United States Patent [19]
`Vonk et al.
`
`lIIIIIllllllll|||IlllllllllIlllllllllllllllllllllllllllllllllllllllllllllll
`5,293,879
`Mar. 15, 1994
`
`US005293879A
`[11] Patent Number:
`[45] Date of Patent:
`
`[54] SYSTEM AN METHOD FOR DETECTING
`TREMORS SUCH AS THOSE WHICH
`RESULT FROM PARKINSON’S DISEASE
`[75] Inventors: B. F. M. Vonk, Wehl; Eugenio
`Johannes W. Van Someren,
`Amsterdam, both of Netherlands
`[73] Assignee: Vitatron Medical, B.V., Netherlands
`[21] Appl. No.: 763,621
`[22] Filed:
`Sep. 23, 1991
`[51] Int. cu .............................................. .. A61B 5/00
`[52] US. Cl. .................................. .. 128/782; 364/508;
`364/413.05
`[58] Field of Search ............. .. 128/782, 779,784, 739;
`364/508, 413.01, 413.02, 413.05; 604/50;
`73/570, 579
`
`[56]
`
`References Cited
`U.S. PATENT DOCUMENTS
`
`3,163,856 12/1964 Kirby ................................ .. 128/782
`4,112,926 9/1978 Schulman et al.
`128/782
`4,117,834 10/1978 McPartland et al. .
`4,195,643 4/ 1980 Pratt, Jr. .............. ..
`4,306,291 12/1981 Zilm et a1. ....... ..
`
`
`
`4,353,375 10/1982 Colburn et a1. 4,444,205 4/1984 Jackson ........... ..
`
`
`
`4,520,674 6/1985 Canada et al. 4,822,337 4/1989 Newhouse et a1.
`
`4,951,674 8/1990 Zanakis et a1. ...... ..
`5,031,614 7/1991 Alt ........................... .. 128/419 OPG
`
`FOREIGN PATENT DOCUMENTS
`
`3446338 7/1986 Fed. Rep. of Germany .
`3916500 5/1989 Fed. Rep. of Germany .
`566559 8/1977 U.S.S.R. ............................ .. 128/782
`
`632345 11/1978 U.S.S.R. ............................ .. 128/782
`
`OTHER PUBLICATIONS
`
`Zilm et al. “The Nature . . . Tremor”, IEEE Transations
`on Biomedical Engineering, vol. BME 26, No. 1, Jan.
`1979, pp. 3-10.
`. Examination”, Medical
`.
`Pau, “Data Compression .
`and Biol. Eng. and Comput. vol. 15, May 1977, pp.
`292-297.
`.
`Colburn et al., “An Ambulatory Activity Monitor . .
`Memory”, ISA Transactions, vol. 15, No. 2, pp.
`149-154 (1976).
`Primary Examiner-Stephen C. Pellegrino
`Assistant Examiner-Guy V. Tucker
`Attorney, Agent, or Firm—Woodcock, Washburn,
`Kurtz, Mackiewicz & Norris
`[57]
`ABSTRACT
`A system and method are provided for accurate deter
`mination of various tremors and movements derived
`from Parkinson’s Disease and other causes. Movement
`signals are generated by an accelerometer, the move
`ment signals being analyzed to determine whether they
`are in a predetermined frequency range, and when there
`occurs an uninterrupted series of such movement sig
`nals that are within the frequency range. By determin
`ing when these two criteria are met, tremor signals are
`reliably discriminated from other activity signals. The
`system includes circuitry for collecting and storing data
`concerning the timing of tremor occurrences, the tim
`ing of other activity occurrences, and amplitude data
`concerning tremor occurrences.
`
`20 Claims, 6 Drawing Sheets
`
`[:1 REST
`
`s\\\\\\\\\ ACTIVITY 2
`
`{2m ACTIVITY 1
`
`-- TREMOR
`
`___ -1..- _... _ W
`
`' HALF-PERIOD
`DURATION
`
`(b)
`
`=
`i
`.
