Dissertations / Theses on the topic 'Plasticità Hebbiana'

Nella presente tesi di dottorato, ho esplorato se fenomeni di apprendimento Hebbiano possano governare il funzionamento dei sistemi cross-modali e sensorimotori del cervello umano. A tal fine, durante il mio dottorato, ho sviluppato e testato due nuovi protocolli Paired Associative Stimulation (PAS), una classe di tecniche di stimolazione cerebrale non invasiva in cui una stimolazione sensoriale periferica viene ripetutamente accoppiata con un impulso di stimolazione magnetica transcranica (TMS) su un’area bersaglio al fine di indurre plasticità associativa Hebbiana. I due protocolli PAS presentati nella mia tesi mirano a due sistemi cerebrali sensoriali-motori con funzionamento a specchio (tactile mirror system e action observation network), sfruttando rispettivamente una via cross-corticale visuo-tattile (cross-modal PAS) e una visuo-motoria (mirror PAS). Nel primo capitolo del presente lavoro, dopo una breve introduzione al concetto di plasticità associativa Hebbiana, fornirò una revisione esaustiva dei protocolli PAS che mirano ai sistemi sensorimotori, proponendo una classificazione in tre macro-categorie (within-system, cross-systems e cortico-cortical), a seconda delle caratteristiche delle stimolazioni accoppiate. Nel secondo capitolo descriverò le principali proprietà del sistema dei neuroni specchio (MNS) considerando anche le sue proprietà cross-modali visuo-tattili ed i meccanismi di plasticità neuronale che sono stati ipotizzati alla base dello sviluppo dei neuroni specchio. Nel terzo capitolo, introdurrò il cross-modal PAS (cm-PAS), un nuovo cross-systems PAS sviluppato per sfruttare le proprietà visuo-tattili della corteccia somatosensoriale primaria, al fine di indurre plasticità associativa Hebbiana in tale regione sensoriale. In una serie di tre esperimenti, ho testo la dipendenza temporale (Esperimento 1), la specificità corticale (Esperimento 2) e visiva (Esperimento 3) del protocollo, misurando possibili cambiamenti nell'acuità tattile dei partecipanti. Nell'esperimento 3, ho valutato anche possibili cambiamenti neurofisiologici all'interno di S1, registrando i potenziali evocati somatosensoriali. Infine, in un quarto esperimento, la dipendenza temporale del cm-PAS è stata ulteriormente studiata, testando l'ipotesi che meccanismi anticipatori di tipo predittivo possano svolgere un ruolo centrale nell'efficacia del protocollo. Nel quarto capitolo introdurrò un secondo cross-systems PAS: il mirror PAS (m-PAS) che sfrutta le proprietà ‘mirror’ visuo-motorie del cervello umano. A differenza del cm-PAS, questo secondo protocollo sfrutta la natura associativa dell'integrazione visuo-motoria all'interno del MNS, mirando a indurre un nuovo, atipico, fenomeno di risonanza motoria attraverso apprendimento Hebbiano. In tre esperimenti ho testato la dipendenza temporale (Esperimento 1), la specificità visiva (Esperimento 2) e corticale (Esperimento 3) del protocollo registrando i potenziali evocati motori durante la visione di semplici movimenti (i.e., risonanza motoria). Inoltre, nel terzo esperimento, ho esplorato anche possibili effetti comportamentali dell’m-PAS, utilizzando un compito di compatibilità imitativa che sfrutta il fenomeno dell'imitazione automatica. Infine, nel capitolo conclusivo, discuterò i risultati teorici, metodologici e clinici e le prospettive future che derivano da questi due protocolli.
