Objective and personalized longitudinal assessment of a pregnant patient with post severe brain trauma

Elizabeth B. Torres, Brian Lande

Following severe trauma to the brain (whether internally generated by seizures, tumors or externally caused by collision with or penetration of objects) individuals may experience initial coma state followed by slow recovery and rehabilitation treatment. At present there is no objective biometric to track the daily progression of the person for extended periods of time. We introduce new analytical techniques to process data from physically wearable sensors and help track the longitudinal progression of motions and physiological states upon the brain trauma. Setting and Participant: The data used to illustrate the methods were collected at the hospital settings from a pregnant patient in coma state. The patient had brain trauma from a large debilitating seizure due to a large tumor in the right pre-frontal lobe. We registered the wrist motions and the surface-skin-temperature across several daily sessions in four consecutive months. A new statistical technique is introduced for personalized analyses of the rates of change of the stochastic signatures of these patterns. We detected asymmetries in the wrists' data that identified in the dominant limb critical points of change in physiological and motor control states. These patterns could blindly identify the time preceding the baby's delivery by C-section when the patient systematically brought her hand to her abdominal area. Changes in temperature were sharp and accompanied by systematic changes in the statistics of the motions that rendered her dominant wrist's micro-movements more systematically reliable and predictable than those of the non-dominant writst. The new analytics paired with wearable sensing technology may help track the day-by-day individual progression of a patient with post brain trauma in clinical settings and in the home environment.