Importance of Objective Tremor Measurements

by Michael P. Caligiuri, PhD

Traditional assessments of tremor severity may employ several tasks. Handwriting, finger-to-nose testing, and drawing Archimedes spirals are the most common. Resting tremor is observed while the patient is in repose or when engaged in various cognitive tasks. Postural tremor is observed when a patient is asked to hold an object or extend arms forward suspended against gravity. Action or intention tremor is observed when a patient is asked to pick up an object or track a moving object with his finger. While these techniques are convenient, their mode of administration is nonstandard and their reliability may be questioned. As with all judgments of severity, the ratings are subjective and often influenced by examiner experience or bias.

Numeric rating scales are an attempt to increase objectivity while retaining the advantage of convenience. In general, tremor severity is rated along a 4- or 5-point ordinal scale ranging from absent to severe (such as with the Unified Parkinson’s Disease Rating Scale or UPDRS). However, rating scales are imprecise to the extent that they cannot distinguish between various forms of tremor such as parkinsonian or essential tremor. Moreover, ordinal scales are nonlinear making it difficult to detect change over time or following an intervention and are difficult to use across patients or clinics.

Important attributes which distinguish instrumental approaches from traditional ratings scales include the capability of a yielding linear measurement of the movement under study, increased resolution for detecting mild tremor, and selectivity to tremor phenomena.

Linearity

Instrumental systems are usually designed to measure movement abnormalities in terms of a continuous variable on an interval scale. In contrast, multi-item ordinal scales such as the UPDRS categorize the movements by degree of severity. Instrumental systems are designed to yield linear measurement of severity, that is, measurements in which the outcome variable is directly proportional to the input. For example, a clinical rating of “2” may not represent tremor severity that is twice as severe as a rating of “1”; however, a power value from the Tremorometer of “100” is exactly twice as severe as a power value of “50”.

Selectivity

The greatest challenge facing clinicians who work with tremor patients lies in their ability to distinguish one form of tremor from another. Quantitative instrumental approaches assess tremor in terms of its frequency characteristics which are used to identify the tremor etiology. It is not uncommon for patients to present with multiple forms of tremor which cannot be easily separated visually. Accelerometer-based instruments coupled with spectral analyses can rapidly portray individual or multiple tremor components.

While considerable progress has been made toward the development of instrumental systems for quantifying tremor, there are limitations. Multi-dimensional recording systems are often bulky and introduce inertial loading to the limb under study. Most laboratory systems require signal amplification and digital acquisition hardware which place additional demands on limited resources. This restriction also limits access to the technology by non-ambulatory patients. Prior to the development of the Tremorometer, available instrumental systems required some familiarity with electronics and signal processing. It was common for movement disorders laboratories to include personnel knowledgeable in the fields of motor control, biomedical instrumentation, or digital signal processing.

High Resolution

Instrumental assessment can reveal subtle motor abnormalities below the threshold of detection during a clinical evaluation. This aspect of quantification allows for the detection of mild tremors that occur early in a disease process or small changes in tremor severity related to intervention.

Standardized Tremor Assessment

Since almost every tremor measurement system is unique, there has been no standardized method for assessing the patient nor have the data analyses been comparable across laboratories. The Tremorometer® provides a standard method for measuring and quantifying tremor characteristics which we believe will have a positive effect on both basic and clinical tremor research.

Tremorometer Applications

The Tremorometer has multiple applications in clinical and research settings:

1. Rapid assessment of drug-induced side effects
2. Evaluation of anti-parkinsonian medications and surgical interventions
3. Diagnosing tremor in patients on tremor-inducing drugs
4. Separating tremor from non-tremor hyperkinetic movements
5. Large-scale multi-site clinical trials or basic tremor research where reliability and standardization are important
6. Tremor screening of large clinical and research populations
7. Immediate tremor readings may be used to assist in setting parameters for brain stimulators to minimize tremor

FlexAble Systems, Inc. completed a 2-year NIMH-sponsored multi-site study in August 2001 that evaluated the validity and sensitivity of the Tremorometer. Six sites participated. Patients with a variety of tremors including idiopathic Parkinson’s disease, neuroleptic-induced parkinsonian tremor, lithium-induced tremor, essential tremor, tremors related to substance abuse, and tremors of unknown etiology were studied. In addition, important demographic, diagnostic, pharmacologic, and historical data were obtained from each participant. Data from over 300 patients and healthy comparison subjects were pooled into a single database for analyses. Some of the more important results were:

1. Tremorometer readings are reliable over time with test-retest correlation coefficients of 0.95 for frequency and 0.82 for amplitude of postural tremor.
2. Analyses support the criterion validity of the Tremorometer. Tremor power and percent scores discriminate patients with mild versus moderate tremor based on clinical impression with nearly 90% accuracy.
3. Patients with Parkinsonian and drug-induced tremors, but not essential tremor, show statistically significant increase in tremor frequency when tested during rest versus tested during posture. This finding may help distinguish parkinsonian from essential tremors.