In the present doctoral thesis, I have explored whether Hebbian learning may rule the functioning of cross-modal and sensory-motor networks of the human brain. To this aim, during my doctorate, I have developed and tested two novel Paired Associative Stimulation (PAS) protocols, a class of non-invasive brain stimulation techniques in which a peripheral, sensory, stimulation is repeatedly paired with a Transcranial Magnetic Stimulation (TMS) pulse to induce Hebbian associative plasticity. The two PAS protocols presented in my thesis target sensory-motor networks with mirror functioning, exploiting a visuo-tactile (cross-modal PAS), and a visuo-motor pathway (mirror PAS), respectively. In the first chapter of the present work, after a brief introduction to the concept of Hebbian associative plasticity, I will provide an exhaustive review of PAS protocols targeting sensory-motor systems, proposing a classification in three macro-categories: within-system, cross-systems, and cortico-cortical protocols, according to the characteristics of the paired stimulations. In the second chapter, I will describe the principal properties of the Mirror Neuron System (MNS) also considering its cross-modal (i.e., visuo-tactile) characteristics and the plastic mechanisms that are been hypothesize at the ground of the development of mirror neurons’ matching properties. In the third chapter, I will introduce the cross-modal PAS (cm-PAS), a novel cross-systems PAS developed to exploit the visuo-tactile mirroring properties of the primary somatosensory cortex (S1) to induce Hebbian associative plasticity in such primary sensory region. In a series of three experiments, timing dependency (Experiment 1), cortical (Experiment 2), and visual specificity (Experiment 3) of the protocol have been tested, by measuring changes in participants’ tactile acuity. In Experiment 3, also possible neurophysiological changes within S1 has been assessed, recording somatosensory-evoked potentials (SEP). Then, in a fourth experiment, cm-PAS timing dependency has been further investigated, testing the hypothesis that anticipatory, predictive-like, mechanisms within S1 may play a central role in the effectiveness of the protocol. In the fourth chapter, a second cross-systems PAS will be introduced: the mirror PAS (m-PAS) which exploits visuo-motor mirroring properties of the human brain. Differently from the cm-PAS, this second protocol targets visuo-motor integration within the MNS and aims at induce a novel, atypical, motor resonance phenomena (assessed recording motor-evoked potentials – MEPs) following Hebbian learning. In three experiments, timing dependency (Experiment 1), visual (Experiment 2), and cortical specificity (Experiment 3) of the protocol have been tested. Furthermore, in the third experiment, the behavioral effects of the m-PAS are explored, using an imitative compatibility task exploiting automatic imitation phenomenon. Finally, in the conclusive chapter, I will discuss theoretical, methodological, and clinical outcomes and future perspectives that arise from these two protocols and the related results.

Information is transferred between neurons in the brain via electrochemical transmission at specialized cell-cell junctions called synapses. These structures are far from being static, but rather are influenced by plasticity mechanisms that alter features of synaptic transmission as means to build routes of information flow in the brain. Hebbian forms of synaptic plasticity – long-term potentiation and long-term depression – have been well studied and are considered to be the cellular basis of learning and memory, although their positive feedback nature is prone to instability. Neurons are also endowed with homeostatic mechanisms of synaptic plasticity that act to stabilize neural network functions by globally tuning synaptic drive. Precisely how neurons orchestrate this adaptive homeostatic response and how it influences Hebbian forms of synaptic plasticity, however, remains only partially understood. Using a combination of whole-cell electrophysiology, two-photon imaging and glutamate uncaging in organotypic hippocampal slices, I have expanded upon the known repertoire of homeostatic mechanisms that increase excitatory synaptic drive when CA1 hippocampal neurons experience a prolonged period of diminished activity. I found that the subunit composition of AMPA and NMDA receptors, the two major glutamate receptor subtypes at excitatory synapses, are altered which, in addition to increasing synaptic strength, are predicted to change the signaling and integrative properties of synaptic transmission. Moreover, I found that the amount of glutamate released from presynaptic terminals during evoked-transmission is enhanced and that this mechanism might, in part, underlie the uniform cell-wide homeostatic increase in synaptic strengths. Lastly, I found that homeostatic strengthening of synaptic transmission reduced the potential for CA1 synapses to exhibit long-term potentiation, and that this was caused by altered presynaptic release dynamics that impeded plasticity induction. Together, this work highlights several mechanistic strategies employed by neurons to increase excitatory synaptic drive during periods of activity deprivation which, in addition to balancing cellular excitability, alters the metaplastic state of synapses.

