通過腦組織脈動估算大腦性質

Abstract

由於難以獲得無創的體內數據，理解腦組織的機械性質仍然具有挑戰性。本研究通過利用心臟周期引起的腦組織脈動（BTP）來估算腦部性質。研究開發了一個耦合固-液的數學模型，並使用經顱組織多普勒（TCTD）技術測量20名健康個體的腦組織位移和相應的血壓，從而獲得數據。傅里葉變換被用來推導位移和壓力信號之間的傳遞函數。 該數學模型假設腦組織可以被模擬為一個耦合固-液系統。球坐標系被應用於模型中以簡化控制方程。為了擬合數據，嘗試了兩種模型，包括耦合固-液模型及其具有多個組分的修正模型。使用MATLAB的'fminsearch'進行模型擬合，優化了包括楊氏模量（E）、泊松比（ν）、比儲量（Q）和透氣度/粘度（κ/μ）等關鍵參數。 總共嘗試了四種方法。前三種方法包括原始模型擬合、為避免負量級進行的對數變換模型擬合和修正模型擬合，這些方法擬合了由物理參數組成的無量綱參數組。然後，無量綱參數組可以計算出三個量綱參數。最後一種方法是直接擬合原始三個量綱參數，因為在無量綱參數組轉化為三個量綱參數的過程中存在問題。 擬合結果顯示所有方法的擬合曲線和損失都與實驗數據有很強的相關性。然而，參數值並不總是符合預期的量級。這種差異部分是由於缺乏已建立的比較標準。此外，這也間接表明了準確模擬腦組織性質的複雜性和挑戰性。 總結來說，雖然本研究沒有顯示出預期的結果，但仍然排除了幾種方法。儘管存在這些挑戰，本研究為利用腦組織脈動估算腦部性質的潛力和局限性提供了寶貴的見解。某些方法和模型的排除突顯了進一步改進模型以更好地考慮腦組織非線性、粘彈性和各向異性特性的必要性。該研究的方法還顯示出使用無創技術準確估算腦組織機械性質的潛力。 未來的研究應該著重於通過引入更多數據集來提高這些模型的準確性，不僅包括健康志願者，還包括患者，以獲得更準確的結果。此外，未來的研究應該探索其他可能的建模技術。通過解決本研究中識別的局限性，研究人員可以提高無創方法估算腦組織性質的可靠性。此外，將這種方法擴展到各種生理和病理狀態，以增強其臨床應用性，特別是在診斷和治療腦部疾病方面，也將是重點。

Understanding the mechanical properties of brain tissue remains challenging due to difficulties obtaining non-invasive in vivo data. This study addresses this by utilising brain tissue pulsations (BTP) from cardiac cycles to estimate cerebral properties. This study developed a coupled solid-fluid mathematical model fitted to data from 20 healthy individuals using Transcranial Tissue Doppler (TCTD) to measure brain tissue displacement and corresponding blood pressure. The Fourier transformation was used to derive transfer functions between displacement and pressure signals. The mathematical model assumes that the brain tissue can be modelled as a coupled solid-fluid system. The spherical coordinate is applied to the model to simplify the governing equation. Two models have been tried to fit the data, including a coupled solid-fluid model and its revised model with multiple compartments. Model fitting, using MATLAB's 'fminsearch,' optimised vital parameters, including Young’s modulus (E), Poisson’s ratio (ν), specific storage (Q), and permeability over viscosity (κ/μ). Four ways are tried in total. The first three, including original model fitting, logarithmic transformation model fitting to avoid negative magnitude, and revised model fitting, are fitting the non-dimensional groups of parameters formed by physical parameters. Then, the dimensional three can be calculated by non-dimensional groups. The last way is to directly fit the original dimensional three because problems exist from non-dimensional groups to the original three parameters. The fitting results correlate with experimental data across all methods according to the fitting curves and loss. However, the parameter values don’t always perform well with expected magnitudes. This discrepancy is partly due to the absence of established standards for comparison. This also indirectly demonstrates the complexity and challenging nature of accurately modelling brain tissue properties. In conclusion, although an expected result hasn’t been shown in this study, several ways are still excluded. Despite these challenges, this study provides valuable insights into the potential and limitations of using brain tissue pulsations for estimating cerebral properties. The exclusion of certain methods and models underscores the need for further refinement of more models that can better account for the nonlinear, viscoelastic, and anisotropic nature of brain tissue. This study’s approach also shows the potential for using non-invasive techniques to estimate brain tissue's mechanical properties accurately. Future research should focus on enhancing the accuracy of these models by incorporating more datasets, not only for healthy volunteers but also for patients, to get more accurate results. Also, future research should explore other possible modelling techniques. By addressing the limitations identified in this study, researchers can improve the reliability of non-invasive methods for estimating brain tissue properties. Besides, extending this methodology to various physiological and pathological states to enhance its clinical applicability, particularly in diagnosing and treating brain disorders, will also be a focus.