Institutes outside Greece
Systematic evaluation of memory retrieval in computational brain microcircuits
Abstract
dc:descriptionSince ancient times, the human being could not stop wondering ‘‘Who am I?’’, “Why do I exist?”, “What is the purpose of life?”, “From where are all these thoughts coming?”. Our brain. This magnificent assemblage of neurons cannot stop wondering and exploring. But all these would mean nothing if they cannot be remembered. So, memory is needed. Memory is a function that defines who we are. Without our memories, we are beings without identity. Losing it can be very harmful to our survival. The centre of memory is the medial temporal lobe (MTL) and especially the hippocampus. Hippocampus is part of the limbic system and is related to short- and long-term memory (STM, LTM) functions. In my thesis, a more flexible brain-inspired computational model of the CA1 region of the mammalian hippocampus was upgraded and used to examine how the retrieval of information could be affected under different conditions. The retrieval process in a memory system plays a critical role in how the already stored information can be accessed. If the retrieval process is not efficient, it can lead to difficulties in accessing the required memory and eventually, have a negative impact on thinking and decision making. Additionally, understanding the retrieval process would help in developing strategies for improving memory and preventing memory loss. Brain waves have also a vital role in the proper function of circuits, especially the theta rhythm. Theta rhythm modulates hippocampal excitation and inhibition separating encoding and retrieval into two independent processes. The model consisted of different types of excitatory and inhibitory interneurons, which were enervated by excitatory and inhibitory inputs from neighbouring regions. Neurons were theta-modulated and had a preferred phase of theta. To examine which biophysical mechanisms could help the improvement of recall performance by removing spurious activity more accurately, the feedforward and feedback excitation which targets specific excitatory and inhibitory cells in the neural circuit is modulated. Six different models (models 1-6) were created by modulating different excitatory and inhibitory pathways: in model 1 the CA3 feedforward excitation to bistratified cells was increased, in model 2 the bistratified feedforward inhibition to pyramidal cells was increased, in model 3 the pyramidal feedback excitation to bistratified cells was increased, in model 4 a combination of models 1 and 2 was used, in model 5 a combination of models 1 and 3 was used, and in model 6 a combination of models 2 and 3 was used. Simulations showed that the increase in the strength of the feedforward excitation to inhibitory neurons was the most effective way to reduce or even eliminate spurious activity. That allows the system to access stored memories more accurately. The factors that damage the performance of the model were examined too. It was found that the number of neurons that decoded a memory affected the complexity of the context, and along with the number of neurons that decoded more than one memory (interference between memories) were the most significant factors which affected negatively the recall performance of the models. The mean recall quality (MRQ) was decreased when memories with more complex context were stored. Additionally, when the gained knowledge was used in less similar contexts, a significant decrease in the recall performance was noticed, whereas the similarity between already stored information and a newly one defined the performance of the models. On the other hand, the number of memories that stored in the network didn’t affect the recall performance, whereas a slight increase was observed when the size of the network was increased. Lastly, the number of inhibitory synapses affected the recall performance of the models only when both of them were reduced. On the contrary, the absence of oriens lacunosum-moleculare (OLM) inhibition didn’t affect the recall performance of the models under any condition. Because the results of my research will provide fundamental insights into the synaptic, cellular and network mechanisms of memory formation, this will lead to more efficient strategies for treatments and the prevention of memory-related disorders such as Alzheimer’s disease, schizophrenia and epilepsy.
Degree
thesis:*- Grantor dc:publisher
- Institutes outside Greece
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Andreakos, Nikolaos
- Ανδρεάκος, Νικόλαος
Subjects
dc:subject × 18- Νευρωνικά δίκτυα
- Μακρόχρονη μνήμη
- Μνημονική ανάκληση
- Μικροκυκλώματα εγκεφάλου
- Εγκέφαλος
- Βραχύχρονη μνήμη
- Neural networks
- Long-term memory
- Memory retrieval
- Brain microcircuits
- Brain
- Short-term memory
- Βασική Ιατρική
- Ιατρική και Επιστήμες Υγείας
- Νευροεπιστήμες
- Basic Medicine
- Medical and Health Sciences
- Neurosciences
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/55917