KI-Glossar
Das vollständige Wörterbuch der Künstlichen Intelligenz
Leaky Integrate-and-Fire (LIF)
Simplified neuronal model that integrates input currents with a leak representing membrane charge loss, generating a spike when the potential reaches a predefined threshold before resetting.
Hodgkin-Huxley
Mathematical model describing action potential dynamics through coupled differential equations representing time-varying ionic conductances.
Izhikevich
Hybrid neuronal model combining the dynamic richness of Hodgkin-Huxley with the computational efficiency of LIF, capable of reproducing multiple neuronal firing regimes.
Action potential
Transient and brief electrical signal generated by neurons when they exceed their excitation threshold, propagating along the axon to transmit information.
Firing threshold
Critical value of membrane potential beyond which a neuron generates an action potential, determining neuronal sensitivity to incoming stimuli.
Refractory period
Time interval following a spike during which the neuron is temporarily unable to generate a new action potential, divided into absolute and relative phases.
Synaptic current
Flow of electrical charges across synapses in response to neurotransmitters, modifying the postsynaptic membrane potential according to their excitatory or inhibitory nature.
Membrane conductance
Measure of neuronal membrane permeability to ions, varying dynamically according to the opening and closing of voltage-dependent ion channels.
Membrane time constant
Parameter characterizing the speed of membrane potential change in response to stimuli, determined by membrane resistance and capacitance.
Stochastic differential equation
Mathematical formulation incorporating random terms to model neuronal noise and intrinsic variability of neuronal responses to stimuli.
FitzHugh-Nagumo model
Two-dimensional simplification of the Hodgkin-Huxley model capturing the essence of neuronal excitability with a recovery variable and a fast activation variable.
Spike-timing-dependent plasticity (STDP)
Learning mechanism where synaptic modification depends on the relative timing of pre- and postsynaptic spikes, strengthening or weakening connections according to their temporal correlation.
Temporal encoder
Mechanism for converting analog information into temporal spike trains, where information is encoded in the precise timing of action potential firing.
Adaptive exponential integrate-and-fire neuron (AdEx)
Neuronal model extending LIF with exponential adaptation dynamics to reproduce regular, bursting, and adaptive firing regimes observed in vivo.
Morris-Lecar model
Two-variable model describing neuronal excitability with calcium and potassium dynamics, particularly suited for studying oscillations and bifurcations.
Excitatory postsynaptic potential (EPSP)
Gradual depolarization of the postsynaptic membrane potential increasing the probability of spike generation, resulting from the activation of excitatory synapses.
Potentiel postsynaptique inhibiteur (IPSP)
Hyperpolarisation ou shunt inhibiteur du potentiel membranaire postsynaptique réduisant la probabilité de décharge neuronale, médié par les neurotransmetteurs inhibiteurs comme GABA.
Dynamique temporelle
Évolution des variables neuronales dans le temps, incluant les changements de potentiels membranaires, les états de canaux ioniques et les patterns de spikes.
Variable de récupération
État interne du neurone modélisant les processus d'inactivation lente et d'adaptation, contrôlant la période réfractaire et la régulation de la fréquence de décharge.
Conductances ioniques
Paramètres décrivant la perméabilité sélective de la membrane aux différents ions (Na+, K+, Ca2+ Cl-), régulés par des canaux voltage-dépendants ou chimiquement contrôlés.