大腦視丘神經元之連發波訊號的自動偵測與分類; Automatic Detection and Classification of Bursts in Brain Thalamus Neurons

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/43902

Title:	大腦視丘神經元之連發波訊號的自動偵測與分類;Automatic Detection and Classification of Bursts in Brain Thalamus Neurons
Authors:	歐靜文;Jing-wen Ou
Contributors:	數學研究所
Keywords:	神經訊號;連發波;burst;nerve impulse
Date:	2010-06-25
Issue Date:	2010-12-08 14:26:03 (UTC+8)
Publisher:	國立中央大學
Abstract:	在神經訊號中，最常見到的是單一棘波訊號（spikes），而生物學家在大腦視丘裡紀錄到另一種與 spikes 不同的訊號，稱之為連發波訊號（bursts）。本文主要在探討 bursts 訊號的特徵，並將特徵具體量化，以利於自動偵測以及分類 bursts 訊號。首先我們先介紹神經細胞如何產生動作電位，以及神經訊號的相關背景知識。然後，根據生物學家憑經驗挑選出來的 bursts 訊號，我們歸納其特徵並建立了三個篩選條件。這三個篩選條件分別是：與時間相關的「間隔條件」、與振幅相關的「遞減條件」、以及與波形相關的「形似條件」。有了篩選條件之後，接著將這些條件實際運用到一筆原始訊號中去偵測 bursts。在探測訊號之前，會先對原始訊號做一些基本的前置處理，包含降低取樣頻率以及>濾波。前置處理完後，利用上述三個篩選條件來篩選 bursts 訊號，並利用主成份分析（簡稱 PCA）為 bursts 訊號做分類。用生物學家憑經驗挑選出來的 bursts 訊號做測試，經由篩選條件過濾後的結果是被生物學家所接受的。而篩選條件自原始訊號中所偵測到的 bursts 訊號，也通過了生物學家的檢驗。由此可以看到篩選條件是能夠反應 bursts 訊號的特徵，並有效的偵測出 bursts 訊號。 The most common signals found in nerve signals are isolated spikes. However, biologists have detected signals other than spikes in the thalamus. These are known as bursts. This paper will mainly be examining the characteristics of bursts. Moreover, the data found will be quantified in order to automatically detect and categorize bursts. First, we will introduce how nerve cells generate action potentials and the background information of nerve signals. Next, according to the bursts selected by a biologist through experience, we will generalize the characteristics and establish three screening conditions. These three screening conditions are as follows: associated with time, the gap condition, associated with amplitude, the decay condition, and associated with waveforms, the shape condition. With these conditions, we will then apply them on to a set of raw data to detect bursts. Before detecting the signals, we will first process the raw data. This includes down sampling and filtering. After processing the raw data, we will automatically detect the bursts using the three screening conditions mentioned above. In addition, using the Principal Component Analysis (PCA), we will then classify the bursts. Testing with the bursts that the biologist has selected based on his experience, the results collected through the filtering is confirmed. Moreover, the bursts detected from the raw data using the filtering criteria also pass the test. Therefore, we can see that the filtering criteria can reflect the characteristics of bursts and effectively detect them.
Appears in Collections:	[Graduate Institute of Mathematics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	677	View/Open

社群 sharing

Loading...