`
`i
`
`:i
`-:
`.,
`
`ii
`
`(0) 0145000001
`
`
`
`0 ///i4 o b c
`
`d
`
`e
`
`f
`
`0
`
`500
`
`1000
`
`1500
`
`msec.
`
`2000
`
`25'00
`
`3000
`
`FITBIT, INC. v. LOGANTREE LP
`Ex. 1011 / Page 1 of 12
`
`

`
`US. Patent
`
`Mar. 15, 1994
`
`Sheet Id 6
`
`5,293,879
`
`\ \ \
`
`
`
`255 o~\\ 5:22:32
`
`z E ass: 22 3 8E 52:; . games; 3%
`
`I!
`
`a _N\ a
`
`..§.~k
`
`a a a
`
`
`
`“=5. \ 5:: 2a:
`
`<2:
`
`5% 5:8
`
`Ex. 1011 / Page 2 of 12
`
`

`
`U.S. Patent
`
`Mar. 15, 1994
`
`Sheet 2 of 6
`
`5,293,879
`
`N»:>_5<
`
`5:H
`
`$55
`
`_:_>_:<
`
`82
`
`
`
`222%
`
`s=E.:s_Ev
`
`zo_E:_§sQ
`
`Ex. 1011 /Page3 Of12
`
`Ex. 1011 / Page 3 of 12
`
`
`

`
`US. Patent
`
`Mar. 15, 1994
`
`Sheet 3 of 6
`
`5,293,879
`
`60
`
`START,
`
`k
`
`WAIT F0 R INTERRUPL/
`63
`
`NO
`
`EDGE
`YES
`START TIMER I /ee
`INITIALIZE N=() /68
`INITIALIZE sd=oj
`
`EDGE
`YES
`READ TIMER 11,
`START TIMER
`
`TIMER OVERFLOW
`
`N0
`
`19
`
`Y
`
`5d
`Ad+
`
`WM
`
`NO
`
`8?
`
`84
`YES
`
`85
`
`J1
`
`8
`
`/89
`
`r@?1 YES
`
`N0
`
`881
`
`L
`II!’ 25’
`
`Ex. 1011 / Page 4 of 12
`
`

`
`U.S. Patent
`
`M
`
`Maw
`
`as:
`
`.-S..283 .
`
`888:3.2:
`2:u8._3.28E5:g-E;.2:
`E2_m
`
`M3
`
`
`
`.s.,22.5E95->9,
`
`S
`
`5,293,879
`
`
`
`
`525%:22%Fso
`MIn:-W.IIIIIan
`1.~zH__zH.
`
`w.3
`
`3mm;9
`
`Mo.h.~rK
`
`Ex. 1011 /Page5 Of12
`
`Ex. 1011 / Page 5 of 12
`
`
`
`
`
`

`
`US. Patent
`
`Mar. 15, 1994
`
`Sheet 5 of 6
`
`5,293,879
`
`Tb
`
`.44
`
`i
`
`HOUR 0F DAY
`
`'2
`
`I 1
`
`0
`
`I
`I
`I
`I
`l
`o o c O
`o
`_<_r_
`<2 c co <0 <- N
`
`I
`
`l
`
`Ex. 1011 / Page 6 of 12
`
`

`
`5,293,879
`
`Fig.4B
`
`HOUR0FDAY
`
`0123456789|0||l2|3|4|5A|6|7|8|9202|222324
`
`(“W”) SEIIHNIN
`
`Ex. 1011 / Page 7 of 12
`
`

`
`1
`
`SYSTEM AN METHOD FOR DETECTING
`TREMORS SUCH AS THOSE WHICH RESULT
`FROM PARKINSON’S DISEASE
`
`BACKGROUND OF THE INVENTION
`
`-
`
`1. Field of the Invention
`This invention relates to a system and method for
`detecting tremors such as derived from Parkinson’s
`Disease and, more particularly, to such systems and
`methods which involve detecting movement and pro
`cessing signals representative of movement so as to
`discriminate tremor'derived signals from activity and
`non-tremor signals.
