本文關(guān)鍵詞：神經(jīng)元的發(fā)放閾值及能量效率研究　出處：《蘭州大學(xué)》2016年博士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 神經(jīng)元 閾值 分界線 信息率 能量效率

【摘要】：神經(jīng)元是神經(jīng)系統(tǒng)實(shí)現(xiàn)其功能的基本單元。理解神經(jīng)元的功能是理解大腦智能來源的基礎(chǔ)。神經(jīng)元的功能簡單來說就是整合接收到的大量信息,然后決定是否產(chǎn)生輸出信號(hào)。神經(jīng)元的輸出信號(hào)稱為動(dòng)作電位,又稱發(fā)放。一般認(rèn)為,神經(jīng)元細(xì)胞膜兩側(cè)電勢(shì)差超過一個(gè)閾值就會(huì)產(chǎn)生動(dòng)作電位。所以神經(jīng)元發(fā)放的閾值在神經(jīng)元的信息整合中起著關(guān)鍵作用。實(shí)驗(yàn)廣泛發(fā)現(xiàn),動(dòng)作電位的閾值是可變的,而且這種可變性對(duì)神經(jīng)元處理信息有重要影響。目前,關(guān)于可變閾值的來源并沒有統(tǒng)一的意見,甚至引起不小的爭議。神經(jīng)元的發(fā)放要消耗大量的能量。發(fā)放耗能在神經(jīng)系統(tǒng)的總耗能中占據(jù)了很大的比例。神經(jīng)系統(tǒng)處理信息是極其耗能的過程,然而對(duì)于動(dòng)物的生存至關(guān)重要,所以其發(fā)放的能量效率可能在漫長的進(jìn)化過程中已經(jīng)得到了優(yōu)化。神經(jīng)元發(fā)放的閾值決定著神經(jīng)元如何編碼信息,而發(fā)放的能量效率則關(guān)系著編碼信息的效率。本文第一章介紹本文課題的研究背景;第二章中總結(jié)闡述神經(jīng)元及動(dòng)作電位的生物學(xué)基礎(chǔ)知識(shí);第三章給出常見的神經(jīng)元模型及后面分析中需要用到的動(dòng)力學(xué)基礎(chǔ)知識(shí)。第四章詳細(xì)討論神經(jīng)元發(fā)放的閾值問題。我們提出閾值現(xiàn)象可以按機(jī)制不同分為“參數(shù)閾值”和“狀態(tài)閾值”,通常所說的閾值是一種狀態(tài)閾值,是由狀態(tài)空間中的“廣義分界線”(簡稱分界線)決定的。我們認(rèn)為分界線普遍地存在于神經(jīng)元模型的狀態(tài)空間中,通過構(gòu)建一個(gè)一般性的可激神經(jīng)元模型,我們得到了普適的分界線表達(dá)式,進(jìn)而得到普遍的閾值隨時(shí)間演化的方程。閾值演化方程在相應(yīng)條件下可以很自然地約化得出與前人的工作一致的結(jié)果,而在前人工作中一般是直接假設(shè)或者做了一些不太自然的簡化才能得到。在此基礎(chǔ)上,我們的神經(jīng)元?jiǎng)恿W(xué)研究還揭示,閾值電壓在不同刺激下的變化是由于系統(tǒng)在狀態(tài)空間中跨越分界線上的不同點(diǎn)造成的,神經(jīng)元的分界線和刺激條件決定了閾值電壓變化的范圍。狀態(tài)空間中的分界線跨越機(jī)制是電壓閾值和刺激后的參數(shù)閾值的普遍的內(nèi)在動(dòng)力學(xué)機(jī)制。我們還系統(tǒng)地在從一維到四維的多個(gè)模型中檢驗(yàn)了我們的結(jié)果,其中二維和三維的模型中解析的分界線,將對(duì)以后的閾值研究有重要意義,而對(duì)有著現(xiàn)實(shí)對(duì)應(yīng)的四維的經(jīng)典Hodgkin-Huxley模型的研究也發(fā)現(xiàn)了一些有意思的新現(xiàn)象。分界線跨越理論為閾值可變性提供了一個(gè)普適的機(jī)制,是閾值可變性的一個(gè)重要來源。在第五章中,我們研究神經(jīng)元發(fā)放的能量效率,也就是一個(gè)神經(jīng)元在處理信息時(shí)消耗單位能量能夠傳遞的信息量。我們計(jì)算了Hodgkin-Huxley神經(jīng)元在有噪聲的環(huán)境中不同溫度不同刺激強(qiáng)度下的信息率和能量效率。我們發(fā)現(xiàn)在特定的溫度下,信息率或能量效率最大。盡管信息率和能量效率不能同時(shí)達(dá)到最大化,但是我們也發(fā)現(xiàn)神經(jīng)元在最優(yōu)能量效率的溫度下保持了較高的信息處理能力。我們認(rèn)為處理信息的能量效率可能對(duì)神經(jīng)系統(tǒng)進(jìn)化起到更重要的作用。最后,我們對(duì)目前的工作做了總結(jié),并展望未來相關(guān)的可能的研究方向。
[Abstract]:Neuron is the basic unit of the nervous system to realize its function. Understanding of neuronal function is the base of understanding the brain intelligence sources. A large amount of information the function of neurons is simply the integration of the received, and then decide whether to generate an output signal. The output signal of neurons called potential, also called release. Generally, the neuronal membrane on both sides the potential difference exceeds a threshold will produce action potentials. So neuron threshold neurons in the integration of information plays a key role in the experiment. Widely found that action potential threshold is variable, and the variability of neuronal processing has an important influence on the source of information. At present, the variable threshold and no unified opinion, even caused no small controversy. The neuron spikes to consume large amounts of energy. Energy consumption in the total energy release of the nervous system occupies a large The proportion of the nervous system. Information processing is extremely energy, however, is vital for the survival of the animal, so the energy efficiency issue may have been optimized in the long evolutionary process. Neuron threshold determines how