Ischaemic stroke remains a major killer and the first cause of acquired handicap in the UK, costing billions of pounds yearly. Preventing death and limiting handicap are therefore major goals. This can be achieved only if the pathophysiology of infarct expansion is properly understood. Work in the animal, especially the non human primate, has shown that following occlusion of the middle cerebral artery (MCA) – the most frequent and prototypical stroke – there is a gradient of reduction in tissue blood flow from the periphery of the cortical MCA territory towards its centre, from perfusion being normal or only mildly reduced in the former to very low (but never abolished) in the latter. Tissue fate has been shown to locally depend on the severity of hypoperfusion and duration of occlusion, with a fraction of the MCA territory being initially in a “penumbral” state, i.e, electrically silent but still salvageable by early reperfusion. Translating this knowledge in the human has proven a long and difficult journey, but was eventually achieved using physiological quantitative imaging, specifically PET. However, PET in MCA stroke has also revealed the presence of considerable pathophysiological heterogeneity from patient to patient, largely unpredictable from either elapsed time since onset or clinical deficit. At time of scanning, some patients show spontaneous recanalization of the MCA, with efficient and actually excessive reperfusion, and these patients do well and end up with very small infarcts and excellent functional recovery. The remaining patients still have MCA occlusion, but within this group, two main patterns are encountered: i) small “core” of already irreversible damage but extensive “penumbral” hypoperfusion; and ii) extensive “core” with little or no remaining penumbra. The latter patients have poor prognosis, and often develop “malignant” MCA stroke, while the former have unpredictable outcome, with the amount of clinical recovery proportional to the volume of penumbra that escapes infarction. Importantly, all three pathophysiological patterns can be seen within 3 hrs of clinical onset, while persistent penumbra can be encountered as late as 16-18 hrs after onset, suggesting a prolonged window for therapeutic intervention. While these observations underpinned key trials of thrombolysis, which have shown clinical benefit up to 3 hrs (now part of clinical routine), they also indicate that not all patients may be rational candidates for thrombolysis, and that only those who are likely to benefit should be exposed to the risks of this intervention. It is increasingly recognized that this underlying pathophysiological heterogeneity needs to be accounted for if thrombolysis is contemplated, and accordingly pathophysiological diagnosis with imaging is rapidly becoming an essential component of stroke assessment and customized management, replacing the clock by individual pathophysiology. To implement this, diffusion- and perfusion-weighted MR (DWI-PWI) and CT-based perfusion imaging are increasingly used, although still undergoing formal validated against gold-standard PET. Beyond thrombolysis per se, knowledge of the individual pathophysiology also guides management of physiological variables like blood pressure, blood glucose, oxygen saturation and temperature, which can otherwise precipitate the penumbra into the core, and the mildly hypoperfused tissue into the penumbra. We propose that future trials of neuroprotection use physiological imaging to select the patient category that best matches the drug’s presumed mode of action, rather than lumping together patients with entirely different pathophysiological patterns in so-called “large trials”, which have all failed so far.