Harmonicity of the Movement as a Measure of Apraxic Behaviour

in Stroke Survivors

Marta Bieńkiewicz

, Philipp Gulde

Georg Goldenberg

and Joachim Hermsdörfer

Technische Universität München, Lehrstuhl für Bewegungswissenschaft,

Uptown München-Campus D Georg-Brauchle-Ring 60/62 D-80992 München, Germany

Städtisches Klinikum München, Klinik für Neuropsychologie, Englschalkinger Straße 77, 81-925 München, Germany

Keywords: Apraxia, Smoothness of Movement, Harmonicity, Kinematic Patterns, Stroke Rehabiliation.

Abstract: Due to the brain damage caused by stroke, apraxic patients suffer from tool use impairment, and sequencing

actions during daily tasks (ADL). Patients fail to use tools in a purposeful manner, often adopting an

inappropriate speed of the movement and a disrupted movement path (Laimgruber et al., 2005). The core of

this symptom lies in the compromised ability to access the appropriate motor program relevant to the task

goal (Hermsdörfer et al., 2006). Although many studies have explored kinematic and spatial features of

apraxia both in object and non-object related motor tasks, there is a niche in the research to provide a

spatiotemporal biomarker for this behaviour. We propose a novel approach based on dynamical systems

framework (Bootsma et al., 2004), looking into the temporal and spatial components of movements.

Preliminary data shows that this measure has a potential to encapsulate the disrupted motor behaviour in

those patients. We put forward a circular-fit based model to quantify deviations from the regular movement

pattern. The application of this study is to create a measure of motor behaviour to be implemented in the

autonomous assistance system (CogWatch) that could facilitate performance of ADL both in the clinical and

home-based setting.

1 INTRODUCTION

The cerebrovascular accident (CVA), whether it is

caused by bleeding or ischemia, causes a permanent

damage of brain tissue. Stroke survivors suffer from

a range of disruptions in motor circuitry such as

spasticity or loss of control over limb

(paresis/plegia). In addition, stroke can cause

sensory deficits as well as language comprehension

and production problems. The main focus of this

paper is apraxia disorder, which describes a

compromised ability of CVA patients to use objects

in an accurate, goal directed manner and in turn,

carry out ADL (Goldenberg et al., 1996). In this

study, we propose a novel quantitative approach for

capturing subtle differences in motor control on the

spatiotemporal dimension between patient group and

healthy elderly.

2 BACKGROUND

The CogWatch (www.cogwatch.eu) project is

designed to create an autonomous assistance system

to aid ADL independence in stroke cohort. The

primary scope of the project is addressing patients

who suffer from impaired ability to use everyday

tools, due to left brain damage (Bieńkiewicz et al.,

2013; Hermsdörfer et al., 2013). That means

inability to access previously mastered knowledge

about action execution, despite a preserved ability to

integrate sensory information from the environment

and execute smooth movement in a goal directed

manner (De Renzi et al., 1982).

The incidence of persistent signs of apraxia in

the population of CVA patients is estimated to be

approximately 24% of all stroke survivors

(Bickerton et al., 2012). The difficulty with the use

of tools is a source of frustration for patients, as it

directly increases the need for the help from

caregivers during ADL. This loss of independence

compounds the problems associated with CVA and

makes the consequences of apraxia more debilitating

(Hanna-Plady et al., 2003). One of the on-going

strands of the project is to identify spatiotemporal

patterns emerging during the production of ADL in

295

Bie

nkiewicz M., Gulde P., Goldenberg G. and Hermsdörfer J..

Harmonicity of the Movement as a Measure of Apraxic Behaviour in Stroke Survivors.

DOI: 10.5220/0004913802950300

In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS-2014), pages 295-300

ISBN: 978-989-758-011-6

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

this group of patients for the purpose of monitoring

online task performance, progress of recovery and

feed in automatized action recognition algorithms

(Hughes et al., 2013).

2.1 Apraxia Non-kinematic Features

The most widely accepted definition of apraxia

describes it as neurological sign of brain damage,

behaviourally observed as the inability to perform

skilled, well-learned motor acts (rothi et al., 1997).

As previously mentioned, this deficit cannot be

however explained by shortfall of motor or sensory

brain functions caused by stroke (figure 1). For

example, features of apraxia are independent from

the loss of motor function of the limb (paresis or

spasticity) or partial loss of visual field (hemianopia)

or compromised visual attention (visual neglect)

(petreska et al. 2007; goldenberg et al., 2007).

