Fine tuning breath-hold-based cerebrovascular reactivity analysis models

Area of Science:

• Neuroimaging
• Vascular Neurology
• Medical Physics

Background:

• Cerebrovascular reactivity (CVR) assessment is crucial for diagnosing neurovascular conditions.
• Existing breath-hold (BH) derived CVR analysis methods have limitations in precision.
• Accurate CVR interpretation is vital for understanding hemodynamic impairment.

Purpose of the Study:

• To evaluate existing and novel analysis methods for breath-hold (BH) derived cerebrovascular reactivity (CVR) measurements.
• To provide insights and models for more exact CVR interpretation.
• To compare model-free methods against traditional sine models for CVR analysis.

Main Methods:

• Utilized five blood-oxygen-level-dependent (BOLD) fMRI datasets from neurovascular patients.
• Calculated temporal lag (phase), percent BOLD signal change (CVR), and explained variance (coherence) maps.
• Compared three sine models with two novel model-free "Optimal Signal" methods.

Main Results:

• All tested models demonstrated significant differences in CVR and coherence between affected and unaffected hemispheres.
• Voxel-wise phase determination significantly increased CVR (P < 0.05).
• Novel model-free "Optimal Signal" methods showed comparable performance to a two-task sine model.

Conclusions:

• Improved CVR and coherence were achieved using voxel-wise phase and frequency adjustment.
• Novel "Optimal Signal" methods offer a robust, feasible, and compliance-independent alternative to sine models.
• The sagittal sinus model-free method is advantageous due to its independence from hemispheric CVR impairment.