`2. Description of the Prior Art
`Monitoring of patients to determine the occurrence
`of Parkinson’s tremors is known in the art, although
`very few long-term tremor registrations have been re
`ported in the literature. The literature has reported
`24-hour recordings, but no easily applicable and useful
`tool has been provided. Prior known monitoring sys
`tems generally comprise some arrangement for taking
`an electromyogram or accelerometer signal and trans
`mitting received signals and/or data to an on-line re
`corder. Data analysis normally takes place off-line,
`which is very time consuming. Other systems, so-called
`actometers, generally count movements of a patient’s
`limb, e. g., arm or leg, but do not accurately discriminate
`between movements representative of tremors and
`those representative of other forms of movement or
`activity. Moreover, no index of the average accelera
`tion of movements is given in such a system. See, for
`example, US. Pat. No. 4,353,375, which discloses the
`general concept of detecting arm or leg movement, but
`I which does not suggest how to separate out tremor
`signal from other activity signals.
`Our investigations have con?rmed the shortcomings
`of the prior art. We have registered movements in ten
`Parkinsonian patients (both sexes, ages 43 to 89) and
`twenty healthy volunteers (both sexes, ages 26 to 67).
`Each patient wore an accelerometer on the wrist, the
`output of which was continuously recorded on a porta
`ble analog instrumentation recorder for 24 hours. Re
`cordings were ambulant, and no restrictions on move
`ments were made. Data were sampled with a frequency
`of 100 Hz for off-line computer analysis. The attempt to
`discriminate tremor from non-pathological movements
`by frequency was not satisfactory: fast Fourier trans
`forms showed that the movement patterns overlapped
`in frequency range. Thus, there remains a need for a
`system and method which reliably discriminates reliable
`tremor signals from activity and non-tremor-induced
`signals. There further remains a need in the art for ob
`raining data by which tremor activity can be classi?ed
`and analyzed so as to optimally indicate desired treat
`ment, i.e., drug level setting by the physician automatic
`drug injection.
`
`20
`
`5,293,879
`2
`motor acceleration (hyperkinesia) as observed in chil
`dren.
`It is a further object to process the tremor signals so
`as to obtain accurate tremor histories, in terms of signal
`characteristics and timing of tremor signals, and to pro
`vide for patient treatment as a function of such histories.
`It is a still further object to provide the basic design
`components of a system for ambulatory monitoring of
`Parkinsonian tremors, which can be adapted into a min
`iaturized wristwatch-type device worn by the patient.
`Our invention is based upon the observation that
`while non-tremor signals overlap the tremor frequency
`range (about 3-12 Hz), it is improbable that a series of
`signals within the tremor range would represent any
`thing but a tremor, so long as the series exceeds a criti
`cal length. Thus, for example, if a critical series length
`(CSL) of eight or more consecutive signals are detected
`within the tremor range, without any intervening peri
`ods, then the signal can reliably be recognized as a
`tremor. In normal limb movement, a series of 3-12 Hz
`periods will seldom reach this CSL, while most series in
`tremor will exceed the CSL.
`In accord with the above objects, there is provided
`by this invention a system and method for monitoring
`tremors in the limb of a patient, the system having an
`accelerometer or like means for detecting limb move
`ments and for generating movement signals representa
`tive thereof, and process circuitry for programmed
`processing of the signals with means to ?lter out signals
`outside of the frequency range of the tremor signals.
`The system further comprises analyzing means, prefera
`bly involving a microcontroller, for determining the
`half-cycle time periods of the movement signals as well
`as a representation of signal amplitude, and means for
`identifying a tremor as constituting a series of consecu
`tive signals having a given range of time periods within
`a programmed range. For example, the system may
`look for movements having time periods within the
`range of about 40-170 ms, and for at least twelve con
`secutive such periods. Further, in a preferred embodi
`ment, tremor amplitude information is stored, along
`with histogram data relating to tremor and non-tremor
`activity. The system comprises utilizing tremor ampli
`tude and timing data, as well as activity data, for provid
`ing indications of patient treatment. The combination of
`discriminating out many of the non-tremor signals, de
`termining the occurrence of consecutive signals within
`the tremor frequency range, and storage and processing
`of such tremor-range signals, results in a system with
`enhanced ability for reliably detecting movement and
`tremor ?uctuations as seen, for example, in patients
`with Parkinson’s disease. The invention is illustrated
`with an embodiment that includes external processing
`capabilities, such as a portable computer, but also em
`braces a portable device of wristwatch size, which can
`automatically control a dispensing unit such as an im
`plantable pump. The invention is illustrated with re
`spect to Parkinsonian tremors, but is equally applicable
`to other types of tremors.