Figure 1: Illustration of the conceptual underpinnings of

the apraxic behaviour. Despite preserved ability to execute

movement and sensory system being functional, patients

have difficulty accessing the motor concepts relevant to

the action goal.

For example, problems with the object use are

present when the task is performed with the non-

affected limb (in the case of right handed

participants with left brain damage, problems with

motor features are present when the task is

performed with the left hand). However, the

problems with daily activities can be enhanced by

these deficits, but are regarded as separate symptoms

from the compromised functionality of motor

schemas. Detailed descriptions of apraxia refer to

three subcategories of symptoms affecting both

object related and non-object related performance

(Petreska et al., 2007; De Renzi et al., 1982, Jason et

al., 1983). This classification refers to the transitive

(object manipulation actions, e.g. using a hammer to

put a nail into wooden board) or non-transitive (such

as gestures, imitation and pantomime) (Goldenberg

& Hagmann, 1997). The non-transitive movements

involve gesture production and recognition for the

meaningful gestures (such as waving goodbye) and

non-meaningful ones (such as copying finger or

hand postures) (Goldenberg et al., 1996). These two

different subtypes of apraxia are usually described in

the body of research as separate – conceptual

apraxia and ideomotor apraxia respectively

(Goldenberg, 2003). The problems with smooth

performance of the task are referred in the literature

as a third category, which is limb apraxia. Limb

apraxia is defined as disruption of kinematic pattern

of the movement, with preserved gesture and tool

knowledge (Petreska et al., 2007). Those subtypes

however, although differentiated as separate

symptoms of apraxic behaviour, often coincide. In

addition to apraxia Action Disorganisation

Syndrome is distinguished in many other

neurological disorders apart from stroke and

regarded as difficulty with sequencing of the motor

acts (Cooper et al., 2005). That means performing

the action in an efficient and organised manner,

despite preserved tool knowledge. The distinction

between ADS and apraxia is however still widely

discussed in the body of literature.

Due to apraxia and ADS, patients are prone to

conceptual, spatial and temporal errors during daily

activities that can lead to potential health and safety

issues (e.g., grasping the knife by the sharp end,

pouring boiling water onto the kitchen desktop).

Common errors include problems with sequencing

in multistep actions (e.g., action or ingredient

addition, omission, anticipation and perseveration

errors) along with conceptual errors (e.g., misuse of

objects, object substitution, hesitation, toying and

mislocation) (Petreska et al., 2007). The cognitive

aspect of apraxia (i.e., the loss of knowledge how

the action is performed) is often accompanied by

changes in the kinematic pattern of the movement in

the unimpaired hand. During the pantomime and

gesture production, patients show irregularities

usually in the direction of the movement, amplitude,

speed and spatial position. Therefore, pantomime

performance is one of the hallmarks in the

neuropsychological examination of patients, due to

its high sensitivity. Some patients might not exhibit

a difficulty with the tool use, but fail to pantomime

the performance. The plausible explanation for this

phenomenon is that the priopioceptive information

from grasping the tool provides additional

sensorimotor input, which reinforces the selection of

the appropriate motor schema (Hermsdörfer et al.,

2006). The kinematic characteristics of apraxia

syndrome will be further discussed in the following

section.

2.2 Apraxia Spatio-temporal Features

There are several studies looking into the kinematic

BIOSIGNALS2014-InternationalConferenceonBio-inspiredSystemsandSignalProcessing

296

hallmarks of apraxic tool use. Unlike the studies

using video-based approach in assessment of apraxia

(Schwartz et al., 1995), or neuropsychological

batteries in assessment of ADL in CVA patients

(Vanbellingen et al., 2011; Graessel et al., 2009),

research using motion capture recordings allows one

to measure the subtle differences in the motor

control. In the seminal study by Laimgruber et al.

(2005), several variables were distinguished as

sensitive measures of the differences between

spatiotemporal features of task performance between

the CVA patients and elderly controls in a

pantomime task of taking a sip of water from a glass.

Those variables were: movement time, peak

velocity, deceleration phase and grip aperture.

Deficits in the speed of the movement were also

shown in other tasks such as the pantomime of

sawing (Hermsdörfer et al., 2006) and pantomime

and use of a hammer (Hermsdörfer et al., 2012). In a

scooping motion task involving CVA patients with

left brain damage, another study has reported

deficits in the amplitude of the movement and

reduced hand roll (Hermsdörfer et al., 2012). In

addition Clark et al. (1994) have demonstrated

imprecise plane of motion and trajectory shape in

the pantomime and tool use of a knife when slicing a

piece of bread, along with the impaired coupling in

the hand velocity and trajectory shape. This was also

shown by Poizner et al. (1995) in the same task

scenario, which highlights the impaired joint

coordination in a slicing movement. The disruptions

in the kinematic features of the movement are linked

to its more observable features such as perplexity,

indicating a difficulty with accessing the appropriate

motor plan (Hermsdörfer et al., 2006).

