Detection of hypokinetic and hyperkinetic states

• US 9,826,921 B2
• Filed: 06/12/2009
• Issued: 11/28/2017
• Est. Priority Date: 06/12/2008
• Status: Active Grant

• Current Assignee: Global Kinetics Corporation Limited
• Original Assignee: Global Kinetics Corporation Limited
• Inventors: Griffiths, Robert Irwin, Horne, Malcolm Kenneth
• Primary Examiner(s): CERIONI, DANIEL LEE
• Application Number:
  US12/997,540
• Publication Number:
  US 20110098608A1
• Time in Patent Office: 3,091 Days
• Field of Search: 600587, 600595
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: A61B 5/1104 induced by stimuli or drugs; A61B5/1102 takes precedence; investigating medicinal preparations G01N33/15
• CPC Class Codes: A61B 5/1114 Tracking parts of the body
• CPC Class Codes: A61B 5/1123 Discriminating type of movement, e.g. walking or running A61B5/1116, A61B5/112 take precedence
• CPC Class Codes: A61B 5/4082 Diagnosing or monitoring movement diseases, e.g. Parkinson, Huntington or Tourette
• CPC Class Codes: A61B 5/6824 Arm or wrist
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7214 using signal cancellation, e.g. based on input of two identical physiological sensors spaced apart, or based on two signals derived from the same sensor, for different optical wavelengths

Abstract

The present invention relates to an automated method of determining a kinetic state of a person. The method obtains accelerometer data from an accelerometer worn on an extremity of the person and processes the accelerometer data to determine a measure for the kinetic state. The present invention further relates to a device for determining a kinetic state of a person. The device comprises a processor configured to process data obtained from an accelerometer worn on an extremity of the person and to determine from the data a measure for the kinetic state. In the method and system the kinetic state is at least one of bradykinesia, dyskinesia, and hyperkinesia.

Claims

1. An automated method of determining a bradykinetic state of a person, which is a state characterized by slowness of movement relative to a normally kinetic person and in which movements involving peaks of acceleration have strong low frequency components, the method comprising:

obtaining a time series of accelerometer data in real-time from an accelerometer worn on an extremity of the person, the time series of accelerometer data being representative of movement of the person over an extended period of time, wherein the time series of accelerometer data is obtained while the accelerometer is oriented to be sensitive to at least one of pronation and supination movements of the person and is attached to at least a limb of the person;

filtering, by a band pass filter, the accelerometer data to extract filtered data for a band of interest, wherein the band of interest has a lower end cut off frequency selected to remove DC;

extracting, by a computer processor, a plurality of bins of acceleration data from the time series of the filtered accelerometer data;

searching, by the computer processor, the data in each of the plurality of bins to identify a time window of peak acceleration, the time window within a given bin comprising at least two data samples that capture accelerometer data for a single movement having a peak acceleration relative to other movements observed within the plurality of bins; and

processing, by the computer processor, data in the time windows of peak acceleration to determine;

spectral components for each of the single movements having peak acceleration; and

an objective measure of the extent to which the single movements having peak acceleration identified within the plurality of bins of acceleration data are slower than expected from a normally kinetic person and indicative of bradykinesia based upon a comparative statistical analysis of the determined sets of spectral components of the single movements of peak acceleration that occur over the extended period of time.

2. The method according to claim 1, wherein the plurality of bins of acceleration data are of a time duration selected to be small enough that relatively regular measures of the extent to which the single movements having peak acceleration are slower than expected from a normally kinetic person and indicative of bradykinesia are determinable, while being long enough to provide a reasonable likelihood of a significant movement by the person during that bin.

3. The method according to claim 2, wherein the plurality of bins have durations within the range of 15 seconds to four minutes.

4. The method according to claim 1, wherein the searching comprises using a moving mean having a window length in a range of 0.02 seconds to 30 seconds in order to find a maxima.

5. The method according to claim 1, further comprising obtaining a sub-bin which comprises a plurality of data points both before and after a peak acceleration PK_i, and wherein the plurality of data points is a number of data points which is a power of two, and wherein the sub-bin is symmetrically positioned about the peak acceleration, and wherein the data points cover a period of time in a range of 0.5 seconds to 30 seconds.

6. The method according to claim 5, further comprising performing a spectral analysis of the sub-bin to obtain sub-band spectral measures, and wherein the sub-band spectral measures are of a width in the range of 0.6 Hz to 1 Hz.

