Search

[Resumo] Nesta tese preséntanse unha serie de algoritmos complexos a partir de Redes de Neuronas Artificiais (RNA) clásicas deseñados desde unha perspectiva mulitidiciplinar que achega os campos da Intelixencia Artificial e a Neurociencia, buscando representar in silico a sinapsis tripartita. Todos os algoritmos deseñanse como un circuito complexo no que unha RNA modula a outra RNA principal para conseguir mellorar o seu rendemento. O algoritmo central nesta tese e o chamado DAM, o cal conta cun astrocito artificial que tamén modula o circuito de dous RNAs mencionado. Para estudar a influencia dun astrocito artificial sobre o circuito, deseñanse otros dous algoritmos de control. O primeiro deles, DM, componse exclusivamente da segunda RNA modulando a RNA principal; mentres que o segundo, CONTROL, e un algoritmo que como elemento modulador intermedio ten un elemento que ten un funcionamento interno idéntico ao das demais neuronas artificiais que componen o resto do circuito. Deseñaronse un total de catro comportamentos do astrocito artificial inspirados en diferentes realidades biolóxicas para explorar se había algunha das elixidas que mais favorecese ao noso algoritmo a hora de resolver problemas de Aprendizaxe Máquina., [Resumen] En esta tesis se presentan una serie de algoritmos complejos a partir de Redes de Neuronas Artificiales (RNA) clásicas diseñados desde una perspectiva mulitidiciplinar que aúna los campos de la Inteligencia Artificial y la Neurociencia, buscando representar in silico la sinapsis tripartita. Todos los algoritmos se diseñan como un circuito complejo en el que una RNA modula a otra RNA principal para conseguir mejorar su rendimiento. El algoritmo central en esta tesis es el llamado DAM, el cual cuenta con un astrocito artificial que también modula el circuito de dos RNAs mencionado. Para estudiar la influencia de un astrocito artificial sobre el circuito, se diseñan otros dos algoritmos de control. El primero de ellos, DM, se compone exclusivamente de la segunda RNA modulando la RNA principal; mientras que el segundo, CONTROL, es un algoritmo que como elemento modulador intermedio tiene un elemento que tiene un funcionamiento interno idéntico al de las demás neuronas artificiales que componen el resto del circuito. Se diseñaron un total de cuatro comportamientos del astrocito artificial inspirados en diferentes realidades biológicas para explorar si había alguna de las elegidas que más favoreciera a nuestro algoritmo a la hora de resolver problemas de Aprendizaje Máquina., [Abstract] This thesis presents a series of complex algorithms based on classical Artificial Neuron Networks (ANN) designed from a multidisciplinary perspective that brings together the fields of Artificial Intelligence and Neuroscience, seeking to represent in silico the tripartite synapse. All algorithms are designed as a complex circuit in which an ANN modulates a main ANN to improve its performance. The central algorithm in this thesis is the so-called DAM, which relies on an artificial astrocyte that also modulates the circuit of two RNAs mentioned above. To study the influence of an artificial astrocyte on the circuit, two other control algorithms are designed. The first one, DM, is composed exclusively of the second ANN modulating the main ANN; while the second one, CONTROL, is an algorithm that as an intermediate modulating element has an element that has an internal functioning identical to that of the other artificial neurons that compose the rest of the circuit. A total of four behaviors of the artificial astrocyte inspired by different biological realities were designed to explore if there were any of the chosen ones that most favored our algorithm when solving Machine Learning problems.

Astrocytes are active cells involved in brain function through the bidirectional communication with neurons, in which the astrocyte calcium signal plays a crucial role. Synaptically-evoked calcium increases can be localized to independent subcellular domains or expand to the entire cell, i.e., calcium surge. In turn, astrocytes may regulate individual synapses by calcium-dependent release of gliotransmitters. Because a single astrocyte may contact ~100,000 synapses, the control of the intracellular calcium signal propagation may have relevant consequences on brain function by regulating the spatial range of astrocyte neuromodulation of synapses. Yet, the properties governing the spatial dynamics of the astrocyte calcium signal remains poorly defined. Imaging subcellular responses of cortical astrocytes to sensory stimulation in mice, we show that sensory-evoked astrocyte calcium responses originated and remained localized in domains of the astrocytic arborization, but eventually propagated to the entire cell if a spatial threshold of >23% of the arborization being activated was surpassed. Using transgenic IP3R2-/- mice, we found that type-2 IP3 receptors were necessary for the generation of the astrocyte calcium surge. We finally show using in situ electrophysiological recordings that the spatial threshold of the astrocyte calcium signal consequently determined the gliotransmitter release. Present results reveal a fundamental property of astrocyte calcium physiology, i.e., a spatial threshold for the astrocyte intracellular calcium signal propagation, which depends on astrocyte intrinsic properties and governs the astrocyte integration of local synaptic activity and the subsequent neuromodulation.

Antisense oligonucleotide (ASO) therapy for neurological disease has been successful in clinical settings and its potential has generated hope for Alzheimer's disease (AD). We previously described that ablating SNCA encoding for alpha-synuclein (alpha Syn) in a mouse model of AD was beneficial. Here, we sought to demonstrate whether transient reduction of alpha Syn expression using ASO(SNCA) could be therapeutic in a mouse model of AD. The efficacy of the ASO(SNCA) was measured via immunocytochemistry, RT-qPCR and western blotting. To assess spatial learning and memory, ASO(SNCA) or PBS-injected APP and non-transgenic (NTG) mice, and separate groups of SNCA-null mice, were tested on the Barnes circular maze. Hippocampal slice electrophysiology and transcriptomic profiling were used to explore synaptic function and differential gene expression between groups. Reduction of SNCA transcripts alleviated cognitive deficits in male transgenic animals, but surprisingly, not in females. To determine the functional cause of this differential effect, we assessed memory function in SNCA-null mice. Learning and memory were intact in male mice but impaired in female animals, revealing that the role of alpha Syn on cognitive function is sex-specific. Transcriptional analyses identified a differentially expressed gene network centered around EGR1, a central modulator of learning and memory, in the hippocampi of SNCA-null mice. Thus, these novel results demonstrate that the function of alpha Syn on memory differs between male and female brains.