Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Simulating Facilitation in a Spiking Neural Network

Kuzmin, Alexander (2020) MOBK01 20201
Degree Projects in Molecular Biology
Abstract
Many insects are able to detect, isolate and track small targets that move quickly against a dynamic background. This ability is enabled by a group of neurons called small target motion detectors (STMD). These neurons, among other properties, have a type of short term memory called response facilitation. This type of facilitation is a phenomenon in which continued visual stimuli lead to an enhanced response, whereas single signals lead to little or no transmission of the signal. In previous studies on dragonflies, studies have shown that the facilitation waves have an intrinsic ability to travel throughout the neural networks even after the stimulus has stopped, which has not been simulated yet. The goal of this project was to examine... (More)
Many insects are able to detect, isolate and track small targets that move quickly against a dynamic background. This ability is enabled by a group of neurons called small target motion detectors (STMD). These neurons, among other properties, have a type of short term memory called response facilitation. This type of facilitation is a phenomenon in which continued visual stimuli lead to an enhanced response, whereas single signals lead to little or no transmission of the signal. In previous studies on dragonflies, studies have shown that the facilitation waves have an intrinsic ability to travel throughout the neural networks even after the stimulus has stopped, which has not been simulated yet. The goal of this project was to examine which neuron and synapse models are currently available, then use those models to construct simulations of neural networks that could potentially support travelling facilitation waves in the ways that mimic the results of the in vivo studies. After a series of simulations the results show that even though NEST simulator currently has a number of synapse and neuron models that support facilitation, none of them could currently support a model with a travelling facilitation wave. The NEURON simulations were promising but overall proved to be inconclusive and required further experimentation. (Less)
Popular Abstract
Focus: Beyond the Hocus Pocus

Imagine being a dragonfly. Now that would be pretty cool! Sitting on a branch, enjoying the sun, picking out which other nearby insect you would like to turn into a delicacy next. There’s so many of them, too - hundreds of thousands of small flies, mosquitoes and moths flying around in erratic fashion. One could be forgiven for having trouble picking out just one of them. It’s like when you’re in a store and trying to pick a product that is the best exemplar of a bunch of the exact same item, but in this case that “product” flies around like a starfighter from a certain cinematic franchise. It would probably be very hard to survive for very long as a species without acquiring certain adaptations over time... (More)
Focus: Beyond the Hocus Pocus

Imagine being a dragonfly. Now that would be pretty cool! Sitting on a branch, enjoying the sun, picking out which other nearby insect you would like to turn into a delicacy next. There’s so many of them, too - hundreds of thousands of small flies, mosquitoes and moths flying around in erratic fashion. One could be forgiven for having trouble picking out just one of them. It’s like when you’re in a store and trying to pick a product that is the best exemplar of a bunch of the exact same item, but in this case that “product” flies around like a starfighter from a certain cinematic franchise. It would probably be very hard to survive for very long as a species without acquiring certain adaptations over time that would help you with this hectic lifestyle.

Adaptation to the surrounding environment is one of the more fundamental processes that improves an organism’s survival chances. Being able to change the way a part of a body functions on the fly allows the animal to be prepared to face more potentially hazardous situations. This study focused on one of the mechanisms that allows insects like dragonflies and hoverflies to be more effective at finding prey and potential mating partners by using selective attention. Tracking small, quickly moving targets against a spotty background can be quite a monumental task, especially considering that the background is probably also moving quickly - because you’re in mid-flight. The fact that you’re an insect and your visual systems aren’t all that fly don’t help either.

In order to alleviate the constant sensory bombardment, dragonflies and some other insects use a group of nerve cells called small target motion detectors (STMD). These neurons specifically react to movements of small objects, while completely ignoring anything larger than themselves (which does make sense, considering that other critters of the same or smaller size will either be food or a potential date). One of the things that these neurons are capable of is called facilitation, which is the ability to enhance the stimulatory signal, if the frequency of the signal is high enough. During live experiments with dragonflies and hoverflies a very peculiar facilitation pattern has been detected - it appeared to move throughout the brain in a wave-like motion. The specific goal of this study was to try to simulate that specific facilitation wave pattern using the currently available digital tools, like NEST simulator and NEURON simulator.

Unfortunately, while some facilitation models were successfully created during the project, the exact facilitation patterns seen during live experiments remain elusive. It appears that there are currently no models that are advanced enough to create a travelling facilitation wave within the NEST simulator platform. Simulations with the NEURON simulator appear to be more promising due to the platform being more robust and models more flexible. In fact, some facilitation-like effects were achieved using a graded AMPA synapse, but further simulations using the hypothetically compliant models with NMDA and AMPA neurotransmitter synapses are needed.

Examensarbete för kandidatexamen i Biologi/Molekylärbiologi 15 hp 2020;
Biologiska institutionen, Lunds Universitet.
Handledare: Bo Bekkouche & David O’Carroll.
Avdelning: Functional Biology. (Less)
Please use this url to cite or link to this publication:
author
Kuzmin, Alexander
supervisor
organization
course
MOBK01 20201
year
type
M2 - Bachelor Degree
subject
language
English
id
9020148
date added to LUP
2020-06-18 11:45:14
date last changed
2020-06-18 11:45:14
@misc{9020148,
  abstract     = {{Many insects are able to detect, isolate and track small targets that move quickly against a dynamic background. This ability is enabled by a group of neurons called small target motion detectors (STMD). These neurons, among other properties, have a type of short term memory called response facilitation. This type of facilitation is a phenomenon in which continued visual stimuli lead to an enhanced response, whereas single signals lead to little or no transmission of the signal. In previous studies on dragonflies, studies have shown that the facilitation waves have an intrinsic ability to travel throughout the neural networks even after the stimulus has stopped, which has not been simulated yet. The goal of this project was to examine which neuron and synapse models are currently available, then use those models to construct simulations of neural networks that could potentially support travelling facilitation waves in the ways that mimic the results of the in vivo studies. After a series of simulations the results show that even though NEST simulator currently has a number of synapse and neuron models that support facilitation, none of them could currently support a model with a travelling facilitation wave. The NEURON simulations were promising but overall proved to be inconclusive and required further experimentation.}},
  author       = {{Kuzmin, Alexander}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Simulating Facilitation in a Spiking Neural Network}},
  year         = {{2020}},
}