Dynamic cerebral autoregulation (CA) describes the transient response of cerebral blood flow (CBF) to rapid changes in arterial blood pressure (ABP). We tested the hypothesis that the efficiency of dynamic CA is increased by brain activation paradigms designed to induce hemispheric lateralization. CBF velocity [CBFV; bilateral, middle cerebral artery (MCA)], ABP, ECG, and end-tidal Pco(2) were continuously recorded in 14 right-handed healthy subjects (21-43 yr of age), in the seated position, at rest and during 10 repeated presentations (30 s on-off) of a word generation test and a constructional puzzle. Nonstationarities were not found during rest or activation. Transfer function analysis of the ABP-CBFV (i.e., input-output) relation was performed for the 10 separate 51.2-s segments of data during activation and compared with baseline data. During activation, the coherence function below 0.05 Hz was significantly increased for the right MCA recordings for the puzzle tasks compared with baseline values (0.36 +/- 0.16 vs. 0.26 +/- 0.13, P < 0.05) and for the left MCA recordings for the word paradigm (0.48 +/- 0.23 vs. 0.29 +/- 0.16, P < 0.05). In the same frequency range, significant increases in gain were observed during the puzzle paradigm for the right (0.69 +/- 0.37 vs. 0.46 +/- 0.32 cm.s(-1).mmHg(-1), P < 0.05) and left (0.61 +/- 0.29 vs. 0.45 +/- 0.24 cm.s(-1).mmHg(-1), P < 0.05) hemispheres and during the word tasks for the left hemisphere (0.66 +/- 0.31 vs. 0.39 +/- 0.15 cm.s(-1).mmHg(-1), P < 0.01). Significant reductions in phase were observed during activation with the puzzle task for the right (-0.04 +/- 1.01 vs. 0.80 +/- 0.86 rad, P < 0.01) and left (0.11 +/- 0.81 vs. 0.57 +/- 0.51 rad, P < 0.05) hemispheres and with the word paradigm for the right hemisphere (0.05 +/- 0.87 vs. 0.64 +/- 0.59 rad, P < 0.05). Brain activation also led to changes in the temporal pattern of the CBFV step response. We conclude that transfer function analysis suggests important changes in dynamic CA during mental activation tasks.