Pendant des décennies, la plasticité synaptique a été considérée comme le substrat principal de la plasticité fonctionnelle cérébrale. Récemment, plusieurs études expérimentales indiquent que des régulations à long terme de l’excitabilité intrinsèque participent à la plasticité dépendante de l’activité. En effet, la modulation des canaux ioniques dépendants du potentiel, lesquels régulent fortement l’excitabilité intrinsèque et l’intégration des entrées synaptiques, a été démontrée essentielle dans les processus d’apprentissage. Cependant, la régulation, dépendante de l’activité, du courant ionique activé par l’hyperpolarisation (Ih) et ses conséquences sur l’induction de futures plasticités reste à éclaircir, tout comme la présence d’une régulation de conductances dépendantes du potentiel dans les neurones inhibiteurs. Dans la première partie de ma thèse, nous caractérisons les mécanismes d’induction et d’expression de la plasticité à long terme de l’excitabilité (LTP-IE) dans les interneurons en panier de la région CA1 exprimant la parvalbumine. Dans une seconde partie, le rôle de Ih dans la régulation homéostatique de l’excitabilité neuronale induite par des manipulations de l’activité neuronale dans sa globalité a été étudié. Dans la troisième étude, nous montrons que la magnitude de la Dépression à Long Terme (LTD) détermine le sens de la régulation de Ih dans les neurones pyramidaux de CA1. En conclusion, cette thèse montre qu’à la fois dans les neurones excitateurs et inhibiteurs, les régulations des conductances dépendantes du potentiel aident à maintenir une relative stabilité dans l’activité du réseau
Synaptic plasticity has been considered for decades as the main substrate of functional plasticity in the brain. Recently, experimental evidences suggest that long-lasting regulation of intrinsic neuronal excitability may also account for activity-dependent plasticity. Indeed, voltage-dependent ionic channels strongly regulate intrinsic excitability and inputs integration and their regulation was found to be essential in learning process. However, activity-dependent regulation of the hyperpolarization-activated ionic current (Ih) and its consequences for future plasticity remain unclear, so as the presence of any voltage-dependent conductances regulation in inhibitory neurons. In the first part of this thesis, we report the characterization of the induction and expression mechanisms of Long-Term Potentiation of Intrinsic Excitability (LTP-IE) in CA1 parvalbumin-positive basket interneurons. In a second part, the role of Ih in the homeostatic regulation of intrinsic neuronal excitability induced by global manipulations of neuronal activity was reported. In the third experimental study, we showed that the magnitude of Long-term Depression (LTD) determines the sign of Ih regulation in CA1 pyramidal neurons. In conclusion, this thesis shows that in both excitatory and inhibitory neurons, activity-dependent regulations of voltage-dependent conductances help to maintain a relative stability in the network activity

L'homme est capable d'utiliser des stratégies ou règles concurrentes selon les contraintes environnementales. Nous étudions un modèle plausible pour une tâche nécessitant l'apprentissage de plusieurs règles associant des stimuli visuels à des réponses motrices. Deux réseaux de populations neurales à sélectivité mixte interagissent. Le réseau décisionnel apprend les associations stimulus-réponse une à une, mais ne peut gérer qu'une règle à la fois. Son activité modifie la plasticité synaptique du second réseau qui apprend les statistiques d'évènements sur une échelle de temps plus longue. Lorsque des motifs entre les associations stimulus-réponse sont détectés, un biais d'inférence vers le réseau décisionnel guide le comportement futur. Nous montrons que le mécanisme de Hebb non-supervisé dans le second réseau est suffisant pour l'implémentation des règles. Leur récupération dans le réseau de décision améliore la performance. Le modèle prédit des changements comportementaux en fonction de la séquence des réponses précédentes, dont les effets sur la performance peuvent être positifs ou négatifs. Les prédictions sont confirmées par les données, et permettent d'identifier les sujets ayant appris la structure de la tâche. Le signal d'inférence corrèle avec l'activité BOLD dans le réseau fronto-pariétal. Au sein de ce réseau, les n¿uds préfrontaux dorsomédial et dorsolatéral sont préférentiellement recrutés lorsque les règles sont récurrentes: l'activité dans ces régions pourrait biaiser les circuits de décision lorsqu'une règle est récupérée. Ces résultats montrent que le mécanisme de Hebb peut expliquer l'apprentissage de comportements complexes en contrôle cognitif
