Today there is no strict definition of consciousness. That is why, it is so hard to define is patient aware himself, or no. And that's why doctors, who works in neurological biomedicine, make too much mistakes in their practice. In most cases, if patient, who has brain injury, is not in consciousness after long time, can't turn back to normal life. This is a one reason, why it is so important strict determine the consciousness of a patient. Journal 'Brain' tries to give for doctors a little help in this. The article details a machine learning algorithm, that can answer, by input data about brain activity of a patient, is this patient in consciousness. This algorithm make this better, then a doctor-human. However, the idea of determining if human is conscious is not new.  The scientists was find a method, that use PET signals, but it's expensive too much, that's why it is not everywhere.  A more accessible alternative is electroencephalography or EEG. The researchers collected data of EEGs about 300 patients in middle-conscience and gave it to machine learning algorithm - DOC Forest.  It can strict determine more that 3/4 of all diagnoses. But today, we really don't know about what consciousness is it, and for this reason, we can not be 100 percent sure that this method will give an accurate diagnosis.