Classification of somatosensory-evoked potentials recorded from patients with severe head injuries

Abstract
The problem of sensory evoked potential (EP) assessment in the critical care setting to isolate damage in specific neuronal pathways, as manifested by certain abnormalities in the response waveform is being addressed using neural networks. An existing visually based grading scheme (GGS, for Greenberg grading system) for somatosensory EPs collected from patients with severe head injuries is being automated. The collection of data used in this research, which consist of somatosensory-evoked potential (SEP) waveforms collected from patients with head injuries is described. The way the system (called Pathfinder) works is described, and results obtained with it are presented.