Depending on environmental demands, humans performing in a given task are able to exploit multiple concurrent strategies, for which the mental representations are called task-sets. We examine a candidate model for a specific human experiment, where several stimulus-response mappings, or task-sets, need to be learned and monitored. The model is composed of two interacting networks of mixed-selective neural populations. The decision network learns stimulus-response associations, but cannot learn more than one task-set. Its activity drives synaptic plasticity in a second network that learns event statistics on a longer timescale. When patterns in stimulus-response associations are detected, an inference bias to the decision network guides successive behavior. We show that a simple unsupervised Hebbian mechanism in the second network is sufficient to learn an implementation of task-sets. Their retrieval in the decision network improves performance. The model predicts abrupt changes in behavior depending on the precise statistics of previous responses, corresponding to positive (task-set retrieval) or negative effects on performance. The predictions are borne out by the data, and enable to identify subjects who have learned the task structure. The inference signal correlates with BOLD activity in the fronto-parietal network. Within this network, dorsomedial and dorsolateral prefrontal nodes are preferentially recruited when task-sets are recurrent: activity in these regions may provide a bias to decision circuits when a task-set is retrieved. These results show that Hebbian mechanisms and temporal contiguity may parsimoniously explain the learning of rule-guided behavior

Cortical and subcortical microcircuits are continuously modified throughout life. Despite ongoing changes these networks stubbornly maintain their functions, which persist although destabilizing synaptic and nonsynaptic mechanisms should ostensibly propel them towards runaway excitation or quiescence. What dynamical phenomena exist to act together to balance such learning with information processing? What types of activity patterns do they underpin, and how do these patterns relate to our perceptual experiences? What enables learning and memory operations to occur despite such massive and constant neural reorganization? Progress towards answering many of these questions can be pursued through large-scale neuronal simulations.    In this thesis, a Hebbian learning rule for spiking neurons inspired by statistical inference is introduced. The spike-based version of the Bayesian Confidence Propagation Neural Network (BCPNN) learning rule involves changes in both synaptic strengths and intrinsic neuronal currents. The model is motivated by molecular cascades whose functional outcomes are mapped onto biological mechanisms such as Hebbian and homeostatic plasticity, neuromodulation, and intrinsic excitability. Temporally interacting memory traces enable spike-timing dependence, a stable learning regime that remains competitive, postsynaptic activity regulation, spike-based reinforcement learning and intrinsic graded persistent firing levels.    The thesis seeks to demonstrate how multiple interacting plasticity mechanisms can coordinate reinforcement, auto- and hetero-associative learning within large-scale, spiking, plastic neuronal networks. Spiking neural networks can represent information in the form of probability distributions, and a biophysical realization of Bayesian computation can help reconcile disparate experimental observations.

QC 20170421

L’integrazione multisensoriale è la capacità del sistema nervoso di utilizzare molteplici sorgenti sensoriali. Una tra le più studiate forme di integrazione è quella tra informazioni visive ed acustiche. La capacità di localizzare uno stimolo acustico nello spazio è un processo meno accurato ed affidabile della localizzazione visiva, di conseguenza, un segnale visivo è spesso in grado di “catturare” (ventriloquismo) o di incrementare (enhancement multisensoriale) la performance di localizzazione acustica. Numerose evidenze sperimentali hanno contribuito ad individuare i processi neurali e le aree cerebrali alla base dei fenomeni integrativi; in particolare, un importante contributo viene dallo studio su soggetti con lesioni cerebrali. Tuttavia molti aspetti sui possibili meccanismi coinvolti restano ancora da chiarire. Obiettivo di questa tesi è stato lo sviluppo di un modello matematico di rete neurale per fare luce sui meccanismi alla base dell’interazione visuo-acustica e dei suoi fenomeni di plasticità. In particolare, il modello sviluppato è in grado di riprodurre condizioni che si verificano in-vivo, replicando i fenomeni di ventriloquismo ed enhancement in diversi stati fisiopatologici e interpretandoli in termini di risposte neurali e reciproche interazione tra i neuroni. Oltre ad essere utile a migliorare la comprensione dei meccanismi e dei circuiti neurali coinvolti nell’integrazione multisensoriale, il modello può anche essere utile per simulare scenari nuovi, con la possibilità di effettuare predizioni da testare in successivi esperimenti.

