Near infrared spectroscopy (NIRS) enables continuous monitoring of dynamic cerebrovascular autoregulation, but this methodology relies on invasive blood pressure monitoring (iABP). We evaluated the agreement between a NIRS based autoregulation index calculated from invasive blood pressure monitoring, and an entirely non-invasively derived autoregulation index from continuous non-invasive blood pressure monitoring (nABP) using the Finometer photoplethysmograph. Autoregulation was calculated as the moving correlation coefficient between iABP and rSO2 (iTOx) or nABP and rSO2 (nTOx). The blood pressure range where autoregulation is optimal was also determined for invasive (iABPOPT) and non-invasive blood pressure measurements (nABPOPT). 102 simultaneous bilateral measurements of iTOx and nTOx were performed in 19 patients (median 2 per patient, range 1-9) with different acute pathologies (sepsis, cardiac arrest, head injury, stroke). Average iTOx was 0.01 ± 0.13 and nTOx was 0.01 ± 0.11. The correlation between iTOx and nTOx was r = 0.87, p < 0.001, 95 % agreement ± 0.12, bias = 0.005. The interhemispheric asymmetry of autoregulation was similarly assessed with iTOx and nTOx (r = 0.81, p < 0.001). Correlation between iABPOPT and nABPOPT was r = 0.47, p = 0.003, 95 % agreement ± 32.1 mmHg, bias = 5.8 mmHg. Coherence in the low frequency spectrum between iABP and nABP was 0.86 ± 0.08 and gain was 1.32 ± 0.77. The results suggest that dynamic cerebrovascular autoregulation can be continuously assessed entirely non-invasively using nTOx. This allows for autoregulation assessment using spontaneous blood pressure fluctuations in conditions where iABP is not routinely monitored. The nABPOPT might deviate from iABPOPT, likely because of discordance between absolute nABP and iABP readings.