Cortical spreading depression (CSD) and periinfarct depolarization (PID) have been shown in various experimental models of stroke to cause secondary neuronal damage and infarct expansion. For decades it has been questioned whether CSD or PID occur in human ischemic stroke. Here, we describe CSD and PID in patients with malignant middle cerebral artery infarction detected by subdural electrocorticography (ECoG).


Centres of the Co-operative Study of Brain Injury Depolarisations (COSBID) recruited 16 patients with large middle cerebral artery infarction. During surgery for decompressive hemicraniectomy, an electrode strip was placed on the periinfarct region, from which four ECoG channels were acquired.


A total of 1,638 hours was recorded; mean monitoring time per patient was 109.2 hours. A total of 127 CSD and 42 PID events were observed. In CSD, a stereotyped slow potential change spreading between adjacent channels was accompanied by transient depression of ECoG activity. In PID, a slow potential change spread between neighboring channels despite already established suppression of ECoG activity. Most CSDs and PIDs appeared repetitively in clusters. CSD or PID was observed in all but two patients. In these two patients, the electrode strip had been placed over infarcted tissue, and accordingly, no local ECoG or recurrent transient depolarization activity occurred throughout the observation period.


CSD and PID occurred spontaneously with high frequency in this study of patients with malignant middle cerebral artery infarction. This suggests that the large volume of experimental studies of occlusive stroke that implicate spreading depolarizations in its pathophysiology can be translated, with appropriate caution, to patients and their treatment. Ann Neurol 2008