| 摘要: | 本論文研究中，我們應用全息希爾伯譜分析(HHSA)方法來分析心電圖中的心律變異(HRV)。另外也使用典型相關(CCA)，分析了人類聽覺穩態響應(ASSR)，結合全息希爾伯譜分析(HHSA)。 典型相關分析(CCA)用於提取聽覺穩態響應(ASSR)相關信號特徵，並使用全息希爾伯譜分析(HHSA) 將提取的人類聽覺穩態響應(ASSR)，分解為調幅（AM）分量和調頻（FM）分量，其中FM頻率代表快速變化的模內頻率，AM頻率代表緩慢變化的模間頻率。
在心律變異(HRV)的分析當中，全息希爾伯譜分析 (HHSA) 打算將測量的心跳間隔 (IBI) 信號解釋為乘法振盪信號（載波）與調製信號（包絡波）的組合。瞬間採用頻率分析（IFA）計算頻率調製（FM）和幅度調製(AM) 每個時間點的頻率。整個測量 IBI 信號的 FM-AM 分佈概率累積並顯示在全息-希爾伯特譜 (HHS) 上。
我們招募了 20 名健康受試者參與我們的心電圖研究。每個受試者被要求觀看兩個視頻片段，包括一個放鬆的視頻片段和一個恐怖視頻剪輯。我們觀察到VHF頻段由~0.05 Hz AM 頻率調製在全息希爾伯譜(HHSA)分析上。VHF功率值顯著降低，觀看恐怖視頻剪輯的情況與觀看輕鬆視頻剪輯的情況相比。我們還發現在觀看恐怖視頻剪輯時，VHF和高頻 (HF) 頻段的功率都降低了。
而在ASSR的分析當中，我們目的在研究AM和FM光譜37赫茲，穩態聽覺刺激中的 人類聽覺穩態響應(ASSR)反應。招募了二十五名健康受試者參與這項研，每個受試者被要求參加兩次聽覺刺激的實驗，包括一隻右耳和一隻左耳，單耳的穩態聽覺刺激。在全息希爾伯譜分析 (HHSA)圖形中，37 Hz（基頻）和 74 Hz（一次諧波頻率）聽覺已成功提取響應。檢查 AM 頻譜、37 Hz 和 74 Hz 聽覺響應由不同的 AM 光譜調製，每個 AM 光譜至少具有三個複合頻率。與傳統傅里葉譜的結果相反，在 37 Hz 處出現分頻，並且在傅里葉光譜中，光譜峰在 74 Hz 處被遮擋。所提出的方法有效地糾正了時變幅度變化引起的分頻問題。
最後全息希爾伯譜分析(HHSA)，提供了一個全新的視角來分析心律變異(HRV)和穩態響應 (SSR)， 因此是一個有用的分析判讀工具，在穩態響應(SSR)中，可以減少避免，因傳統傅里葉頻譜中的幅度調製引起的錯誤、或誤導性的判讀。
關鍵字：典型相關分析、穩態響應、全息希爾伯譜分析
;In this thesis, we applied the Holo-Hilbert spectral analysis (HHSA) method to analyze heart rhythm variability (HRV) in ECG. Human Auditory steady-state response analysis (ASSR) Canonical Correlation Analysis (CCA) combined with Holo-Hilbert spectral analysis (HHSA) was also performed using canonical correlation analysis. CCA is used to extract ASSR-related signal features, and the extracted ASSR response is decomposed into amplitude modulation (AM) components and frequency modulation (FM) components using Holo-Hilbert spectral analysis (HHSA), where FM frequency represents a rapidly changing intra-mode frequency, AM frequencies represent slowly varying intermodal frequencies. In the analysis of heart rhythm variability (HRV), Holo-Hilbert spectral analysis (HHSA) intends to interpret the measured heartbeat interval (IBI) signal as a combination of a multiplicative oscillatory signal (carrier wave) and a modulating signal (envelope wave). Instantaneous frequency analysis (IFA) was used to calculate the frequency of frequency modulation (FM) and amplitude modulation (AM) at each time point. The FM-AM distribution probabilities of the entire measured IBI signal are accumulated and displayed on the Holographic-Hilbert Spectrum (HHS).
Twenty healthy subjects were recruited for our ECG study. Each subject was asked to watch two video clips, including a relaxing video clip and a horror video clip. We observed that the VHF band was modulated by ~0.05 Hz AM frequency on Holo-Hilbert spectral analysis(HHSA). VHF power values were significantly lower when watching a horror video clip compared to watching a relaxing video clip. We also found reduced power on both the VHF and high frequency (HF) bands when watching horror video clips.
In the ASSR analysis, we aimed to study the ASSR responses to 37 Hz, steady-state auditory stimuli in the AM and FM spectra.
Twenty-five healthy subjects were recruited to participate in this study, and each subject was asked to participate in two auditory stimulation experiments, including one right ear and one left ear, and a steady-state auditory stimulus for one ear. Hearing responses at 37 Hz (fundamental frequency) and 74 Hz (first harmonic frequency) were successfully extracted in the Holo-Hilbert spectral Analysis (HHSA) graph. Examining the AM spectrum, the 37 Hz and 74 Hz auditory responses are modulated by different AM spectra, each with at least three composite frequencies. Contrary to the results of conventional Fourier spectroscopy, frequency division occurs at 37 Hz, and in Fourier spectroscopy, the spectral peak is blocked at 74 Hz. The proposed method effectively corrects the frequency division problem caused by time-varying amplitude changes.
Finally, Holo-Hilbert spectral Analysis (HHSA) provides a new perspective to analyze HRV and Steady State Response (SSR), and thus is a useful analytical interpretation tool, in Steady State Response (SSR), avoidance can be reduced, since False, or misleading interpretations caused by amplitude modulation in conventional Fourier spectrum.
Keywords: Heart rhythm variability、Auditory steady-state response、Canonical correlation analysis, Holo-Hilbert spectral Analysis |