`
`15
`
`25
`
`35
`
`45
`
`SUMMARY OF THE INVENTION _
`It is a primary object of this invention to provide a
`system and method for monitoring tremors of Parkin
`son’s Disease and the like, the system having capability
`for discriminating against activity and other non-tremor
`65
`signals so as to provide signals reliably re?ective of
`tremor. As used in describing this invention, tremor also
`includes so-called motor slowing (bradykinesia) and
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`FIG. 1 is block diagram showing the primary system
`components of this invention.
`FIG. 2A is a graph illustrating a limb movement
`signal over 3 seconds and the logical analysis for distin
`guishing rest, activity and tremor. FIG. 2B is a simpli
`?ed flow diagram of the logical steps for distinguishing
`these states.
`
`Ex. 1011 / Page 8 of 12
`
`

`
`10
`
`25
`
`30
`
`40
`
`45
`
`5,293,879
`3
`4
`FIG. 3 is a block diagram showing the primary hard
`In the logical analysis of tremor signals according to
`the invention, ?rst, the period time of the incoming
`ware components of the apparatus of this invention.
`FIG. 4A is a representation of a histogram showing
`signal is speci?ed to be between certain limits, e.g.,
`mean tremor amplitude as obtained in the system of this
`about 40-170 ms. Note that these durations, represent
`invention.
`ing half cycles as seen in curve (a) of FIG. 2A, corre
`FIG. 4B is a graph illustrating histogram data repre
`spond to frequencies of about 12 and 3 Hz respectively,
`senting tremor and activity occurrence as obtained with
`i.e., a range of 3-12 Hz.
`the system of this invention.
`In addition, a further basic step is taken, namely car
`rying out of a sequence analysis, to count the number of
`repetitions (without break) that meet the period criteria.
`For example, if at least twelve half-periods are uninter
`rupted by “rest” or “activity", these sequences are clas
`si?ed as tremors. Depending on the frequency of the
`tremor, this criterion requires an uninterrupted tremor
`signal somewhere between about 0.48 seconds (for a
`12.5 Hz tremor) and 2.04 seconds for a 2.94 Hz tremor.
`Still referring to FIG. 2A, the signal is shown with
`respect to a zero reference line, and with respect to
`dashed threshold lines. Zero crossings are used to re
`duce the signal to durations and maximum absolute
`values between successive zero crossings, i.e., half-peri
`ods and their amplitudes. Plus and minus thresholds are
`arbitrarily established at low levels, to avoid scoring of
`minor fluctuations around zero as movement or tremor.
`Thus, whenever the signal between two or more zero
`crossings does not exceed a threshold, as seen at epi»
`sodes (c) and (e) of curve (c), it is determined to be
`“rest.” Of the remaining signals, the half-periods with a
`duration between 40 and 170 ms are selected as indi
`cated at line (b). Half-periods with a shorter or longer
`duration (as seen at (b) and (a) respectively in line (c))
`are classi?ed as “activity 1.” The next step, based on the
`above assumed characteristics of tremor and other
`movements, is to select from the remaining signals only
`sequences of at least twelve half-periods uninterrupted
`by “rest” or “activity 1." These sequences are classi?ed
`as tremor, as seen at (f) on line (c) of FIG. 2A, while
`shorter sequences are classi?ed as “activity 2", as seen
`at (d) on line (c). From these classi?cations, total time of
`epochs determined as tremor, activity 1 and/or activity
`2 can be determined, e.g., hourly, as well as the average
`amplitude of signal during these epochs and the mean
`duration of the epochs.
`Referring now to FIG. 28, there is shown a flow
`diagram illustrating an algorithm for carrying out the
`logical steps in this invention in determining a tremor.
