Method and apparatus for continuous measurement of motor symptoms in parkinson's disease and essential tremor with wearable sensors
Disclosed embodiments include an apparatus for generating a plurality of movement impairment indices from one or more kinematic signals to characterize movement disorders. Additionally we disclose methods for generating a plurality of movement impairment indices from one or more kinematic signals obtained from one or more kinematic sensors, said methods implemented in a digital computer with one or more processors in order to characterize movement disorders based on spectral analysis, regularity metrics, and time-frequency analysis.
1. A method for generating a plurality of movement impairment indices from one or more kinematic signals obtained from one or more kinematic sensors, said method implemented in a digital computer with one or more processors in order to characterize movement disorders, said method comprising:
(a) collecting said one or more kinematic signals from said one or more kinematic sensors in a wearable inertial monitoring system configured for continuous measurement of movement;
(b) processing said one or more kinematic signals to reject artifacts and noise resulting in one or more enhanced kinematic signals;
(c) finding a measure of power spectral density of said one or more enhanced kinematic signals; and
(d) calculating one or more impairment indices to characterize movement disorders in Parkinson's disease or essential tremor from said measure of power spectral density.
2. The method of claim 1 , further comprising detrending said one or more enhanced kinematic signals prior to finding said measure of power spectral density of said one or more enhanced kinematic signals to eliminate low frequency components of said one or more kinematic signals.
3. The method of claim 2 , further comprising generating one or more total magnitude bandlimited signal powers from said one or more enhanced kinematic signals.
4. The method of claim 3 , wherein said one or more total magnitude bandlimited signal powers correspond to a total magnitude of acceleration or total rotational rate within a given frequency band characterizing specific impairment conditions.
5. The method of claim 4 , further comprising postprocessing said plurality of impairment indices from said measure of power spectral density to generate a plurality of smoother impairment indices for continuous quantification of tremor, dyskinesia, and bradykinesia as a function of time based on said one or more kinematic sensors.
6. The method of claim 1 , wherein calculating said plurality of impairment indices from said measure of power spectral density involves computing a plurality of regularity metrics.
7. The method of claim 6 , wherein said plurality of regularity metrics are chosen from the group consisting of spectral flatness, eigenvalue spread, Lempel-Ziv complexity, Approximate Entropy, Sample Entropy, Multi-Scale entropy, nonlinear metrics, and complexity metrics.
8. The method of claim 1 , wherein finding said measure of power spectral density of said one or more enhanced kinematic signals includes generating a time-frequency analysis plot.
9. The method of claim 8 , wherein said time-frequency analysis plot is a spectrogram showing spectral changes as a function of time.
10. The method of claim 1 , wherein finding said measure of power spectral density of said one or more enhanced kinematic signals involves using a power spectral estimation method chosen from the group consisting of periodogram, modified periodogram, Barletts's method, Welch's method, Blackman-Tukey's method, autoregressive spectrum estimation, moving average spectrum estimation, autoregressive moving average spectrum estimation, and adaptive spectrum estimation.
11. The method of claim 1 , wherein said one or more kinematic sensors comprise an accelerometer and a gyroscope.
12. A method for analyzing and visualizing kinematic data to characterize movement disorders, comprising:
(a) collecting one or more kinematic signals from one or more kinematic sensors in a wearable inertial monitoring system configured for continuous measurement of movement; and
(b) calculating and displaying a time-frequency analysis plot of said one or more kinematic signals to characterize movement disorders in Parkinson's disease or essential tremor as a function of time.
13. The method of claim 12 , wherein said time-frequency analysis plot is a spectrogram showing spectral changes as a function of time.
14. An apparatus for generating a plurality of movement impairment indices from one or more kinematic signals to characterize movement disorders, comprising:
(a) one or more kinematic sensors in an untethered wearable inertial monitoring system configured for continuous measurement of movement by collecting one or more kinematic signals; and
(b) a processor configured for 1) processing said one or more kinematic signals to reject artifacts and noise resulting in one or more enhanced kinematic signals, 2) finding a measure of power spectral density of said one or more enhanced kinematic signals, and 3) calculating one or more impairment indices to characterize movement disorders in Parkinson's disease or essential tremor from said measure of power spectral density.