http://www.diva-portal.org
Preprint
This is the submitted version of a paper presented at Twentieth International Congress of Parkinson's Disease and Movement Disorders, Berlin, Germany, June 19-23, 2016.
Citation for the original published paper:
Somayeh, A., Memedi, M., Nyholm, D., Senek, M., Medvedev, A. et al. (2016)
Quantification of upper limb motor symptoms of Parkinson's disease using a smartphone. In: Twentieth International Congress of Parkinson's Disease and Movement Disorders
N.B. When citing this work, cite the original published paper.
Permanent link to this version:
Objective
P’ is a multidimensional and complex disorder affecting motor and non6 motor functionalitiesV Essessments of P’ symptoms are usually done by clinical rating scalesV One of them is the Unified P’ Rating Scale ]UP’RST5 developed to provide comprehensive5 efficient and flexible means to monitor P’6related disability and impairmentV It has been the most commonly used rating scaleV It is composed of four main parts where the third part is designed for rating of motor symptomsV However5 UP’RS has relatively poor inter6rater reliability [05 9]V Enother scale that is used to grade motor function of patients is Treatment Response Scale ]TRST[k5 7]V E limitation is that there is no general agreement on which parts of the symptomatology should be included in the TRS score [B]V However5 the scales have low intra6 and inter6rater reliability and their use is limited5 as they are only used during in6clinic observations [8]V
Background
Methods
Results
The correlation coefficients between SVM predictions and mean clinical ratings were as followsF :VB< for TRS5 :V8 for dyskinesia score5 :VB9 for item 9k of UP’RS ]finger tapsT5 :V7W for item 9B of UP’RS ]rapid alternating movements of handsT5 and :VBW for item k0 of UP’RS ]body Öradykinesia and HypokinesiaTV The spiral test had a good test6retest reliability with a coefficient of :VC75 indicating that spiral scores are stable and consistent over timeV When assessing the ability of the PGs to distinguish between patients and healthy controls the means of k out of W PGs ]PG05 PG9 and PG7T were different between the two groups ]p•:V::BTV @igure 9 shows clinical and predicted scores for two representative patientsV
Conclusions
The upper limb motor tests of the smartphone were able to
capture important and relevant symptom information of the
clinical rating scalesV The methods for quantifying the upper limb
motor symptoms of P’ patientsF
• Had adequate correlations to clinical ratings
• Were able to differentiate between movements of patients and
healthy controls5 and
• ]Spiral testsT had good test6retest reliabilityV
Somayeh Eghanavesi 05 Mevludin Memedi 0595 ’ag Nyholm k5 Marina Senek k5 Elexander Medvedev 75 Håkan Eskmark k5 Sten6Magnus Equilonius k5 @ilip Öergquist B5
Radu Gonstantinescu 85 @redrik Ohlsson W5 Jack Spira C5 Sara Lycke <5 and Enders *ricsson W
0 Gomputer *ngineering5 ’alarna University5 9 Informatics5 School of Öusiness5 Örebro University5 k Neuroscience5 Neurology5 Uppsala University5 7 Information
Technology5 Uppsala University5 B ’eptV of Pharmacology5 University of Gothenburg5 8 ’eptV of Glinical Neuroscience5 University of Gothenburg5 W Ecreo Swedish IGT5
C Sensidose EÖ5 < Genvigo EÖ
The aim of this study is to develop and evaluate methods for quantifying motor symptoms in Parkinson’s disease ]P’T using combined upper limb motor test data5 collected during tapping and spiral drawing tasks by a smart phoneV
Figure 1. Spiral drawing and Taping tests on the smartphone.
