De nombreuses études s’intéressent aux comportements décisionnels et d’apprentissage ainsiqu’aux structures qui les sous-tendent. Il a été montré que le Cortex Préfrontal (CPF) ainsiqu’un réseau de structures sous-corticales, les Ganglions de la Base (GB), étaient impliquésdans ces processus. Néanmoins, le rôle respectif de chacun n’est pas définit. Deux hypothèsessont émises. La première stipule que les deux structures fonctionnent indépendamment. LesGB seraient impliqués dans les comportements habituatifs tandis que le CPF se chargerait descomportements planifiés. La seconde hypothèse considère que les deux structures collaborent: les GB contrôleraient un processus d'apprentissage à cinétique lente dans le CPF et sedésengageraient progressivement au fur et à mesure de l’apprentissage. Ceci reviendrait d'unecertaine façon à inverser les rôles : les GB seraient nécessaires aux processus de décision tantque le CPF n'aurait pas fini son apprentissage. Celui-ci fonctionnerait ensuite sur un modeautomatique. Le principal obstacle à l’étude respectif du rôle des GB et du CPF dans cesprocessus intervient dans les paradigmes expérimentaux qui ne dissocient pas la prise dedécision per se des processus d’apprentissage. Notre premier objectif a donc été d’élaborerune tâche expérimentale qui permette de différencier les phases d’apprentissage des phases deprise de décision. Nous avons ensuite supprimé l'influence des GB sur le cortex, en inhibantleur structure de sortie, le Globus Pallidus interne (GPi) par des injections intracérébrales demuscimol chez le primate non-humain effectuant une tâche comportementale : le "two armedbandit task". Nous montrons que les animaux sont toujours capables de prendre des décisionsaprès inhibition du GPi mais qu’ils sont incapables d’apprendre la valeur de nouvelles cibles.Ces résultats confirment que, chez le primate en tous les cas, les GB et le CPF sont bienimpliqués dans un processus collaboratif : l'intégrité de l'ensemble du circuit est nécessairepour l'apprentissage alors que le cortex seul peut suffire une fois que le choix se situe dans uncontexte habituel. / Many studies are interested in decision making and learning processes and in brainareas which are engaged in. Among them, the implication Prefrontal Cortex (PFC) and a subcortical structures’ network, the Basal Ganglia (BG) has been shown. Nevertheless, theprecise role of each structure has not yet been defined. There are two main hypotheses. Thefirst one holds that GB and PFC function independently. BG would support habitualbehaviors and PFC planned behaviors. The second hypothesis proposes that both structuresare collaborating: the basal ganglia drive a low kinetic learning process in the prefrontalcortex and become less and less engaged as the task is learned. It means reversing the roles:BG would be necessary for decision making processes as soon as PFC finishes its learning.This latter would then function as an automatic mode. The main problem which avoids us todisentangle the role of each structure is the experimental paradigms used which mix uplearning and decision making. Our first aim was to design an experimental task in which therewas learning phase and decision making phase per se. Then, we blocked basal gangliainfluence on PFC by inhibiting their exit structure, the Globus Pallidus internal, withintracerebral muscimol injections in non-human primates performing a “two-armed bandittask”. Our results show that monkeys are able to do decision making after GPi inhibition butthey are unable to learn new values. These results confirm that, in non-human primates, BGand PFC are well involved as co-workers in one process: integrity of all the circuit isnecessary for learning whereas only cortex is sufficient once the choice is in habitual context.

L-Dihydroxyphenylalanine (L-DOPA), the anti-parkinsonian drug affording the greatest symptomatic relief of parkinsonian symptoms, is still misunderstood in terms of its neurotoxic potential and the mechanism by which generated dopamine (DA) is able to exert an effect despite the absence of DA innervation of target sites in basal ganglia. This review summaries important aspects and new developments on these themes. On the basis of L-DOPA therapy in animal models of Parkinson's disease, it appears that L-DOPA is actually neuroprotective, not neurotoxic, as indicated by L-DOPAs reducing striatal tissue content of the reactive oxygen species, hydroxyl radical (HO•), and by leaving unaltered the extraneuronal in vivo microdialysate level of HO•. In addition, the potential beneficial anti-parkinsonian effect of L-DOPA is actually increased because of the fact that the basal ganglia are largely DA-denervated. That is, from in vivo microdialysis studies it can be clearly demonstrated that extraneuronal in vivo microdialysate DA levels are actually higher in the DA-denervated vs. the intact striatum of rats - owing to the absence of DA transporter (i.e., uptake sites) on the absent DA nerve terminal fibers in parkinsonian brain. In essence, there are fewer pumps removing DA from the extraneuronal pool. Finally, the undesired motor dyskinesias that commonly accompany long-term L-DOPA therapy, can be viewed as an outcome of L-DOPAs sensitizing DA receptors (D1-D5), an effect easily replicated by repeated DA agonist treatments (especially agonist of the D 2 class) in animals, even if the brain is not DA-denervated. The newest findings demonstrate that L-DOPA induces BDNF release from corticostriatal fibers, which in-turn enhances the expression of D3 receptors; and that this effect is associated with motor dyskinesias (and it is blocked by D3 antagonists). The recent evidence on mechanisms and effects of L-DOPA increases our understanding of this benefical anti-parkinsonian drug, and can lead to improvements in L-DOPA effects while providing avenues for reducing or eliminating L-DOPAs deleterious effects.

