腦波非線性訊號分析--以阿滋海默症為例

Abstract

在此研究中，我們試圖以渾沌(chaos)理論為基礎，Sampling entropy為方法，來分阿滋海默症Alzheimer’s disease (AD)不同時期的腦波複雜度，以提供臨床新的客觀工具，以利AD的診斷與預後評估。 失智症是一認知功能逐漸惡化的疾病，且AD佔了其中最大部份，且隨著年齡的增長，其發生的機率也大幅上升，因此，隨著台灣人口老人化的現象，AD病患也預期會增加。但是，AD的臨床診斷，常因病情進展的緩慢，而延後了2~3年。因此，找到客觀方便的方式，來達到及早診斷及早治療的目的，變的刻不容緩。 目前，臨床上對於AD的診斷，是按照DSM-IV（Diagnostic and Statistical Manual of Psychiatric Disorders 4th ed）診斷標準診斷。雖有如此詳盡的評估，早期AD依舊不容易在早期進行客觀的評估。雖然，有些功能性腦部影像檢查，如functional MRI, positron emission tomography (PET), and single photon emission computed tomography (SPECT)，可以提供部分的訊息，以作為臨床診斷的參考，但是，限制於高單價，輻射暴露，及顯影劑注射引起的過敏反應，都使得這些檢查並非極度方便。而腦波圖(Electroencephalography, EEG)為一價格低廉，非侵入性的檢查方式，可以快速取得人類腦部活動的紀錄，且其臨床使用已超過五十年，為一安全性極高的檢查。 但由於過去，對於腦波圖的判讀，是需要倚靠有經驗的神經內科專科醫師，且其對於失智症的評估，只能提供局部慢波的訊息，對於診斷及追蹤，實無多大助益；但是，近年來，由於電腦的發達，及數位訊號處理的進步，數位腦波分析（quantitative EEG）的概念，例如使用傅立葉轉換（Fourier analyses），開始被提出並應用於臨床失智病人腦波的分析。但是，這些基於線性系統的數學模式分析方式，卻只能提供病人與正常人之間的差異，並無法有效提供進一步的應用。而基於非線性系統的近似熵（approximate entropy），可以分析訊號的複雜度，訊號越複雜，其值越高。是在數學模式上比較符合人類生理訊號的非線性特徵，在過去的研究中，也發現在AD的病人中，其雙側顳葉的近似熵會有下降的趨勢。但是，單純的近似熵卻會受到資料的長短，及其潛在的趨勢，而得不到一個穩定的結果，而限制了其應用。在此方面，採樣熵（sampling entropy）則是避免了資料長短的誤差，再藉由此研究中，利用經驗解構模式，來去除潛在的趨勢，使得所得結果更能符合臨床資料，並有機會使腦波圖可以成為一客觀評估失智症的工具。 另一個問題為臨床醫生依舊無法在治療個別病人時，事前預見其治療效果。在這篇論文中，我們試圖用多尺度熵（multiscale entropy）來分析對於乙醯膽鹼酯酶抑製劑治療有效及無效病患腦波圖的差異。以期在治療的初期即能為病患訂定有效的治療方針。 腦波圖，通過非線性數位訊號處理的幫助下，腦波圖的潛藏信息可以被提取用於臨床評估，追蹤，甚至預測治療效果。

In this dissertation, we will develop the computational tools, on the basis of sampling entropy, to evaluate different patterns of complexity activation of the electroencephalography (EEG) in longitudinal changes in the Alzheimer's disease (AD) after the acetylcholinesterase inhibitor (AChE inhibitor). Alzheimer’s disease is the most common form of dementia. The cause and progression of AD are not well understood. One hypothesis is that AD is caused by reduced synthesis of the neurotransmitter, acetylcholine (ACh). AChE inhibitors are proved as an effective therapy. The diagnosis and evaluation in the early stage dementia is challenging in clinical medicine. Quantitative electroencephalographs (qEEG) provide a potential method to objectively quantify the cortical activations in AD, but they are too insensitive to probe the alteration of EEG in the early AD. The approximate entropy, which is a non-linear statistic and is able to quantify the irregularity of a time series, was significant lower in the bilateral temporal region in AD patients, in whom the basic pathology is hippocampus atrophy. But there were some bias in approximate entropy, such as inconsistent results, and depending on the data length. Therefore, in order to evaluate different patterns of complexity activation of the EEG in longitudinal changes in AD after the acetyl cholinesterase inhibitor therapy, it is necessary to develop a better method with the sampling entropy. However, a technical issue which has been ignored by most researchers is that the signal should be stationary. In order to resolve the non-stationarity of SaEn in EEG to improve the sensitivity, an empirical mode decomposition (EMD) was applied for detrending in this dissertation. Twenty-seven AD patients (9M/18F; mean age 74.0±1.5 years) were included. Their initial Minimal Mental Status Examination was 19.3±0.7. They received the first resting awake 30-mine EEG before the therapy. Five of them received a follow-up examination within 6 months after the therapy. The 30-s EEG data without artifacts were selected and analyzed with a new proposed method,“EMD-based detrended-SaEn” to attenuate the influence of intrinsic non-stationarity. The correlation factors in 27 AD patients showed a moderate correlation (0.361-0.523, p < 0.05) between MMSE and EMD-based detrended SaEn in Fp1, Fp2, F4 and T3. There was a high correlation (Correlation coefficient = 0.975, p < 0.05) between the changes of MMSE and the changes of EMD-based detrended-SaEn in F7 in 5 follow-up patients. The dynamic complexity of EEG fluctuations is degraded by pathological degeneration, and EMD-based detrended SaEn provides an objective, non-invasive and non-expensive tool for evaluating and following AD patients. The other issue, for clinician, it is still not predictable effect in individual patient. In this dissertation, we tried to use multiscale entropy (MSE) in EEG to predict the efficacy of AChE inhibitor. Seventeen newly diagnosed AD patients (9M/8F; mean age 74.6±7.4 years) were enrolled in this study, with an initial MMSE of 18.8±4.5. After 12 months’ therapy of AChE inhibitor, 7 patients (3M/4F; mean age 76.1±7.9 years) were responsive (responder) and 10 patients (6M/4F; 73.5±7.3 years) were non-responsive (non-responder). The major difference between two groups is Slope2 (MSE6 to 20). The area under curve (ROC curve) of Slope2 is 0.871(95% CI = 0.69 - 1). The sensitivity is 85.7% and the specificity is 60% while the cutoff value of Sloep2 is -0.024. MSE of EEG, especially Slope2, is able to be an objective tool to predict the efficacy of AChE inhibitor before the therapy. By the assistance of non-linear digital signal processing, the embedded information of EEG in AD could be extracted for the clinical evaluation, following up and even prediction the therapeutic effect.