We carried out baseline and short-term follow-up MRI, including perfusion-weighted imaging (PWI) and tests of neurologic and cognitive function on 15 consecutive patients with large-vessel ischemic stroke who showed a persistent large perfusion-diffusion mismatch at enrollment up to seven days after the onset of symptoms. Of these, ten underwent induced blood pressure elevation with phenylephrine and oral medications (in eight) or intravenous fluids (in two) with the goal of improving perfusion; five had no such treatment. Significant functional improvement was defined by a reduction of 3 or more points on the NIH stroke scale (NIHSS). Significant improvement in perfusion was defined by a reduction in the volume of hypoperfused brain by 30 cc on PWI using time-to-peak (TTP) maps, without enlargement of the infarct. There was a strong, statistically significant association between improved function and improved perfusion: six (75%) of eight patients who improved in function, but none of the seven who did not, showed a reduction in volume of hypoperfused brain. All six patients who met the perfusion goal, and only two (22%) of nine who did not showed significant functional improvement (Fisher's exact: P < 0.01). There were no differences between patients who improved functionally and those who did not with respect to age, initial volume of abnormality on DWI or PWI, initial NIHSS, or changes on DWI. These findings indicate that reduction in volume of hypoperfused brain on PWI is a marker of response to treatment to improve perfusion even in subacute stroke and that partial reperfusion of regions of salvageable but dysfunctional tissue is a mechanism of improved function associated with induced blood pressure elevation.