The contribution of tremor studies to diagnosis of Parkinsonian and essential tremor: a statistical evaluation

Abstract

This study statistically evaluated a set of commonly measured tremor parameters to determine their individual and combined ability to discriminate between essential tremor (ET) and Parkinsonian tremor (PT). Accelerometer and surface electromyographic (EMG) records of moderate to severe upper limb tremor in 20 patients with ET and 22 patients with PT were used to quantitatively compare tremor amplitude, frequency and pattern of muscle bursting in two resting and three non resting postures. The group statistics showed significant differences between ET and PT with respect to tremor frequency in all five postures, tremor amplitude at rest and muscle bursting patterns. Discriminant function analysis showed that no single parameter or combination of parameters was able to correctly classify all patients. Frequency was much more discriminating than amplitude or muscle bursting patterns in all limb postures. The best amplitude discrimination was obtained when the hand and forearm were both fully supported. Muscle bursting patterns were poorly discriminating and did not assist in correct classification of single patients. Group statistics confirmed a highly significant biological difference between the two tremor types. Optimal classification of single PT (86% correct) and ET (95% correct) patients was obtained using frequency and two selected amplitude parameters from the resting limb. Limb posture was an important variable in optimising the discriminative ability of tremor studies. The implications for routine tremor studies are summarised.