The ability of the visual system for object recognition is remarkable. A better understanding of its processing would lead to better computer vision systems and could improve our understanding of the underlying principles which produce intelligence. We propose a computational model of the visual areas V1 and V2, implementing a rich connectivity inspired by the neocortical circuit. We combined the three most important cortical plasticity mechanisms. 1) Hebbian synaptic plasticity to learn the synapse strengths of excitatory and inhibitory neurons, including trace learning to learn invariant representations. 2) Intrinsic plasticity to regulate the neurons responses and stabilize the learning in deeper layers. 3) Structural plasticity to modify the connections and to overcome the bias for the learnings from the initial definitions. Among others, we show that our model neurons learn comparable receptive fields to cortical ones. We verify the invariant object recognition performance of the model. We further show that the developed weight strengths and connection probabilities are related to the response correlations of the neurons. We link the connection probabilities of the inhibitory connections to the underlying plasticity mechanisms and explain why inhibitory connections appear unspecific. The proposed model is more detailed than previous approaches. It can reproduce neuroscientific findings and fulfills the purpose of the visual system, invariant object recognition.
Das visuelle System des Menschen hat die herausragende Fähigkeit zur invarianten Objekterkennung. Ein besseres Verständnis seiner Arbeitsweise kann zu besseren Computersystemen für das Bildverstehen führen und könnte darüber hinaus unser Verständnis von den zugrundeliegenden Prinzipien unserer Intelligenz verbessern. Diese Arbeit stellt ein Modell der visuellen Areale V1 und V2 vor, welches eine komplexe, von den Strukturen des Neokortex inspirierte, Verbindungsstruktur integriert. Es kombiniert die drei wichtigsten kortikalen Plastizitäten: 1) Hebbsche synaptische Plastizität, um die Stärke der exzitatorischen und inhibitorischen Synapsen zu lernen, welches auch „trace“-Lernen, zum Lernen invarianter Repräsentationen, umfasst. 2) Intrinsische Plastizität, um das Antwortverhalten der Neuronen zu regulieren und damit das Lernen in tieferen Schichten zu stabilisieren. 3) Strukturelle Plastizität, um die Verbindungen zu modifizieren und damit den Einfluss anfänglicher Festlegungen auf das Lernergebnis zu reduzieren. Neben weiteren Ergebnissen wird gezeigt, dass die Neuronen des Modells vergleichbare rezeptive Felder zu Neuronen des visuellen Kortex erlernen. Ebenso wird die Leistungsfähigkeit des Modells zur invariante Objekterkennung verifiziert. Des Weiteren wird der Zusammenhang von Gewichtsstärke und Verbindungswahrscheinlichkeit zur Korrelation der Aktivitäten der Neuronen aufgezeigt. Die gefundenen Verbindungswahrscheinlichkeiten der inhibitorischen Neuronen werden in Zusammenhang mit der Funktionsweise der inhibitorischen Plastizität gesetzt, womit erklärt wird warum inhibitorische Verbindungen unspezifisch erscheinen. Das vorgestellte Modell ist detaillierter als vorangegangene Arbeiten. Es ermöglicht neurowissenschaftliche Erkenntnisse nachzuvollziehen, wobei es ebenso die Hauptleistung des visuellen Systems erbringt, invariante Objekterkennung. Darüber hinaus ermöglichen sein Detailgrad und seine Selbstorganisationsprinzipien weitere neurowissenschaftliche Erkenntnisse und die Modellierung komplexerer Modelle der Verarbeitung im Gehirn.

