靜息態功能性連結與功率頻譜之不同分析方法比較

Evaluations of different analyses in resting-state functional connectivity and power spectrum

洪于涵(Yu-Han Hong)

指導教授 ： 翁駿程

中山醫學大學/醫學科技學院/生物醫學科學學系碩士班/碩士(2013年)

https://doi.org/10.6834/CSMU.2013.00004

摘要

靜息態功能性磁振造影（Resting state functional magnetic resonance imaging, rs-fMRI）近來的發展逐漸受到重視，但在先前的文獻中使用的影像分析方式不盡相同，本研究主要目的為探討rs-fMRI中可信度及穩定性較高的分析方式，以期未來廣泛應用於臨床疾病的研究。
　　實驗中我們採用人腦（n=18）及鼠腦（n=5）的靜息態功能性影像進行分析，分別對人腦數據進行獨立成分分析（Independent components analysis, ICA）、MNI種子點感興趣區分析（MNI seed ROI analysis）及任務刺激種子點感興趣區分析（Task-based seed ROI analysis）；鼠腦數據則進行ICA及手動種子點感興趣區分析（Manual seed ROI analysis）。以這幾種不同的分析方法，來分析腦功能區之間的功能性連結（Functional connectivity）與其功率頻譜（Power spectrum），藉此比較不同分析方式之優劣，再進一步觀察人腦與鼠腦的分析趨勢是否具有一致性。
　　我們首先以task-based seed ROI analysis的結果作為黃金標準與MNI seed ROI analysis進行比較，發現這兩種分析方式之間的結果差異並不大，因此我們認為MNI seed ROI analysis應用在大量的臨床研究上結果是可信的。ICA的結果較為客觀，但其自動計算出的腦神經網路區可能會因為包含到非主要觀察的腦功能區而影響到分析結果。雖然MNI seed ROI analysis比ICA還要主觀，但其可細分出較精細的腦功能區並提供腦功能區之間的相關性（Correlation）。這樣的結果表現在人腦與鼠腦影像分析上趨勢是一致的。
　　結果顯示rs-fMRI的各種分析方式各有其優缺點，我們建議於研究中可先使用ICA尋找未知的腦神經網路區，若已有確定的腦神經網路區，可改用MNI seed ROI analysis做更進一步的探討。而選擇適當的rs-fMRI影像分析方式，未來將可更廣泛應用於人腦或動物臨床疾病的研究上。

關鍵字

靜息態功能性磁振造影 ； 獨立成分分析 ； 種子點感興趣區分析 ； 功能性連結 ； 功率頻譜

並列摘要

Resting-state functional magnetic resonance imaging (rs-fMRI) is recently a major direction in clinical studies, but various kinds of image analyses were used in the previous reports. Therefore, the main purpose of this study was to find a high stable and reliable analysis method in resting state fMRI. Specifically, we tried to compare the power spectrum and region correlation of independent components analysis and seed ROI analysis. The stable and reliable analysis method in resting state fMRI was expected to be widely used in the clinical disease research.
In this study, we performed resting-state fMRI in the human brain (n = 18) at 1.5 tesla MRI scanner and in the rat brain (n = 5) at 7.0 tesla MRI scanner. Independent components analysis, MNI seed ROI analysis, and task-based seed ROI analysis were then used in the human brain data; independent components analysis and manual seed ROI analysis were used in the rat brain data, respectively. The functional connectivity between brain regions and the power spectrum were obtained in these various analyses. The advantages and disadvantages of these analyses were discussed, and the results obtained between the human brain and the rat brain were compared and discussed.
In human brain, the result of the task-based seed ROI analysis was used as the gold standard and it was compared with MNI seed ROI analysis. We found there was no significant difference between these two analyses. Therefore, we believed the MNI seed ROI analysis was reliable to be performed in large numbers of clinical researches. The result of ICA was more objective, but it could be affected by other functional brain regions classified into the interesting brain networks. Although, MNI seed ROI was more subjective than ICA, the information of the correlations between the brain sub-regions could be obtained. The result found in the rat brain was consistent with the human brain.
Our results showed there were advantages and disadvantages in the different analyses of rs-fMRI. We suggested ICA could be first used to find the alteration in unknown brain network in the study. For specific brain network, MNI seed ROI analysis could be performed further. Appropriate image analysis methods will be useful in the clinical disease of both human and animal studies.

並列關鍵字

Resting-state functional magnetic resonance imaging (rs-fMRI) ； independent components analysis ； seed ROI analysis ； functional connectivity ； power spectrum

參考文獻

3. 謝政學。(2012) 在Resting-state fMRI下以Seed-Based分析評估不同磁場、掃描時間和Seed ROI半徑造成的功能性連結變化。中山醫學大學生物醫學科學學系碩士論文。104頁。

4. Abler, B., Erk, S., Herwig, U., & Walter, H. (2007). Anticipation of aversive stimuli activates extended amygdala in unipolar depression. Journal of psychiatric research, 41(6), 511-522.

6. Aso, T., Urayama, S. I., Fukuyama, H., & Le Bihan, D. (2013). Comparison of diffusion-weighted fMRI and BOLD fMRI responses in a verbal working memory task. NeuroImage, 67, 25-32.

7. Bai, F., Zhang, Z., Yu, H., Shi, Y., Yuan, Y., Zhu, W., ... & Qian, Y. (2008). Default-mode network activity distinguishes amnestic type mild cognitive impairment from healthy aging: a combined structural and resting-state functional MRI study. Neuroscience letters, 438(1), 111-115.

8. Bandettini, P. A., Wong, E. C., Jesmanowicz, A., Hinks, R. S., & Hyde, J. S. (1994). Spin-echo and gradient-echo EPI of human brain activation using BOLD contrast: a comparative study at 1.5 T. NMR in Biomedicine, 7(1-2), 12-20.