Algorithms for the Clinical Analysis of Nystagmus Eye Movements

Two algorithms used in the clinical analysis of nystagmus are described. Their development was necessitated by the greater complexity of the nystagmus waveforms in response to system identification types of vestibular and optokinetic stimuli as compued to the less complex response to a step input. Practical considerations for clinical application also influenced their development. The first algorithm converts nystagmus data into a regulary sampled estimate of slow phase velocity (SPV), an important feature of the signal. It uses a new set of fast phase detection conditions which allow for automatic processing of reversals in nystagmus direction and for wide variability for clinical data. The second algorithm detects noise induced spikes in this SPV estimate using an adaptive criterion based upon a measure of the overall "noisiness" of the data compared to the stimulus.