Cases underwent a comprehensive tremor questionnaire. During the questionnaire, age of onset and duration of tremor symptoms were assessed, as well as complaints of leg tremor (“Do you often have an uncontrollable tremor in your leg?”) and family history of tremor (first- and second-degree relatives). Reported frequency of falls in the last year was assessed (“How many falls have you had in the past year?”).
The videotaped neurological examination included detailed assessments of postural, kinetic, and intention tremor in the arms. The severity of postural and kinetic tremor in the arms (total tremor score: range = 0–36) was based on the 0 to 3 ratings of six tasks with each arm. As in previous studies, intention tremor (i.e., tremor that occurs with goal-directed movement [finger-nose-finger movement] and worsens when approaching the target) was rated as 0 (absent), 0.5 (probable), and 1 (definite); severity of intention tremor (two arms combined) was therefore graded from 0 to 2. Presence of head and voice tremor was assessed during the videotaped neurological examination.
The study began in 2000; however, assessment of leg tremor during the videotaped neurological examination did not begin until August 2007. Starting at that point, the videotaped neurological examination included an assessment of postural tremor of the legs, as previously described, as well as 10 repetitive toe-to-target movements with each foot. Patients were examined while seated and asked to remove their shoes and socks, although a few preferred to keep their socks on. They were asked to raise their foot from the ground to reach the target (a tongue blade) and touch it with their big toe. The tongue blade was placed at least 16 inches from the ground level (see Videos 1 and 2). The videotapes were reviewed independently by two movement disorder neurologists (M.K. and E.D.L.), who noted the presence versus absence of intention tremor in the legs (i.e., tremor that worsened terminally as the patient's foot was in close proximity to the target). The more junior of the two neurologists (M.K.) was trained by the more senior neurologist (E.D.L.) to conservatively mark the presence of intention tremor in the legs. In the case of disagreements between the neurologists, the videotapes were coreviewed until a consensus was reached. There were 37 instances in which the junior neurologist marked the presence of intention tremor of the legs, but on further assessment by the senior neurologist, the tremor was deemed to be no more than kinetic tremor. There were no other sources of disagreement between the two neurologists.
A 9-m walk was assessed by asking the patient to walk that distance. This task was performed twice and the average time was calculated.
Two subjects with intention tremor of the legs (1 male and 1 female, both with intention tremor of the arms as well) were selected to undergo computerized quantitative tremor analysis using an ultralight piezoresistive miniature accelerometer (±25 g; weight, 1.2 g) with linear sensitivities of approximately 4.5 mV/g in the biological tremor range (0–25 Hz). The accelerometer was attached to the distal phalanx of the great toe. Movement signals were obtained while performing a targeting task with the foot, moving from the floor to a pressure sensitive target cue 80 cm in height.
The tremor acquisition setup allowed for relatively unrestrained activity of the leg and foot throughout testing, approximating the clinical state. Six 10-second trials were obtained per patient. Data from the more affected foot were used in the calculations of averages. Total testing time was approximately 1 hour.
Two channels recorded the accelerometric and target sensor signals. Recordings were digitized at 700 Hz with a 16-bit A/D board, smoothed, and processed. Displacement (tremor amplitude) was derived offline by double integration of accelerometric data after filtering out low-frequency voluntary movements (less than 2 Hz) and averaging. Displacement data were calculated from 0.5-second epochs taken from 100 ms after movement onset and 100 ms before reaching the target sensor. Tremor frequencies were calculated from accelerometry using a fast Fourier transform (FFT) to generate auto spectra.