The ‘penumbra’ is a concept coined in animal experiments suggesting that functionally impaired tissue can survive and recover if sufficient reperfusion is re-established within a limited time period, which depends on the level of residual flow. In an ischaemic territory, irreversible damage progresses over time from the centre of the most severe flow reduction to the periphery with less disturbed perfusion. This centrifugal progression of irreversible tissue damage is characterised by a complex cascade of interconnected electrophysiological, molecular, metabolic and perfusion disturbances. Waves of depolarisations, the peri infarct spreading depressions, inducing activation of ion pumps and liberation of excitatory transmitters play an important role in the drastically increased metabolic demand during reduced oxygen supply causing hypoxic tissue changes and lactacidosis, which further damage the tissue. Positron emission tomography allows the quantification of regional cerebral blood flow, the regional metabolic rate for oxygen and the regional oxygen extraction fraction, which can be used to identify regions with a critical reduction in these physiologic variables as indicators of penumbra and irreversible damage within ischaemic territories in animal models and patients with stroke. These positron emission tomography methods require arterial blood sampling and due to the complex logistics involved, are limited for routine application. Therefore, newer tracers were developed for the noninvasive detection of irreversible tissue damage (flumazenil) and of hypoxic tissue changes (fluoromisonidazole). As a widely applicable clinical tool, diffusion/perfusion-weighted magnetic resonance imaging is used; the ‘mismatch’ between perfusion and diffusion changes serves as a surrogate marker of the penumbra. However, in comparative studies of magnetic resonance imaging and positron emission tomography, diffusion-weighted imaging showed a high false-positive rate of irreversible damage, and the perfusion-weighted–diffusion-weighted mismatch overestimated the penumbra as defined by positron emission tomography. Advanced analytical procedures of magnetic resonance imaging data may improve the reliability of these surrogate markers but should be validated with quantitative procedures.