Animals respond to changes in the environment and internal states to modify their behavior. The basal ganglia, including the striatum contribute to action selection by integrating sensory, motor and reward information. Therefore, dysregulation of striatal function is common in many neuropsychiatric disorders, including Parkinson’s disease, Huntington disease, schizophrenia, and addiction. Here, using fiber photometry, pharmacology, and behavioral approaches in transgenic mice, I explored the cellular and circuit mechanisms underlying key striatal functions. In Chapter 1, I begin by presenting the existing literature on the anatomy and physiology of the striatum. Next, I review the important functions of the striatum. Within this general review, I highlight the specific roles that striatal (DA) and acetylcholine (ACh) play in striatal circuitry and function. In Chapter 2, I demonstrate the naturally evoked ACh dip has a DA component and a non-DA component. Specifically, I show that DA via cholinergic DA D2 receptors (D2Rs) modulate the length of the ACh dip and rebound ACh levels following the dip. In addition, I show that DA coordinates the activity between DA and ACh during behavior. Finally, I present data that supports a role for ACh in motivated behavior. In Chapter 3, I show that cholinergic D2Rs are not necessary for reward learning but do facilitate reversal learning in a probabilistic choice task. In addition, I show that changes in DA and ACh levels contribute to reversal learning in a probabilistic choice task. Finally, in Chapter 4, I discuss the general conclusions and study implications, as well as future directions.

The early postnatal period is marked by the rapid acquisition of sensorimotor processing capabilities. Initially responding to a limited set of environmental stimuli with a restricted repertoire of behaviors, mammals exhibit a remarkable proliferation of sensorimotor abilities in the early postnatal period. Central to action selection, reinforcement, and contingency learning are a subcortical set of evolutionarily conserved nuclei called the basal ganglia. The striatum, which is the primary input nucleus of the basal ganglia, receives afferent innervation from throughout the CNS. Its projection neurons (SPNs) integrate these diverse inputs, regulating movement and encoding salient cue-outcome contingencies. Here, using electrophysiological, electrochemical, imaging, and behavioral approaches in mice, I will explore the postnatal maturation of the striatal cholinergic interneuron (ChI), a critical modulator of dopamine signaling, afferent excitation, and SPN excitability. In Chapter 1, I will set the stage for this exploration by reviewing the current literature on striatal postnatal development, including cellular physiology, axonal elaboration and synapse formation, and plasticity expression. I will survey striatal deficits observed in clinical neurodevelopmental conditions such as autism, ADHD, tic disorders, and substance use disorders. I will additionally summarize evidence that the striatum is uniquely vulnerable to physiological and immunological insult, as well as early life adversity. In Chapter 2, I turn my focus specifically to the striatal ChI, uncovering fundamental cell-intrinsic changes that occur postnatally in this population. I will also elaborate on the postnatal maturation of dopamine release properties and regulation thereof by cholinergic signaling from the ChI. In Chapter 3, I investigate the circuit connectivity and circuit-driven firing dynamics of ChIs as they mature postnatally. I utilize a brain slice preparation retaining thalmostriatal afferents in order to assay the ChI pause, a synchronized transient quiescence in ChIs thought to facilitate cue learning and behavioral flexibility. I find that the ChI pause is refined postnatally, dependent on developmental changes in thalamic input strength and the cell- intrinsic expression of specific ionic conductances. Finally, in Chapter 4, I present preliminary evidence that ChI circuit maturation as defined in preceding chapters is delayed by chronic stress exposure postnatally. Following the maternal separation model of early life stress, ChI intrinsic characteristics mature normally, but they retain heightened thalamic innervation and thalamus-driven pause expression.