`In FIG. 2B, the following symbols are used:
`T=measured half-period time
`Sd=serial duration of half-periods
`Ad=total duration of activity
`Td=total duration of tremor
`N=number of cycles
`CSL=critical series length
`Th=programmable high time for half-period (e.g.,
`170 ms)
`T1=programmable low time for half-period (e.g., 40
`ms)
`The algorithm starts as indicated at 60, and waits for
`an interrupt at 62. Upon an interrupt, it is determined at
`63, whether the edge of a half signal, or crossover has
`been detected. If no, the interrupt pertains to another
`matter, and the routine switches to interrogate the
`marker, RS 232 or RTC. If yes, the routine proceeds to
`66, where the timer is started, to time out the interval T
`following the start of the signal. At 68, Sd and N are
`initialized to zero. At 70, it is determined whether an
`other edge has been detected, or whether there has been
`timer over?ow, meaning that a period has elapsed with
`
`DESCRIPTION OF THE PREFERRED
`EMBODIMENTS
`Referring now to FIG. 1, there is shown an overall
`block diagram of the system of this invention. Block 20
`represents an accelerometer device package with inter
`face electronics, which is suitably attached to the pa
`tient's arm or other limb. The accelerometer passes its
`signal to the signal processing and storage apparatus 21,
`which generally comprises hardware and software for
`processing the accelerometer output, as well as memory
`storage for holding algorithm software and stored data
`20
`obtained from the accelerometer. In practice, apparatus
`21 may be a portable processing unit packaged sepa
`rately from the acceleometer sensor 20. However, the
`invention embraces packaging all of this apparatus in a
`wristwatch-like device, capable of registering and stor
`ing tremor data on a ?xed interval, e.g., hourly basis,
`over a duration of several weeks.
`Further as shown in FIG. 1, there is shown an audio
`output 22 which receives output signals from block 21,
`and provides an audio output of a predetermined nature
`to the patient. The audio output may, for example, sig
`nal to the patient when the battery is low, or may pro
`vide coded signals representing a variety of device con
`ditions. Patient data input block 23 represents a means
`whereby the patient can transmit signals to device 21,
`either for immediate use or storage. Interface 26 is in
`two way communication with device 21, and in turn is
`in two way communication with processor 27, which
`may suitably be a personal computer. The personal
`computer can be used to program processing variables
`in apparatus 21, and may also be used in a situation
`where the device is interrogated to read out data that
`has been stored over a duration of time, for analysis and
`display. The computer may further be used to control a
`pump 28, which is used for dispensing medicine such as
`L-Dopa at an optimum rate and amount as determined
`by processing of the patient data. The invention also
`embraces the design of processor 27 as a dedicated unit
`for carrying out the treatment analysis, in which case it
`may be incorporated into a single miniaturized system
`package worn by the patient. In such a case, processor
`27 may directly control an implantable pump 28.
`Detection and discrimination of the abnormal move
`ments constituting a tremor caused by Parkinson’s Dis
`ease, or another cause as contrasted to normal activity
`movements, is based primarily on two criteria, leading
`to a period amplitude sequence analysis (PASA). It is
`known that the dominant frequency for Parkinson’s
`disease, for example, varies between 3 and 7 Hz for a
`rest tremor, and up to 12 Hz for a postural tremor.
`However, as noted above, a tremor cannot be reliably
`distinguished from normal movements on the basis of
`frequency above. Such tremors do differ from normal
`movements in the number of repetitions of signals
`within a con?ned frequency band. We de?ne tremor
`operationally as a sequence of repetitive movements
`both within a given frequency band and enduring for a
`critical length of time.
`
`55
`
`60
`
`65
`
`Ex. 1011 / Page 9 of 12
`
`

`
`15
`
`25
`
`35
`
`40
`
`5,293,879
`6
`5
`tered signals. At the end of every incoming event,
`out any crossover. If another signal edge is detected,
`whether tremor or activity, the peak value is converted
`then at 71 the timer is read and restarted. The elapsed
`by an AD convertor of the microcontroller, and such
`time T of the timer, is compared at 72 to T;,, and at 73
`values are averaged over a predetermined programma
`to T1. If T is within the two limits, then at 74 N is incre
`ble period and stored in an EEprom. The peak detector
`mented by l, and the serial duration of half-periods (Sd)
`is incremented by the time of the interval just elapsed,
`is reset after an A/D conversion by a logic signal from
`an appropriate port line of the controller.
