Optimal measurements of hemodynamic response latency in fNIRS using the jackknife approach

详细信息    查看全文

• 作者：Manon Maheux ; É ; tienne Bisaillon-Sicotte ; Shirin Tabrizi ; Jorge L. Armony ; Jean-Marc Lina and Pierre Jolicoeur
• 刊名：Psychophysiology
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：54
• 期：1
• 页码：62-73
• 全文大小：1054K
• ISSN：1469-8986

文摘

Functional near-infrared spectroscopy (fNIRS) permits measurements of changes in the concentration of oxygenated and deoxygenated hemoglobin, typically with a higher sampling rate than with other imaging methods based on the hemodynamic response. We examined the potential of the fNIRS technique to estimate variations in the latency of hemodynamic responses to experimental events and sought optimal methods to maximize the reliability and reproducibility of latency effects. We used Monte Carlo simulations using subsamples of real fNIRS measures to estimate the statistical power of different approaches (such as fixed threshold, percent of peak, fractional-area latency, for both individual-subject estimates and estimates from jackknife averages) to detect a known simulated latency shift. The simulations used measures of hemodynamic responses in the temporal lobe from two groups of young adult participants who listened to auditory stimuli, one with a blocked presentation design and one with an event-related design. We estimated the relative sensitivity of different latency measures and approaches to the measurement of latency effects of different magnitudes using realistic noise and signal-to-noise characteristics. In general, the jackknife approach provided the greatest statistical power to detect a known latency shift, without inflation of Type I error.