to M1. A great deal of uncertainty remains regarding the basal ganglia-recipient part of VL (VLa). Our provisional hypothesis is that activity in VLa is driven by M1, but sculpted or biased by the inhibitory signals received from the basal ganglia. We will use non-human primates trained on sequential arm movement and arm perturbation tasks that are predicted to differentiate the activity of neurons in VLp and VLa. The membership of individual thalamo-cortical and cortico-thalamic neurons to an arm-related M1-VL circuit will be determined by antidromic identification. Macroelectrodes implanted at arm-related sites in M1, or in VLp and VLa, will be used to evoke antidromic spikes and, alternatively, to monitor local field potentials (LFPs). The direction of information flow between VL and M1 will be estimated using a Granger causality analysis of LFP and spike data. Aim 1 will test whether the task-related activity of M1-projecting neurons in VLp and VLa differ as current theory would predict from the subcortical inputs received: VLp neurons may encode kinematics and goal-appropriate feedback signals, consistent with the role hypothesized for cerebellum in predicting sensory consequences of motor commands. VLa neurons may instead signal task context, consistent with a role for the basal ganglia in context-dependent selection. Aim 2 will determine if cortico-thalamic neurons in M1 transmit different task-related information to VLp and VLa. A comparison of the results from Aims 1 and 2 will determine if the task information encoded in VLp and VLa can be explained by the information transmitted to those nuclei from M1. Granger causality analyses of LFP and spike data from Aims 1 and 2 will determine the direction of information flow between M1 and VLp/VLa. Finally, Aim 3 will test the causal influence on task performance of focal inactivations in VL. Inactivations in VLp may induce global ataxia-like impairments independent of task context whereas inactivations in VLa may selectively impair context-dependent modulations of task performance. This project will shed light on the basic functions of a circuit that is a central plaer in the pathophysiology of disorders of movement such as Parkinson's disease, dystonia, essential tremor and ataxia. Aim 3 has particular significance because it will identify sequelae that may accompany neurosurgical interventions (i.e., thalamotomy and deep brain stimulation) that target the VLp and VLa.

The fluid, seemingly effortless execution of sequences of movements is a ubiquitous feature of everyday motor behavior. Ample evidence for the importance of this ability comes from the common human diseases (Parkinson's disease, in particular) in which sequential skills are especially impaired. Long-term motor sequencing skills are formed, most likely, through the cooperation of parallel cortical-sub-cortical circuits involving associative, premotor, and motor regions of the brain. Recent evidence suggests that for each of these brain circuits, the sub-cortical loop through the basal ganglia (BG) contributes selectively to reinforcement-driven modulation of thalamo-cortical plasticity while cortex is well-suited as a site for long-term retention and efficient recall of well-practiced skills. These findings lead to the hypotheses that BG loops play central roles in the acquisition of sequence information whereas the anatomically-connected cortical areas are more important for the storage and use of already-learned information. The specific aims (SAs) of this proposal will test that general hypothesis by using non-human primates: (1) To determine if neurons in the globus pallidus interna (GPi, a primary BG output nucleus) preferentially encode sequence task information during learning and the production of recently learned sequences. Cortical neurons are predicted to not show a preference for recently learned sequences. (2) To test if intact BG circuits are necessary primarily for the learning and production of recently- learned sequences. Cortical circuits, in contrast, are predicted to be necessary even for well-learned sequences. Associative loops through cortex and BG may play greater roles in the fast acquisition and flexible recall of goal- directed sequence information. The premotor and motor loop circuits may mediate slow acquisition of habit-like effector-specific representations. We will infer the circuit membership of individual GPi neurons by stimulating different cortical areas and observing the orthodromic inhibitory effects. Animals will perform a discrete sequence production task alternating in blocks between random, novel-to-familiar and over-trained sequences. SA1 will test if neuronal encoding of task information in associative, premotor, and motor circuits reflects the predicted divergent roles for BG- and cortical-components of these circuits. SA2 will determine if interruptions of BG output (i.e., GPi inactivation or lesion) selectively impair the learning or recall of recently learned sequences. Inactivations of cortex, in contrast, are predicted to also disrupt the recall of well-learned sequences. Results from these experiments will aid in understanding the physiological basis for normal and impaired sequential behavior in humans.

Abstract Deep brain stimulation (DBS) is one of the most effective treatments for patients with advanced Parkinson's disease (PD). Delivery of high frequency electrical stimulation to the subthalamic nucleus (STN) ameliorates parkinsonian motor symptoms, often within seconds, but therapeutic effects wear off quickly if stimulation is stopped, often within minutes. This transient nature of symptomatic relief underscores the fact that existing DBS protocols mask symptoms but do not alleviate underlying circuit dysfunction. A modified DBS protocol, called coordinated reset (CR-DBS), has shown potential to provide long-lasting therapeutic benefits for days to weeks after stimulation, but this protocol has been slow to translate into widespread clinical use because (1) the multi- site, pseudorandom stimulation patterns required to implement it cannot be delivered with existing devices and (2) its mechanisms of action remain obscure, hindering insights into what parameters of CR-DBS should be tuned to ensure engagement of long-lasting effects. Recently, in a mouse model of PD, we discovered a cellular- based strategy to induce long-lasting motor recovery, by using optogenetics to target interventions to specific neuronal subpopulations in the external globus pallidus (GPe), an anatomical neighbor of the STN. Long-lasting motor rescue was induced by interventions that simultaneously increased the firing rates of GPe neurons enriched in parvalbumin (PV-GPe) and decreased the firing rates of GPe neurons enriched in lim homeobox 6 (Lhx6-GPe). Interestingly, at the physiological level, these cell-type specific interventions in the GPe converged upon a similar mechanism as CR-DBS, by ameliorating pathological patterns of neural activity in basal ganglia output nuclei that have been associated with parkinsonian motor deficits. This proposal will use knowledge gained from our discovery of long-lasting rescue through cell-type directed interventions in GPe to guide rational design and interrogation of human-applicable forms of DBS that may yield similarly long-lasting therapeutic benefit. Our experiments will test a novel, mechanistic hypothesis, based on supporting preliminary data, that the pattern of electrical DBS can be tuned to drive cell-type specific responses in the GPe that mirror those previously found to be sufficient to induce of long-lasting motor rescue with optogenetics. Experiments in Aim 1 will investigate the cellular mechanisms through which phasic stimulation in the STN evokes cell-type specific responses in the GPe (Aim 1.1) and use a machine learning approach to identify stimulation protocols that maximize this cell-type specific response (Aim 1.2). Experiments in Aim 2 will test the therapeutic efficacy of phasic stimulation protocols compared to conventional DBS, using behavioral and physiological assays in mouse (Aim 2.1) and primate (Aim 2.2) models of PD. Taken together, these experiments will advance our understanding of the fundamental differences between how conventional vs. phasic stimulation impacts the nervous system, with cell-type specific and synapse-specific resolution, and could provide novel therapeutic strategies that can be rapidly translated into humans.