Press alternatively left and right buttons as fast as possibleV Test takes F’ secondsV Using the right handV Spiral drawing
Figure 2. Results of methods’ predictions versus actual clinical rating scores for two random patients (patient A, first column; patient B, second column)
Participants
Nine‘teen patients diagnosed with P6 and FF healthy controls were recruited in a single center= open label= single dose clinical observational study in SwedenV Simultaneous clinical‘ and smartphone‘based measures were collected up to +M times following a single levodopaOcarbidopa morning dose jM’( over normal to induce dyskinesiaszV
Clinical assessment
Subjects were asked to perform standardized motor tests in accordance with UP6RS and were videotapedV The videos were blindly rated by three movement disorder specialistsV The ratings were given based on Treatment Response Scale jTRSz ranging from ‘; x ‘Very Off’ to ’ x ‘On’ to 3; x ‘Very dyskinetic’= three UP6RS motor items jitem F;= Kinger Taps] item FM= Rapid 9lternating Movements of Xands] item ;+= 4ody 4radykinesia and Xypokinesiaz= and dyskinesia scoreV Means of the three specialists’ assessments per time point on these scales were used in subsequent analysisV
Smartphone-based data collection
On each test occasion= the subjects performed upper limb motor tests jtapping and spiral drawingsz= using a smartphone jKigure +z [B]V The subjects were instructed to perform the tests using an ergonomic pen stylus with the device placed on a table and to be seated in a chairV 6uring tapping tests= the subjects were asked to alternately tap two fields jas shown in the screen of the devicez as fast and accurate as possible= using first right hand and then left handV 8ach tapping test lasted for F’ secondsV 6uring spiral tests= the subjects were instructed to trace a pre‘drawn 9rchimedes spiral as fast and accurate as possible= using the dominant handV The spiral test was repeated three times per test occasionV The smartphone recorded both position and time‘stamps jin millisecondsz of the pen tipV
Data processing and analysis
The raw tapping and spiral data were processed with time series analysis methods= including both time‘ and frequency‘domains methodsV Nineteen and FF spatiotemporal features were extracted from spiral and tapping data= respectivelyV Keatures were calculated to represent various kinematic quantities during the motor tests such as acceleration= speed= time delay= and distanceV The features from both tapping and spiral data were used in a Principal :omponent 9nalysis and B principal components jP:sz were retained= which in turn were used as inputs to a Support Vector Machines jSVMz to be mapped to mean clinical ratingsV The analysis were performed with a stratified +’‘fold cross‘validationV Test‘retest reliability of the spiral tests were assessed after calculating correlations between the first P:s for the three spiral tests and then calculating the mean of all possible correlationsV
References
+V Martinez‘Martin P= Ii‘Nagel 9= Iracia QM=et alV Unified Parkinson’s 6isease Rating Scale characteristics and structureV The cooperative Multicentric IroupV Mov 6isordV +[[A][7BH‘q;V
FV Richards M= Marder W= :ote Q= et alV Ynterrater reliability of the Unified Parkinson’s 6isease Rating Scale motor examinationV Mov 6isordV +[[A][7q[‘[+V
;V Nyholm6= Nilsson Remahl 9Y= 6izdar N= et alV 6uodenal levodopa infusion monotherapy vs oral polypharmacy in advanced Parkinson diseaseV NeurologyV F’’M]HAjFz7F+H‘FF;V
AV Nyholm 6= :onstantinescu R= Xolmberg 4= 6izdar N= 9shmark XV :omparison of apomorphine and levodopa infusion in four patients with Parkinson’s disease with symptom fluctuationsV 9cta Neurol ScandV F’’[]++[jMz7;AM‘;AqV
MV Sampaio := Ioets := Schrag 9= et alV Rating Scales in Parkinson’s diseaseV OXKOR6 UNYV8RSYTY PR8SSV F’++7+’H‘+’BV
HV 8spay 9J= Iiuffrida JP= :hen R= et alV 6ifferential response of speed= amplitude= and rrhythm to dopaminergic medications in Parkinson’s diseaseV Mov 6isord= volV FH= ppV FM’A‘qV
BV Westin J= 6ougherty M= Nyholm 6= et alV 9 home environment test battery for status assessment in patients with advanced Parkinson’s diseaseV :omp Meth Programs 4iomed= volV [q= ppV FB‘;MV
