Neurone formel | EPFL Graph Search

thumb|Représentation d'un neurone formel (ou logique). Un neurone formel, parfois appelé neurone de McCulloch-Pitts, est une représentation mathématique et informatique d'un neurone biologique. Le neurone formel possède généralement plusieurs entrées et une sortie qui correspondent respectivement aux dendrites et au cône d'émergence du neurone biologique (point de départ de l'axone). Les actions excitatrices et inhibitrices des synapses sont représentées, la plupart du temps, par des coefficients numériques (les poids synaptiques) associés aux entrées. Les valeurs numériques de ces coefficients sont ajustées dans une phase d'apprentissage. Dans sa version la plus simple, un neurone formel calcule la somme pondérée des entrées reçues, puis applique à cette valeur une fonction d'activation, généralement non linéaire. La valeur finale obtenue est la sortie du neurone. Le neurone formel est l'unité élémentaire des réseaux de neurones artificiels dans lesquels il est associé à ses semblabl

Using the Green's function to simplify and understand dendrites

Willem Anna Mark Wybo

Neurons are endowed with dendrites: tree-like structures that collect and transform inputs. These arborizations are believed to substantially enhance the computational repertoire of neurons. While it has long been known that dendrites are not iso-potential units, only in the last few decades it was shown experimentally that dendritic branches can transform local inputs in a non-linear fashion. This finding led to the subunit hypothesis, which states that within the dendritic tree, inputs arriving in one branch are transformed non-linearly and independently from what happens in other branches. Recent progress in experimental recording techniques shows that this localized dendritic integration contributes to shaping behavior. While it is generally accepted that the dendritic tree induces multiple subunits, many questions remain unanswered. For instance, it is not known how much separation there needs to be between different branches to be able to function as subunits. Consequently, there is no information on how many subunits can coexist along a dendritic arborization. It is also not known what the input-output relation of these subunits would be, or whether these subunits can be modified by input patterns. As a consequence, assessing the effects of dendrites on the workings of networks of neurons remains mere guesswork. During this work, we choose a theory-driven approach to advance our knowledge about dendrites. Theory can help us understand dendrites by deriving accurate, but conceptually simple models of dendrites that still capture their main computational effects. These models can then be analyzed and fully understood, which in turn teaches us how actual dendrites function computationally. Such simple models typically require less computer operations to simulate than highly detailed dendrite models. Hence, they may also increase the speed of network simulations that incorporate dendrites. The Green's function forms the basis for our theory driven approach. We first explored whether it could be used to reduce the cost of simulating dendrite models. One mathematically interesting finding in this regard is that, because this function is defined on a tree graph, the number of equations can be reduced drastically. Nevertheless, we were forced to conclude that reducing dendrites in this way does not yield new information about the subunit hypothesis. We then focused our attention on another way of decomposing the Green's function. We found that the dendrite model obtained in this way reveals much information on the dendritic subunits. In particular, we found that the occurrence of subunits is well predicted by the ratio of input over transfer impedance in dendrites. This allowed us to estimate the number of subunits that can coexist on dendritic trees. We also found that this ratio can be modified by other inputs, in particular shunting conductances, so that the number of subunits on a dendritic tree can be modified dynamically. We finally were able to show that, due to this dynamical increase of the number of subunits, individual branches that would otherwise respond to inputs as a single unit, could become sensitive to different stimulus features. We believe that this model can be implemented in such a way that it simulates dendrites in a highly efficient manner. Thus, after incorporation in standard neural network simulation software, it can substantially improve the accessibility of dendritic network simulations to modelers.

EPFL2017

Spatial navigation in geometric mazes

Navigation is defined as the capability of planning and performing a path from the current position towards a desired location. Different types, or strategies, of navigation are used by animals depending on the task they are trying to solve. Visible goals can be approached directly, while navigation to a hidden goal usually requires a memorized representation of relative positions of the goal and surrounding landmarks. Neurophysiological and behavioral experiments on rodents suggest that different brain areas are responsible for the expression of different navigation strategies. Specifically, dorsal striatum has been related to storage and recall of stimulus-response associations underlying simple goal-approaching behaviors, whereas hippocampus is thought to store the spatial representation of the environment. Such a representation is built during an unrewarded spatial exploration and appears to be employed in cases when simple stimulus-response strategies fail. Discovery of neurons with spatially correlated activity, i.e. place cells and grid cells, in the hippocampal formation complements behavioral and lesion data suggesting its role for spatial orientation. The overall objective of this work is to study the neurophysiological mechanisms underlying rodent spatial behavior, in particular those that are responsible for the implementation of different navigational strategies. Special attention is devoted to the question of how various types of sensory cues influence goal-oriented behavior. The model of a navigating rat described in this work is based on functional and anatomical properties of brain regions involved in encoding and storage of space representation and action generation. In particular, place and grid cells are modeled by two interconnected populations of artificial neurons. Together, they form a network for spatial learning, capable of combining different types of sensory inputs to produce a distributed representation of location. Goal-directed actions can be generated in the model via two different neural pathways: the first one drives stimulus-response behavior and associates visual input directly to motor responses; the second one associates motor actions with places and hence depends on the representation of location. The visual input is represented by responses of a large number of orientation-sensitive filters to visual images generated according to the position and orientation of the simulated rat in a virtual three-dimensional world. The model was tested in a large array of tasks designed by analogy to experimental studies on animal behavior. Results of several experimental studies, behavioral as wells as neurophysiological, were reproduced. Based on these results we formulated a hypothesis about the influence that the rat's perception of surrounding environment exerts on goal-oriented behavior. This hypothesis may provide an insight into several issues in animal behavior research that were not addressed by theoretical models until now.

EPFL2007

Spatial navigation in geometric mazes

EPFL2007

Using the Green's function to simplify and understand dendrites

Willem Anna Mark Wybo

EPFL2017