7. The method according to claim 6, further comprising applying a weighting to at least a subset of the sub-band spectral measures to produce a weighted mean spectral power (MSP_i), wherein a greater measure of the extent to which the single movements having peak acceleration are slower than expected from a normally kinetic person and indicative of bradykinesia is given when the maximum (MSP_i) is small and exists in lower frequency sub-bands, and a lesser measure is given when the maximum (MSP_i) is high and exists in higher frequency sub-bands.

8. The method according to claim 7, further comprising:

selecting a plurality n of consecutive bins;

determining a peak acceleration (PK_i) and a MSP_i for each selected bin, and from across the n bins selecting the largest value of PK_i (PK_i.max); and

computing at least one bradykinesia score BK according to at least one of the following formulae;

BK=PK_i.max×

MSP_i.max;

and

BK=A×

log_c (PK_i.max×

MSP_i.max)−

B

where A, c and B are selectable tuning constants and MSP_i.max is the largest value of MSP_i.

9. The method according to claim 8, further comprising calculating a cumulative measure comprising a sum of individual measures determined in order to provide a cumulative indication of the bradykinetic state.

10. The method of claim 4 wherein the window length is a fraction of the duration of a human motion.

11. The method according to claim 8, wherein a moving median or moving average of the at least one measure of the extent to which the single movements having peak acceleration are slower than expected from a normally kinetic person and indicative of bradykinesia within a moving window is calculated, to give a probabilistic determination of the likelihood of bradykinetic behaviour.

12. The method according to claim 5, further comprising performing a spectral analysis of the sub-bin to obtain sub-band spectral measures, and wherein the sub-band spectral measures are of a width in the range of 0.1 Hz to 2 Hz.

13. The method of claim 1, further comprising:

displaying, on a visual monitor, an involuntary movement score corresponding to the objective measure determined in the processing step.

14. The method of claim 1, wherein the filtering causes a reduction of accelerometer data.

15. The method of claim 1, wherein the limb is an arm of the person and the accelerometer is worn on a distally located part thereof.

16. A wearable device to assist in determining a bradykinetic state of a person, which is a state characterized by slowness of movement relative to a normally kinetic person and in which movements involving peaks of acceleration have strong low frequency components, the device comprising:

an accelerometer arranged to be worn on at least a limb of the person, wherein the accelerometer is oriented to be sensitive to at least one of pronation and supination movements of the person;

a band pass filter for band pass filtering a time series of accelerometer data obtained in real-time from the accelerometer worn on at least the limb of the person to extract filtered data for a band of interest, the time series of accelerometer data being representative of movement of the person over an extended period of time, wherein the band of interest has a lower end cut off frequency selected to remove DC; and

a processor configured to perform operations comprising;

extracting a plurality of bins of filtered accelerometer data from the time series of the accelerometer data,

searching the data in each of the plurality of bins to identify a time window of peak acceleration, the time window within a given bin comprising at least two data samples that capture filtered accelerometer data for a single movement having a peak acceleration relative to other movements observed within the given bin, and

processing the data in the time windows of peak acceleration to determine;

spectral components for each of the single movements having peak acceleration; and

an objective measure of the extent to which the single movements having peak acceleration identified within the plurality of bins of filtered accelerometer data are slower than expected from a normally kinetic person and indicative of bradykinesia based upon a comparative statistical analysis of the determined sets of spectral components of the single movements of peak acceleration that occur over the extended period of time.

17. A non-transitory computer program product for determining a bradykinetic state of a person, which is a state characterized by slowness of movement relative to a normally kinetic person and in which movements involving peaks of acceleration have strong low frequency components, the computer program product comprising:

a processor;

a memory containing a computer-executable instruction that configures the processor to;

obtain a time series of data in real-time from an accelerometer worn on an extremity of the person, the time series of accelerometer data being representative of movement of the person over an extended period of time, wherein the time series of accelerometer data is obtained while the accelerometer is oriented to be sensitive to at least one of pronation and supination movements of the person and is attached to at least a limb of the person;

band pass filter accelerometer data to extract filtered data for a band of interest, wherein the band of interest has a lower end cut off frequency selected to remove DC;

extract a plurality of bins of acceleration data from the time series of the filtered accelerometer data;

search the data in each of the plurality of bins to identify a time window of peak acceleration, the time window within a given bin comprising at least two samples that capture accelerometer data for a single movement having a peak acceleration relative to other movements observed within the given bin; and

process the data in the time windows of peak acceleration to determine;

spectral components for each of the single movements having peak acceleration; and

an objective measure of the extent to which the single movements having peak acceleration identified within the plurality of bins of acceleration data are slower than expected from a normally kinetic person and indicative of bradykinesia based upon a comparative statistical analysis of the determined sets of spectral components of the single movements of peak acceleration that occur over the extended period of time.