使用Ketamine麻醉臨床表徵以及藥理交互作用的應用 --以腦電波生理變化為主要研究內容

Analysis of General Anesthesia combined with Ketamine—mainly focused on Electroencephalogram

蔡奉芳(Feng-Fang Tsai)

指導教授 ： 范守仁

共同指導教授 ： 陳祈玲(Chih-Ling Chen)

國立臺灣大學/醫學院/臨床醫學研究所/博士(2020年)

https://doi.org/10.6342/NTU202004416

摘要

全身麻醉是由麻醉科醫師在開刀房中創造的可回復性昏迷狀態，與睡眠狀態非常不相同。手術當中使用麻醉深度監測儀已經被證實無法有效地分析病人的意識清醒程度以及避免手術後認知功能障礙；本人 將經驗模態分解方法應用在全身麻醉時期腦波的分析，我們利用此分析方法相較於傳統分析方法較為適合分析非線性以及非固定性生理訊號的優點，利用不同的運算方式做訊號處理以求得臨床上可觀察到的生理意義。首先，本人利用臨床上病人接受全身麻醉時所收集的腦波，運用經驗模態分析方法發現可以找到比較清晰的個別波形能量變化顯示方式，使用在接受propofol 進行全身麻醉的病患可以清楚分辨失去意識 (loss of consciousness) 前後的能量變化。 用線性呈現個別波形能量上的變化 對照傳統傅立葉有相同結果且更容易觀察，研究結果已發表在 Biomedical Physics and Engineering Express 。
接著，針對全身麻醉時期常使用的誘導藥物，探討不同藥理機轉的藥物所表現出腦波的能量差異。針對中樞神經解離劑ketamine其藥理作用為佔住大腦皮質，基底核，海馬迴的抑制性中間神經元受器，造成大腦皮質區與其他部分的不協調，主動性地造成連結混亂以及失去意識。此研究使用希伯特-黃轉換來處理腦波訊號。根據病人所收的全身麻醉分為 接受propofol靜脈麻醉組別以及sevoflurane氣體麻醉組別收案，藥物控制在病人臨床上失去反射後進行利用經驗模態分析以及連續性標準差量化個別頻域的腦波。
使用希伯特-黃轉換分析生理訊號，除了觀察時頻圖了解頻率分布的變化；更重要的是利用希伯特-黃轉換可以保留訊號的原始能量分布以及分析組成函數的瞬時頻率和能量 密度频 譜 來分析特定藥物造成的變化。這在臨床上並非創舉，根據Jianqiang Hu et.al在2015年將正常人的腦波與阿茲海默疾病病患的腦波比較，發現阿茲海默病患組別的腦波在側腦，顳葉以及枕葉區相較於正常人theta波增加了能量密度，在alpha波減少能量密度。量測非線性波形與線性波形的差異性可以做為變化比較的指標。用瞬時頻率變化來表示資訊的調頻改變方式nonlinearity-非線性性，表示亂度增加，用degree of wave distortion (DWD)計算瞬時頻率，量測非線性波形與線性波形的差異性。雖然因為副作用的關係，單一藥物的使用並不常見於臨床工作，但是單一藥物的臨床變化作為兩兩藥物比較的腦波變化至為重要。本研究為了證實臨床研究的數據正確性，增加單一注射propofol組別比較能量密度以及對非線性性的影響，以與並用藥物的組別來做比較；另針對誘導劑量的ketamine也進行能量密度的比較，探求單一藥物的臨床表現。研究腦波不同頻率成分間的關係叫做跨頻率耦合在腦波分析中被注意到，跨頻率耦合的目的是找出腦波低頻訊號中的相位以及高頻訊號中的振幅是否存在著同步或相位延遲關係。現在，認知神經科學家對於工作記憶的研究指出，當大腦需要維持工作記憶時，在顳葉皮質可發現 theta~ gamma 相位振幅耦合（theta 4~8Hz gamma 32Hz 以上）的現象；顯示此現象是大腦正在執行功能的關鍵角色。關於成人的工作記憶與跨顱交流電刺激之研究亦發現，原本表現較差的受試者，在施予跨顱交流電刺激後，表現明顯比電刺激前更佳。相位振幅耦合在麻醉學科上的應用，已經利用相位耦合與否來探討臨床上propofol造成的意識改變，可以觀察到 alpha的振幅最高點會離開 0~1 赫茲delta低頻的相位低谷；意識回復時亦然。本研究除了應證 theta~alpha相位振幅耦合變化發生在使用sevoflurane 麻醉誘導意識改變特徵，也找出使用 sevoflurane並用ketamine 時的相位振幅耦合變化，證實我們可以在麻醉過程中利用分析個別腦波分析病人使用的麻醉藥物。