neurons encoding information, and put the energy efficiency affects the efficiency of the first encoding information. This chapter introduces the background of the research topic; the second chapter summarizes basic knowledge of biology and neuronal action potential; need to use the basic knowledge of mechanical neuron model and the third chapter gives the analysis of the common behind. The fourth chapter discusses neuron threshold problem. We propose a threshold phenomenon can be divided into "according to the mechanism of different parameter threshold" and "state threshold", usually said threshold is a state of the threshold is determined by the state space in the "general line" ( Referred to as the dividing line) decision. We believe that state space boundaries generally exist in the neuron model, through the construction of a generic excitable neuron model, we obtain the general boundary expression of pain, and the time evolution equation is obtained. The universal threshold threshold evolution equation can naturally get reduced with previous work consistent results in the corresponding conditions, while in the previous work in general is a direct assumption or do some of the less natural can be simplified. On this basis, we also study on neuron dynamics reveals the change of threshold voltage in different stimuli are due to different points in the state space system across the dividing line between the neuron boundaries and stimulation conditions determine the range of threshold voltage change. The dividing line in the state space is across the threshold and stimulating mechanism After the threshold parameter universal intrinsic dynamic mechanism. We also systematically in multiple models from one-dimensional to four-dimensional in our test results, the analytical model of 2D and 3D in line, the threshold for future research has important significance, but also found some new interesting phenomenon of study on a classical Hodgkin-Huxley model corresponding to the four-dimensional reality. The dividing line across the theory provides a universal mechanism for threshold variability is an important source of threshold variability. In the fifth chapter, we study the energy efficiency of neuron is a neuron, when processing the information consumption information unit energy transfer. We can calculate the temperature under different stimulus intensities in the noisy environment of the information rate and energy efficiency of Hodgkin-Huxley neurons. We found that at a specific temperature Under the maximum information rate or energy efficiency. Although the information rate and energy efficiency can reach the maximum at the same time, but we also found that neurons maintain information processing ability of higher energy efficiency in the optimal temperature. We believe that the energy efficiency of processing information may be more important to the evolution of the nervous system. Finally, we a summary of the present work and look forward to the future possible research direction.

【學(xué)位授予單位】：蘭州大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2016
【分類號(hào)】：Q42

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 CHEN YueLing;YU LianChun;CHEN Yong;;Reliability of weak signals detection in neurons with noise[J];Science China(Technological Sciences);2016年03期

，

本文編號(hào)：1378204