Altered cerebrovascular-CSF coupling in Alzheimer’s Disease measured by functional near-infrared spectroscopy

Figures

• 

  Figure 1. The heat maps from 10 and 800 Hz data summarise results from this study. The heat maps revealed that SD of coupling pattern from cardiac band involving HbR and its first derivate (dHbR) provided good separation between AD patients and controls, see the bold face pairs. Results from the VLF and respiratory bands from both sampling frequencies looked slightly different.

• 

  Figure 2. Scatter plots of MMSE vs. SD of lag variation analysis with the fitting linear line (A) cardiac band at 10 Hz, (B) cardiac band at 800 Hz. Red dots present AD patients and blue rings healthy controls.

• 

  Figure 3. We measured fNIRS signals (CSF water, oxygenated HbO and deoxygenated HbR blood) from the forehead area using a multiwavelength fNIRS able to sense water concentration (A). Sample signals in the cardiac band taken from the same subject within the same time frame were given. Signal analysis model utilises cross-correlation process for each pair within pre-defined time window, red and blue represent hemodynamic and CSF signals respectively (B). The window is shifted with certain overlap. This diagram represents 25% overlap of the window. For cross-correlation process in each window we tracked the lag value achieving the best correlation coefficient. Collecting these lag values across the whole signals forms a series that is subjected to pattern analysis using standard deviation (SD) (C). Finally, a significant test using Wilcoxon-test at 0.05 significant level was used.