Computational challenges for the analysis of intracranial pressure and heart rate data in traumatic brain injuries patients

Background and motivation: Intracranial pressure (ICP) after severe brain injuries or similar life threatening conditions can be continuously monitored (Hu et al., 2009). The ICP signal contains useful information to predict life threatening conditions such as intracranial hypertension. So far, monitoring approaches are focusing mainly on the relationship between arterial blood pressure and intracranial pressure. Own observation in pediatric patients however, showed that changes in heart rate have direct influence on the ICP. Our hypothesis therefore is that the HR–ICP relationship can be quantified via complex event processing methods. A few works concentrate on the capability of identifying a model describing the intracranial system behaviour. For example, in Hu et al. (2007a), the authors present an estimation algorithm based on hidden state estimation approach and non linear Kalman filters to estimate unobserved variable given some measurements such as ICP and cerebral blood flow velocity (CBFV). What might be interesting is understanding the interelationship between ICP and other measures of the monitored patients. For example, in Hu et al. (2008), the authors present ApEN an algorithm based on the adaptive calculation of approximate entropy, integrated with a causal coherence analysis that is able to exploit the potential interaction between ICP and R wave intervals (Hu et al., 2008). Interesting in this sense is also (Hu et al., 2007b) where the authors extract indices from beat to beat mean intracranial pressure measurements and intervals between consecutive normal sinus heart beats (ICP and RR intervals). Starting from the visual observation that heart rate and ICP present peaks at similar points, we applied several statistical methodologies to identify such co-occurrences of peaks and the relationships existing between the two time series. Moreover we are currently investigating the relationship existing also with the other variables monitored. This preliminary analysis performed appear to be promising and we are now extending our work to perform online peaks detection of ICP peaks considering the relationship between the ICP and HR.