`i.e., T. The routine then loops back to 70 to wait for the
`next edge.
`The microcontroller is suitably a 68HC805B6. This
`controller is indicated because of its low'power con
`Referring back to block 72, if T is greater than T},
`(representing a frequency lower than the low frequency
`sumption, 6 k EEprom program memory, on-board 8
`channel A/D converter, serial interface capability and
`limit), the program branches to 84 where it is deter
`flexible timer/comparator system. Because the proces
`mined whether N has reached CSL. If yes, this means
`sor is a DC-type, its clock frequency can go down to 0
`that a tremor has been detected, and the total tremor
`duration is incremented by Sd. If no, meaning activity
`Hz. For this reason, the clock signal is generated exter
`nally by oscillator circuit 43, which provides a 76.8 kHz
`but no tremor, then the total activity duration is incre
`signal having a good time base for the serial interface
`mented by Sd. Likewise, at 73, if T is less than T}, the
`(1200 Band) and a good measuring value forthe timer
`program branches to 87 where again it is determined
`input. The clock signal can be put on hold by a logic
`whether N is equal to or greater than CSL. If yes, total
`signal on one of the microcontroller lines. The clock
`duration of tremor is incremented by the accumulated
`Sd plus the most recent interval T; if no, total duration
`can be started by several events; a signal from the com
`of activity is incremented by Sd+T.
`parator, a marker input, or a real time clock over?ow
`can start up the clock again.
`Referring now to FIG. 3, the accelerometer unit is
`Block 44 represents a real time clock, which operates
`shown at 30. The accelerometer may be a model 12
`separately from the microcontroller so as not to disturb
`PICOCHIP TM , made by Endevco. This is a miniature
`normal microcontroller routines. The RTC has its own
`accelerometer based on bimorph bender technology,
`32.768 Hz crystal; and has an interrupt capability for
`and has a size of about 3.6X4.6>< 1.5 mm. The sensor is
`interrupting the processor every programmable period,
`basically a charge generating device which is capable of
`e.g., every hour or half hour, to store total counted data
`detecting arm movements. The charge generates a volt
`in EEprom 45. The RTC generates an interrupt vector
`age across a resistor when there is a movement of the
`arm; this voltage is ampli?ed by about 100, and buffered
`which is connected to a suitable line of the microcon
`by a low power operational ampli?er, before outputting
`troller. A second EEprom 46 is also utilized, as shown,
`such that main storage of data is provided by two 2 k
`to the following processing circuitry. The accelerome
`byte EEproms.
`ter and interface electronics are mounted in a metal
`Interface unit 47 is preferably an RS232 transceiver
`case, connected to electrical ground. The output from
`used to connect the device which is mounted on the
`accelerometer 30 is connected to a programmable am
`patient to processor or computer 27.
`pli?er 31, which receives program > signals from mi
`Still referring to FIG. 3, power supply 35 generates
`crocontroller 40 across bus 41. Ampli?er 31 may, for
`required + 5 and -5 voltages for the circuitry out of a
`example, have eight programmable values, 5 X to 40X ,
`9 volt battery, and also generates a Ub/2 signal for
`providing an overall gain in the range from 500 to 4000.
`measuring of actual battery voltage. This voltage is
`These programmable values are held by respective port
`measured suitably every hour by the microcontroller on
`lines of the microcontroller 40.