Autism spectrum disorders (ASDs) are characterized by phenotypic and genetic heterogeneity. Our analysis of functional networks perturbed in ASD suggests that both truncating and non-truncating de novo mutations contribute to autism. Moreover, we find that truncating mutations affecting the same exon lead to strikingly similar intellectual phenotypes in unrelated ASD probands and propose that exons, rather than genes, represent a unit of effective phenotypic impact for truncating mutations in autism. The phenotypic effects are likely mediated by nonsense-mediated decay of splicing isoforms and similar patterns may be observed in other genetic disorders. While multiple cell types and brain areas are affected, the impact of ASD mutations converge on a strongly interconnected system of neural structures that involve basal ganglia loops and the limbic system. We observe that distant projections constitute a disproportionately large fraction of the network composition, suggesting that the integration of diverse brain regions is a key property of the neural circuit. We demonstrate that individual de novo mutations impact several disparate components of the network and may further explain the phenotypic variability. Overall, our study presents a method that, to our knowledge, is the first unbiased approach using genetic variants to comprehensively discover and identify the neural circuitry affected in a psychiatric disorder.

Parkinson’s Disease is a disorder of the midbrain dopaminergic system with characteristic neurodegenerative patterns, recognized for its motor symptoms. The neurodegeneration is most prevalent in the substantia nigra pars compacta, while dopaminergic neurons in neighboring structures are comparatively spared. There are many possible explanations for this disparity, including differences in tolerance to oxidative stress, and vulnerability to α-synuclein aggregates. The substantia nigra is part of the basal ganglia, a network of nuclei in the midbrain and base of the forebrain which are responsible for coordinating voluntary movement. Dopamine has an inhibitory effect in the basal ganglia. It dampens signals to remove noise, so the basal ganglia circuitry is not hyperactive. In the absence of dopamine, the flow of information through the basal ganglia is disrupted. This results in tremor, bradykinesia, and rigidity, known as the classic triad. No cure currently exists and therapies are unable to slow disease progression, so treatments are aimed at symptom management. Degenerative processes in Parkinson’s Disease occur rapidly, early in the disease progression, with about 60% neuronal death in the substantia nigra prior to diagnosis. There is a need for biomarkers or other signs which can be used to clinically to diagnose the disease at an earlier stage. In conclusion this paper provides suggestions for future lines of research.

This dissertation is a collection of a total of seven publications that deal with eye movement disorders in patients with basal ganglia disorders. We obtained normative data for videooculography in healthy individuals. We have described the eye movement evolution during a human life such as the increase of latency, movements become hypometric and antisaccadic error rate increases. We have shown that sex and education do not affect the eye movements. Our study highlighted the asymmetry in the eye movement performance. As the first, we studied the vergence in patients with Parkinson's disease (PN) using videooculography (VOG). We devised and defined a paradigm for this examination and saw that in patients with PN there is a prolonged latency and hypometry of divergence. In patients with ephedrone induced parkinsonism (EP), we were the first who examined eye movements and found that it was possible to identify between this toxic Parkinson's syndrome and PN on the basis of a videooculography. In EP patients, we described velocity decsrease and hypometry in horizontal saccades, prolonged latency in horizontal saccades, and higher error rate in the antisacadic task. Behavioral disorder in REM sleep (RBD) as a prodromal stage of PN leads to impaired eye movement. In the evaluation with PN patients, we...

京都大学 / 新制・論文博士 / 博士(医学) / 乙第13453号 / 論医博第2246号 / 新制||医||1055(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 渡邉 大, 教授 林 康紀, 教授 高橋 淳 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

No description available.

In this work, we develop models of neural circuits and plasticity rules that underlie different forms of learning and memory, with a focus on learning processes that involve multiple brain regions. We begin by surveying the literature on synaptic plasticity rules and implementations of learning algorithms in the brain. Each subsequent chapter presents a model of how a specific aspect of learning is implemented biologically, based on experimental evidence and normative considerations. We first focus on the neural basis of reinforcement learning in the basal ganglia. We show that in order to enable effective learning when control of behavior is distributed across multiple regions (``off-policy reinforcement learning''), classic models of dopamine activity must be adapted to include an additional action-sensitive component. We also show that the known plasticity rules of direct and indirect-pathway striatal projection neurons are inconsistent with existing models of striatal codes for action. We propose and find experimental support for a new model of striatal activity driven by efferent input. This model is functionally compatible with striatal plasticity rules and enables simultaneous multiplexing of action-selection and learning signals, a necessary ingredient for off-policy reinforcement learning. We next use an off-policy reinforcement learning model to explain a new experimental finding about the conditions under which learned motor skills are consolidated to be driven by the dorsolateral striatum in rats. We then shift our focus to consider consolidation more broadly, proposing a general model of the advantages of systems in which memories and learned behaviors are consolidated from short-term to long-term learning pathways. In particular, our model proposes that such architectures enable selective filtering of the set of experiences used for learning, which can be essential in noisy environments with many extraneous stimuli. In the appendices, we explore other factors relevant to learning algorithms, including the interaction between multiple sensory modalities, and the problem of credit assignment in multi-layer neural networks. In summary, this work presents a varied set of models of different forms of learning in the brain, emphasizing the cooperative role of plasticity rules and multi-regional circuit architecture in producing functionally useful synaptic weight updates.