Network requirements for parametric working memory

点击翻译按钮获取中文摘要

DESCRIPTION provided by applicant): The proposal is designed to enable the candidate to become a mature, accomplished researcher in the field of theoretical and computational neuroscience, through individual mentoring, participation in courses and seminars, and particularly through intensive training to complete a high-quality research project. The candidate has produced excellent research in theoretical physics, but requires a period of development to become an independent, productive scientist in the field of neuroscience. He has chosen an institution with an established graduate program in neuroscience, and with faculty of high acclaim in order to gain the best quality of training. The research project will address the cellular mechanisms of working memory and so help elucidate a key process in cognition. The project focuses on the mechanisms of parametric working memory, whereby a quantity (such as frequency) with a continuous range of values can be encoded and memorized in the firing rates of neurons. The candidate will implement a computational model network, containing neurons with biophysical properties, which will reproduce specific sets of data in a parametric working memory task. He will test what profiles of synaptic connectivity can produce the necessary stable network states for parametric working memory, hypothesizing that the strength of synaptic connections to a neuron correlates with the excitability of the neuron. He hypothesizes that long-term plasticity is necessary to achieve the synaptic connectivity necessary for working memory, so will test which kinds of long term plasticity stabilize the network. He hypothesizes that short-term plasticity is necessary to produce time variation in the memory states, as seen in experiment. He will test which forms of short-term plasticity lead to the observed behavior, and whether they contribute to the stability of the mnemonic states.