`an analog line of the microcontroller. The power supply
`The output of ampli?er 31 is connected to a program
`mable ?lter 32 which has programmable low and high
`circuit also generates a power up reset signal, as indi
`pass roll off points. The low pass point is suitably pro
`cated. A sound interface 50 is utilized as a driver for
`piezo electric tone generator 51. Frequencies for several
`grammable to values between 0.5 and 5 Hz with steps of
`45
`tones are generated by the programmable timer
`0.5 Hz. Respective port lines of the microcontroller are
`TCMPl of the microcontroller. For example, when the
`used to provide the program values. In practice, the
`high pass point is programmable between about 8 and
`value of Ub/Z reaches a predetermined level, the mi
`crocontroller provides an alert signal through sound
`12 Hz in steps of 0.5 Hz. The ?lter roll off characteris
`interface 50 and generator 51. The length and frequency
`tics are 12 dB/oktaaf. By this means, the ?lter passes
`of the signal are programmable. There is also provided
`only signals within the desired period range, e.g., 3 to 8
`a marker interface 53, which holds information about
`Hz.
`the switch positions of switches indicated as 1 and 2.
`The output from the ?lter is connected through to
`These marker switches can be used by the patient to
`comparator 33, which suitably operates as a cross-over
`signal predetermined occurrences such as medicine use,
`detector. As seen also in connection with FIG. 3, the
`sleep time, eat time, etc. The marker interface is con
`output of the comparator produces a pulse for the time
`period when the signal has exceeded a given positive
`nected through to suitable lines of the microcontroller,
`and can be reset by a logic signal on another one of the
`threshold above reference value, and likewise when the
`negative swing of the signal exceeds a corresponding
`lines.
`negative value. The comparator is programmable
`The software for operating the microcontroller con
`tains standard routines for handling the following:
`through controller 40 to eight different detection levels,
`1. Interrupt handling
`ranging from 0.25 v to 2 v, which values are provided
`by respective port lines of the microcontroller. The
`. Timer interrupt handling
`output pulses from the comparator 33, which represent
`. RTC interrupt handling
`signals within the programmed period (frequency)
`. Marker interrupt handling
`range, are inputted to microcontroller timer input
`. A/D Conversion
`TCAPl. The output of ?lter 32 is also connected to a
`. Oscillator start and stop
`peak detector 34 for measuring peak values of the ?l
`. Sound generation routine
`
`
`
`50
`
`55
`
`65
`
`‘Pusan crw
`
`Ex. 1011 / Page 10 of 12
`
`

`
`20
`
`5,293,879
`8
`7
`3. The system as described in claim 1, comprising
`5. Serial data storage routine, i.e., transferring data
`means for generating and storing tremor histogram data
`relating to tremor identi?cation, detected amplitude
`from said signals.
`information, and time of tremor from microcontroller
`4. The system as described in claim 1, comprising
`to EEprom.
`amplitude means for determining the amplitude of said
`6. Communication routine for the RS232 interface.
`signals and for generating and storing amplitude data
`The system and method of this invention further
`representative of said amplitudes.
`encompass techniques for evaluating the tremor data, to
`5. The system as described in claim 4, wherein said
`determine desired treatment. Referring to FIGS. 4A ,
`amplitude means comprises means for generating peak
`and 48, there are shown representative histograms
`amplitude data corresponding to said signals.
`which are determined, suitably by personal computer
`6. The system as described in claim 4, comprising
`27, and displayed on an appropriate display, to provide
`information from which a physician may select treat
`treatment means for indicating treatment as a function
`of said signal amplitude data.
`ment. In FIG. 4A, a histogram of mean tremor ampli
`7. The system as described in claim 1, comprising
`tude is presented, and the physician evaluates this pat
`time means for generating and storing time signals cor
`tern of mean amplitude as a function of time of day in
`responding to the time of detected tremors.
`order to determine appropriate treatment. In FIG. 413,
`8. The system as described in claim 7, comprising
`there is shown a dual histogram of tremor and activity,
`with the remaining time being classi?ed as rest. Upon
`treatment means for indicating treatment as a function
`of said time signals.
`the presentation of this histogram, the physician can
`9. The system as described in claim 8, wherein said
`determine appropriate treatment as a function of the
`treatment indicating means comprises means for deter
`time of sensed tremors, or as a function of the time of
`mining data representative of normal patient activity
`the tremors and the time of other sensed activity. Of
`and for indicating treatment as a function of the timing
`course, for a miniaturized totally portable system worn
`of tremors and the timing of normal activity.
