Modeling of potentiation as cascaded gated processes; relevance to learning and seizure
Open AccessPoster presentation
- 630 Downloads
KeywordsAnimal Model Single Location Frequency Pattern Model Neuron Elemental Functionality
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
The neuron has elemental functionalities such as voltage-gated pores and allosterically-gated enzymes. Such functionalities are cascaded in the neuron resulting in complex functionalities. Such a functionality is potentiation. Potentiation is characterized by an excitation frequency/excitatory-postsynaptic-potential (EPSP) slope relationship. The basis of potentiation is thought to be the same as that of brain seizure and learning. I have reduced these gated elemental functionalities with the "Halfgate" device (fig 1). The behavior of the Halfgate is determined by several inputs. There is one output. The Halfgate-Set is a combination of sensors providing inputs to the Halfgate, an actuator receiving an output from the Halfgate and the Halfgate. The actuator modifies a single material in a single location. The "Den", which models biological potentiation, is composed of Halfgate-Sets mimicking concentration gated pores (See Fig 1). Long-term and short-term memories are embodied in the concentrations of solutes. The Den model exhibits frequency/slope behavior like that seen experimentally. In learning simulations, employing a monolayer of Den-based neurons, challenge-induced misfiring of incidental neurons was scored. Long-term memory was demonstrated: misfiring decreased regarding each successive session-start. Short-term memory was demonstrated: within a session misfiring was reduced. First session misfiring at start 50%, end <1%; second session start 3.2%, end <0.1%; third session start 1.8%, end <0.01%. Simulating recruitment in seizure initiation, specific high frequency patterns of excitation caused >0.1% of neurons to fire continuously. Model neurons containing subunits other than the Den are described. Models of experience-modified potentiation, and environmentally and electrically-modified seizure induction are detailed. Details are given of how microcontrollers can be used to produce task-general model brains composed of randomly interconnected neurons, which are comprised solely of cascaded gated pores.
© Adkins; licensee BioMed Central Ltd. 2008
This article is published under license to BioMed Central Ltd.