The EEG of the Newborn Brain - Detection of Hypoxia and Prediction of Outcome

Doctoral thesis, 2005

Brain function requires a continuous flow of oxygen and an insufficient supply (hypoxia) for a few minutes during the first period of life may have severe consequences; consequences which might impact both the life of the individual who has been struck and his or her family. Therefore, there is a need to find a method which can enable the detection of an insufficient oxygen supply at an early stage, before brain damage becomes permanent. A decreased flow of oxygen to the brain can be the result of reduced cerebral circulation, poor oxygenation in the lungs etcetera. Since there is a large variation in how individuals are affected, it is also desirable to have a means of predicting the outcome. The work presented in the current thesis aims at finding methods for preventing permanent brain damage and at predicting the future development if damage has still occurred. The objectives are to investigate if these goals can be reached by the use of various signal processing techniques applied to the electroencephalogram (EEG) and remote consultations between paediatricians and neurophysiologists. A study approaching the latter of these issues is presented in the thesis and the results show that merging the knowledge from the different specialities involved (in this case neonatology, paediatric heart surgery, and neurophysiology) has a great potential value. Not only does the merging improve the care of the individual patients, it also by facilitating the spreading of knowledge among specialists, and possibly also results in new knowledge. There are however obstacles when introducing a remote consultation system in everyday health care and some of these problems are documented in the report. The usefulness of various signal processing techniques applied to EEG is assessed by defining evaluation parameters for both prognostic value and detection properties. The most promising methods for each purpose predict the outcome with high accuracy within 30 min from the end of hypoxia or detect the effects of hypoxia within two minutes with a reasonably low false alarm rate. Earlier papers on EEG interpretation in association with hypoxia show that the properties of periods with alternating high and low amplitude of the EEG are relevant for the assessment of the outcome after hypoxia. Therefore, the frequency content of such high-amplitude periods is studied and comparisons are made between results from healthy and post-asphyctic newborn babies. For the latter of these two groups, the power in the frequencies below the recordable frequency band of conventional EEG amplifiers is also studied for the above mentioned high-amplitude periods. The results presented in this thesis show ways to improve the intensive care of newborn babies. Hopefully, these ideas will reach every day health care and help to improve the quality of life for several children exposed to hypoxia, and for their families.