`by the patient, where the microprocessor or other dedi
`cated circuitry is part of the package worn directly by
`10. The system as described in claim 1, further com
`prising marker interface means for enabling the patient
`the patient, other criteria for evaluating the data may be
`to input time data into said system.
`used so as to directly control a pump 28 or other suit
`11. The system as described in claim 1, further com
`able treatment rendering apparatus.
`prising sound output means for outputting an audio
`It is to be understood that the system and method of
`signal representative of a detected condition of said
`this invention are not limited in scope to the speci?c
`system.
`embodiment illustrated. Thus, while the speci?c mi
`12. The system as described in claim 1, comprising
`crocontroller illustrated can carry out the functions of
`storage means for storing tremor data representative of
`counting pulses, determining the number of pulses in a
`series, and storing pulse amplitude values, these and
`said tremor occurrences, a pump for introducing a me
`dicinal liquid into said patient for treatment of tremors,
`similar functions can be carried out by other well
`and means responsive to said stored data for operating
`known forms of hardware and/or software. Also, while
`the preferred embodiment is a wristwatch-like package
`said pump.
`13. The system as described in claim 1, comprising a
`which carries out all the processing, the invention em
`programmable ?lter for ?ltering said generated repre
`braces other forms, e.g., separate accelerometer and
`processing units.
`sentative signals so as to pass only signal frequencies
`within a programmable range.
`As used in the claims, the terms “period” and “half
`14. The system as described in claim 1, wherein said
`period" are the same, i.e., the pulse formed from a posi
`detecting means comprises an accelerometer.
`tive or negative portion of a signal, between zero cross
`15. The system as described in claim 1, comprising
`ings, can be described as having a period or half period
`storage means for storing tremor data representative of
`T.
`What is claimed:
`said tremor occurrences, computer means for process
`ing said tremor data, and linking means for linking said
`1. A system for monitoring tremors in the limb of a
`patient, comprising
`stored tremor data to said computer means.
`16. The system as described in claim 1, comprising
`means for detecting limb movements and for generat
`threshold determining means for determining when
`ing signals representative of said limb movements,
`signals being analyzed do not exceed a predetermined
`analyzing means for analyzing said signals to deter
`mine signal periods,
`minimum amplitude, indicating a rest period.
`storage means for storing signal period criteria, said
`17. The system as described in claim 1, wherein said
`determining means comprises rest means for determin
`criteria comprising data representing a range of
`ing when there is an interruption of signals having sig
`signal periods within about 40-170 ms and charac
`nal periods within said range, thereby indicating rest or
`teristic of tremors, and a number greater than zero
`inactivity.
`corresponding to a minimum number of consecu
`tive signals with periods in said range which repeat
`18. A method of detecting a tremor of an ambulatory _
`patient, comprising
`uninterruptedly without intervening periods, and
`establishing frequency and duration criteria applica
`determining means for determining when said signal
`periods of said signals are within said range and for
`ble to the movements of a tremor to be detected,
`continuously obtaining signals representative of limb
`determining when said signals with periods within
`movement of said patient,
`said range repeat at least said minimum number of
`determining when said signals meet said frequency
`times without interruption, whereby tremor occur
`criterion, and
`rences are detected.
`determining when said signals meeting requirements
`2. The system as described in claim 1, wherein said
`of said frequency criterion continue substantially
`number is eight.
`
`65
`
`30
`
`40
`
`Ex. 1011 / Page 11 of 12
`
`

`
`10
`9
`?rst In?“ for deter_mining when $_aid_5igna15 meet
`without any intervening periods over a time period
`requirements of said frequency criterion, and
`corresponding to said duration criterion.
`second means for determining when said signals
`.
`_
`_
`meeting requirements of said frequency criterion
`19. A system for detecting a tremor of a patient, said
`continue substantially without any intervening
`system having means for continuously Obtaining and 5
`periods over a_ time period corresponding to said
`analyzing signals representative of limb movement of
`dunno“ cme?on
`said patient, characterized by
`20. The system as described in claim 19, comprising
`accumulating means for accumulating determined dura
`means,for sto