The frequency response of cerebral autoregulation

The frequency-response of pressure autoregulation is not well delineated; therefore, the optimal frequency of arterial blood pressure (ABP) modulation for measuring autoregulation is unknown. We hypothesized that cerebrovascular autoregulation is band-limited and delineated by a cutoff frequency for which ABP variations induce cerebrovascular reactivity. Neonatal swine (n = 8) were anesthetized using constant minute ventilation while positive end-expiratory pressure (PEEP) was modulated between 6 and 0.75 cycles/min (min^-1). The animals were hemorrhaged until ABP was below the lower limit of autoregulation (LLA), and PEEP modulations were repeated. Vascular reactivity was quantified at each frequency according to the phase lag between ABP and intracranial pressure (ICP) above and below the LLA. Phase differences between ABP and ICP were small for frequencies of >2 min^-1, with no ability to differentiate cerebrovascular reactivity between ABPs above or below the LLA. For frequencies of <2 min^-1, ABP and intracranial pressure (ICP) showed phase shift when measured above LLA and no phase shift when measured below LLA [above vs. below LLA at 1 min^-1: 156° (139 -174°) vs. 30° (22-50°); P < 0.001 by two-way ANOVA for both frequency and state of autoregulation]. Data taken above LLA fit a Butterworth high-pass filter model with a cutoff frequency at 1.8 min^-1 (95% confidence interval: 1.5-2.2). Cerebrovascular reactivity occurs for sustained ABP changes lasting 30 s or longer. The ability to distinguish intact and impaired autoregulation was maximized by a 60-s wave (1 min^-1), which was 100% sensitive and 100% specific in this model.