Prediction of human errors by maladaptive changes in event-related brain networks.

Humans engaged in monotonous tasks are susceptible to occasional errors that may lead to serious consequences, but little is known about brain activity patterns preceding errors. Using functional MRI and applying independent component analysis followed by deconvolution of hemodynamic responses, we studied error preceding brain activity on a trial-by-trial basis. We found a set of brain regions in which the temporal evolution of activation predicted performance errors. These maladaptive brain activity changes started to evolve approximately 30 sec before the error. In particular, a coincident decrease of deactivation in default mode regions of the brain, together with a decline of activation in regions associated with maintaining task effort, raised the probability of future errors. Our findings provide insights into the brain network dynamics preceding human performance errors and suggest that monitoring of the identified precursor states may help in avoiding human errors in critical real-world situations.