Neurology-Stroke Division, Neurological Institute of New York, Columbia University Irving Medical Center, New York, NY, United States of America. The first two authors contributed equally to this manuscript.
Physiol Meas. 2020 Mar 6;41(2):024002. doi: 10.1088/1361-6579/ab71f2.
To develop a joint time-frequency analysis technique based on generalized harmonic wavelets (GHWs) for dynamic cerebral autoregulation (DCA) performance quantification.
We considered two groups of human subjects to develop and validate the method: 55 healthy volunteers and 35 stroke-free subjects with unilateral internal carotid artery stenosis (CAS). We determined the mean and coherence-weighted average of the phase shift (PS) of appropriately defined GHW-based transfer functions, based on data points over the joint time-frequency domain. We compared agreement of standard transfer function analysis (TFA) and GHW analyses in healthy subjects using Bland-Altman plots. We assessed sensitivity of each metric to detect the presumed side-to-side difference in DCA function in CAS subjects (with decreased PS on the occluded side), using McNemar's chi square test to compare each metric to the standard TFA approach. An alternative Morlet wavelet-based approach was also considered.
The GHW and TFA methods exhibited strong agreement in healthy subjects. Among CAS subjects, GHW metrics outperformed TFA and Morlet wavelet-based approaches in identifying expected side-to-side differences: TFA sensitivity was 40.0% (95%CI 23.9-57.9), Morlet 60.0% (95%CI 42.1-76.1), and GHW >70% for both metrics (GHW mean PS sensitivity 74.3, 95%CI 56.7-87.5, p = 0.0027 versus TFA; GHW coherence-weighted PS sensitivity 71.4, 95%CI 53.7-85.4, p = 0.0009 versus TFA).
In comparison to the widely used stationary Fourier transform-based TFA and to Morlet wavelet-based analysis, our data suggest that the GHW-based analysis performs better in identifying DCA asymmetry between the two cerebral hemispheres in patients with high grade unilateral carotid stenosis. Our method may provide enhanced confidence in employing DCA metrics as a sensitive diagnostic tool for detecting impaired DCA function in a variety of pathological settings.
基于广义谐波小波（GHW）开发一种联合时频分析技术，用于定量分析动态脑自动调节（DCA）性能。
我们考虑了两组人体受试者来开发和验证该方法：55 名健康志愿者和 35 名无卒中性单侧颈内动脉狭窄（CAS）的受试者。我们基于联合时频域中的数据点，确定了适当定义的基于 GHW 的传递函数的相位滞后（PS）的均值和相干加权平均值。我们使用 Bland-Altman 图比较了健康受试者中标准传递函数分析（TFA）和 GHW 分析的一致性。我们使用 McNemar 的卡方检验比较了每种指标与标准 TFA 方法，评估了每种指标在检测 CAS 受试者（闭塞侧 PS 降低）中 DCA 功能假定的双侧差异的敏感性。还考虑了一种替代的 Morlet 小波方法。
GHW 和 TFA 方法在健康受试者中表现出很强的一致性。在 CAS 受试者中，GHW 指标在识别预期的双侧差异方面优于 TFA 和 Morlet 小波方法：TFA 敏感性为 40.0%（95%CI 23.9-57.9），Morlet 为 60.0%（95%CI 42.1-76.1），而 GHW 为两种指标均>70%（GHW 平均 PS 敏感性为 74.3，95%CI 56.7-87.5，p=0.0027 与 TFA 相比；GHW 相干加权 PS 敏感性为 71.4，95%CI 53.7-85.4，p=0.0009 与 TFA 相比）。
与广泛使用的基于平稳傅里叶变换的 TFA 和 Morlet 小波分析相比，我们的数据表明，在识别高分级单侧颈动脉狭窄患者两侧大脑半球之间的 DCA 不对称方面，基于 GHW 的分析表现更好。我们的方法可能会增强 DCA 指标作为检测各种病理情况下受损 DCA 功能的敏感诊断工具的可信度。