Synaptic plasticity determines the development of functional neural circuits. It is widely accepted as the mechanism behind learning and memory. Among different forms of synaptic plasticity, Hebbian plasticity describes an activity-induced change in synaptic strength, caused by correlated pre- and postsynaptic activity. Additionally, Hebbian plasticity is characterised by input specificity, which means it takes place only at synapses, which participate in activity. Because of its correlative nature, Hebbian plasticity suggests itself as a mechanism behind associative learning. Although it is commonly assumed that synaptic plasticity is closely linked to synaptic activity during development, the mechanistic understanding of this coupling is far from complete. In the present study channelrhodopsin-2 was used to evoke activity in vivo, at the glutamatergic Drosophila neuromuscular junction. Remarkably, correlated pre- and postsynaptic stimulation led to increased incorporation of GluR-IIA-type glutamate receptors into postsynaptic receptor fields, thus boosting postsynaptic sensitivity. This phenomenon is input-specific. Conversely, GluR-IIA was rapidly removed from synapses at which neurotransmitter release failed to evoke substantial postsynaptic depolarisation. This mechanism might be responsible to tame uncontrolled receptor field growth. Combining these results with developmental GluR-IIA dynamics leads to a comprehensive physiological concept, where Hebbian plasticity guides growth of postsynaptic receptor fields and sparse transmitter release stabilises receptor fields by preventing overgrowth. Additionally, a novel mechanism of retrograde signaling was discovered, where direct postsynaptic channelrhodopsin-2 based stimulation, without involvement of presynaptic neurotransmitter release, leads to presynaptic depression. This phenomenon is reminiscent of a known retrograde homeostatic mechanism, of inverted polarity, where neurotransmitter release is upregulated, upon reduction of postsynaptic sensitivity
Das Phänomen der synaptischen Plastizität bestimmt die Entwicklung funktionaler neuronaler Schaltkreise. Die meisten Neurowissenschaftler betrachten synaptische Plastizität als die neuronal Grundlage von Lernen und Gedächtnis. Es gibt viele Ausprägungsarten synaptischer Plastizität, eine davon ist die sogenannte Hebb’sche Plastizität. Diese ist definiert durch eine aktivitätsinduzierte, langanhaltende Veränderung der Stärke einer synaptischen Verbindung, verursacht durch korrelative Aktivierung der Prä- und der Postsynapse. Zusätzlich ist die Ausbreitung der Hebb’sche Plastizität synapsenspezifisch, d.h. nur die Synapsen, die an der korrelativen Aktivierung teilnehmen, erfahren auch die Veränderung. Das Wachstumssignal breitet sich also nicht auf benachbarte Synapsen aus. Der korrelative Wesenszug der Hebb’schen Plastizität macht sie zu einem naheliegenden zellulären Mechanismus assoziativen Lernens. Es wird angenommen, dass synaptische Aktivität und synaptische Plastizität während der Entwicklung neuronaler Schaltkreise eng gekoppelt sind. Das mechanistische Verständnis dieser Kopplung ist jedoch weitgehend unverstanden. In der vorliegenden Arbeit wurde das lichtaktivierbare Kanalrhodopsin-2 verwendet, um Aktivität an der glutamatergen neuromuskulären Synapse in der lebenden, sich frei bewegenden, Drosophila melanogaster Larve auszulösen. Wenn die Prä- und die Postsynapse korrelativ aktiviert wurden, führte dies zur verstärkten Integration von Glutamatrezeptoren des GluR-IIA Typs in die postsynaptischen Rezeptorfelder, was in einer Erhöhung der postsynaptischer Empfindlichkeit mündete. Dieses Platizitätsphänomen wurde als synapsenspezifisch identifiziert und damit als Hebb’sch. Im Gegenzug, wurde der gleiche Rezeptortyp entfernt, wenn Neurotransmitterfreisetzung nicht zu einer erheblichen Depolarisation der Postsynapse führte. Dieser Mechanismus könnte für die Kontrolle des Rezeptorfeldwachstums verantwortlich sein. Es wurde ein physiologisches Modell erarbeitet, bei dem Hebb’sche Plastizität das Wachstum postsynaptischer Rezeptorfelder während der Entwicklung leitet und sporadische, nicht synchronisierte Neurotransmitterfreisetzung die Rezeptorfeldgröße stabilisiert, indem sie das Wachstum Dieser begrenzt. Zusätzlich wurde eine neue Modalität der synaptischen Plastizität an der neuromuskulären Synapse entdeckt: Ein retrograder Signalweg wird aktiviert wenn die postsynaptische Seite, unter Umgehung der Präsynapse, direkt, lichtinduziert aktiviert wird. Dieser Signalweg führt zur präsynaptischen Depression. Das Phänomen erinnert stark an einen bereits bekannten retrograden homöostatischen Mechanismus, reziproker Polarität, bei dem Neurotransmitter Freisetzung hochreguliert wird, wenn die Empfindlichkeit der Postsynapse verringert wird