關鍵字

腦波分析 ； 經驗模態分析 ； 麻醉深度監測 ； 相位振幅耦合 ； 希伯特黃轉換

並列摘要

The electroencephalography (EEG) is a noninvasive medical technique for monitoring and amplifying the electrical activity of cerebral. The most common anesthetic depth bi-spectral index (BIS) monitor was wildly used which recorded forehead EEG, decomposed with Fourier transform and processed as depth of anesthesia. The biggest disadvantage of BIS monitor is that the Fourier transform is not suitable for non-stationary and non-linear signal, a.k.a. EEG. The Hilbert–Huang Transformation(HHT) was proposed in this research in order to decompose EEG signals into intrinsic mode functions (IMF) because we believe HHT can work well with non-stationary and nonlinear data and preserve more messages from those signals. We applied this method in perioperative EEG signal analysis in order to quantify the energy shifting during general anesthesia.
In the first, we collected the patients who received general anesthesia whose EEG data recorded from bi-spectral index (BIS) monitor were collected. We decompose the raw EEG data from BIS monitor with Fourier transform and Hilbert-Huang Transform, and we compared the results of two different methods. With HHT method, it is easier to observe the energy changes of IMFs.
This research was focused on ketamine, because it was a dissociated sedative which was wildly used as an adjuvant in anesthesia; however, there are no monitors can monitor the depth of anesthesia while patients were used ketamine. This research wants to find the brain wave changes after ketamine was injected, and the interaction between different oscillations in EEG. In the part two research, we included patients with inhalation induction and intravenous induction. Under the same anesthesia depth, the experimental group use ketamine 0.5 mg.kg-1; while the control group using normal saline injection. This research used the sliding-window standard deviation to quantify the continuous energy change of each bandwidth EEGs. We found specific energy gathering after ketamine was injectied. Besides that, this research invented different methods to evaluate the intra-wave frequency modulation and amplitude modulation in individual brain waves, as well as phase-amplitude coupling. We applied those flourishing methods on EEG to find more details of neurophysiological changes during general anesthesia.
In the third part research, intravenous induction with propofol combined with ketamine or alfentanil were included. For the purpose to characterizing the dynamic properties of EEG, time-frequency analysis is commonly used to study the dynamics of an EEG signal in both time and frequency domains simultaneously. The changes in power densities of the first 3 IMFs for two groups are totally different. The normalized power density of IMF2 after anesthesia induction is significantly different from that of baseline. This implies that the normalized power density of IMF2 is a potential assessment correlated to the level of consciousness. These results imply that two anesthesia agents monitor the depth of anesthesia while patients were used ketamine. These results imply that two anesthesia agents cause different responses in EEG signals, which reflect different underlying physiological mechanisms in a human brain.
Finally, we used phase-amplitude coupling method to find the differences between anesthetic medications. The phase for different combination of coupling components showed significant changes in anesthetic-induced unconsciousness. The coupling between the delta-band phase and the theta-band amplitude changed from in-phase to out-phase coupling during the induction process from consciousness to unconsciousness. The changes in the coupling phase in EEG signals were abrupt and sensitive in anesthetic-induced unconsciousness.
The understanding of differences between anesthetic agents can contribute to the safety of drug dose and drug choice. This research used new method in order to improve the anesthetic depth monitoring in operation theater.

並列關鍵字

Electroencephalography ； Hilbert-Huang Transformation ； empirical mode decomposition ； phase-amplitude modulation

參考文獻

1. A.A. Petrosiana, D. V. P., W. Lajara-Nansona, R.B. Schiffera (2001 ). "Recurrent neural network-based approach for early recognition of Alzheimer's disease in EEG." Clinical Neurophysiology 112 1378-1387.

2. Aho, A. J., K. Kamata, V. Jantti, A. Kulkas, S. Hagihira, H. Huhtala and A. Yli-Hankala (2015). "Comparison of Bispectral Index and Entropy values with electroencephalogram during surgical anaesthesia with sevoflurane." British Journal of Anaesthesia 115(2): 258-266.

3. Akeju, O., S. E. Kim, R. Vazquez, J. Rhee, K. J. Pavone, L. E. Hobbs, P. L. Purdon and E. N. Brown (2016). "Spatiotemporal Dynamics of Dexmedetomidine-Induced Electroencephalogram Oscillations." PLoS One 11(10): e0163431.

4. Akeju, O., M. B. Westover, K. J. Pavone, A. L. Sampson, K. E. Hartnack, E. N. Brown and P. L. Purdon (2014). "Effects of sevoflurane and propofol on frontal electroencephalogram power and coherence." Anesthesiology 121(5): 990-998.

5. Berger, M., J. B. Mark and M. Kreuzer (2020). "Of Parachutes, Speedometers, and EEG: What Evidence Do We Need to Use Devices and Monitors?" Anesth Analg 130(5): 1274-1277.