Evaluation of dynamic cerebral autoregulation might yield a physiologically more adequate measure of cerebral hemodynamic impairment in carotid artery stenosis than CO2-reactivity. This study re-evaluates and compares the Valsalva maneuver (VM) and phase shift during deep breathing. Nineteen patients with severe carotid artery stenosis and 17 age-matched controls were examined using transcranial Doppler sonography and non-invasive blood pressure recordings (Finapres). Phase shift was determined by cross-spectral analysis, responses to VM were graded by the formerly-introduced autoregulation slope index (ASI) and the new Valsalva time index (VTI). Phase shift and autoregulatory indices were significantly reduced on the affected side (p < 0.001). Correlations with CO2-reactivity were significant when pooling values of controls and patients (r from 0.54 to 0.78; p < 0.001). Correlations except for the VTI (r = -0.65; p = 0.002) were not significant considering only the affected side in patients. Correlations of pooled values between phase shift and VM-derived indices were significant (VTI r = -0.62; p < 0.001; ASI r = 0.49; p < 0.001), within patients only when comparing side-to-side differences (VTI r = -0.58; p = 0.009; ASI r = 0.52; p = 0.023). In conclusion, detection of impaired cerebral autoregulation is possible both by deep breathing and VM. The new VTI seems to be more suitable than the conventional ASI. Inter-method agreement concerning the extent of impairment is only acceptable for intra-individual side-to-side differences. Since absolute values of one autoregulation testing method or CO2-reactivity alone might fail, various tests should be combined for comprehensive assessment of cerebral hemodynamic impairment.