Other studies looking at goal-directed movement

without tool use, such as pointing task, have

reported impaired reaction times, acceleration

deficits and prolonged movement times in the task

performance by apraxic individuals (Fisk &

Goodale, 1988; Hermsdörfer et al., 1999;

Hermsdörfer et al., 2003; Haaland & Harrington,

1994). In grasping movements, impaired prehension

and awkward hand rotation were noted as spatial

features of kinematic impairments in patients

(Hermsdörfer et al., 1999; Tretriluxana et al,. 2009).

2.3 Smoothness of Movement

Differences in the movement organisation in terms

of spatiotemporal characteristics are usually limited

to the presented above approaches, taking into

consideration velocity and acceleration profiles,

movement times and movement path. In this study,

we focus on the movements that are naturally

cyclical in both spatial and temporal dimensions

(sawing, hammering and circular toothbrushing). We

have chosen this particular task, due to the plethora

of research investigating the oscillatory arm

movements in healthy adults. Bootsma, Fernandez

and Mottet (2004) have demonstrated that self-paced

cyclical arm movements performed back and forth

between two targets are normally represented by

velocity curves that resemble a repetitive sinusoid

oscillation over time. This natural harmonicity of the

movement can be represented by circular shaped

phase planes, when the velocity of the movement is

plotted against position. This can be plotted as a

semi-circle on either side of the x/y/z axis,

representing one pointing movement. The

assumption is that the more phase plane deviates

from a regular circular shape, the less harmonic the

movement. Lower harmonicity of the movement can

implies a less natural the pace of the movement or a

lesser degree of control (Bootsma et al., 2004).

2.4 Research Aims

The purpose of this research is to explore the

feasibility of new biomarkers based on the

harmonicity measure to capture the apraxic features

in the movement.

3 METHODS

3.1 Experimental Design

In the study 20 healthy elderly, age-matched with

patients will be tested. All of the healthy participants

are to be right handed, 10 will be tested with right

hand, 10 with the left hand. In the clinical group, 10

patients will be tested that suffered from first CVA,

affecting primarily areas in the left brain

hemisphere.

Control and patient groups will be tested under

two modes of execution:

A. Actual action execution

B. Pantomime with action object visible

Three daily tasks will be tested:

i) Sawing a piece of wood

ii) Hammering

iii) Toothbrushing

Each of those conditions will have two trials of

repetition and the experimental design will be

counterbalanced using Latin Squares. The practice

trial will include a task of pouring a glass of water

HarmonicityoftheMovementasaMeasureofApraxicBehaviourinStrokeSurvivors

297

from a jug (pantomime versus tool use). Motor

performance will be recorded using passive marker

setup and Qualisys Motion Capture system.

Pantomime and tool use in addition will be assessed

using the Goldenberg & Hagmann (1998) 2 point

scale. The following kinematic variables will be

analysed alongside: movement time, peak velocity,

movement path, frequency of the movement,

number of acceleration zerocrossings (jerks),

deceleration phase, grip aperture and orientation.

Number of errors committed and kinematic features

of the movement will be compared across conditions

for each patient and groups between patients and

age-matched controls.

3.2 Phase Portraits and Circular Fit

Matlab script (Mathworks, 2012) was developed to

provide the derivates of spatiotemporal positional

data and create phase plot data. The algorithm was

based on the Bootsma et al. (2004) study looking

into the harmonicity of aiming movement. We have

adopted the approach proposed by the authors and

normalised for Aω – peak velocity of an ‘idealised’

harmonic movement at given amplitude and

movement time and for the amplitude of the

movement. This can be mathematically expressed

as:

Aω=(A* π)/(MT*2) (1)

where A denotes distance travelled and MT

movement time.

To create a mathematical fit for the phase plot

data we applied the ‘Taubin’ method of curve and

surface fitting (1991) and incorporated it into Matlab

script. This method is based on a geometric-fitting

approach and minimization of the approximate mean

square distance:







∑





































∑































(2)

Where x

and y

refer to consecutive points from the

phase plot data for each trial.

3.3 Squared Error as Candidate

Measure of Harmonicity

For the purpose of calculating the deviations from

the harmonic movement pattern, we have adopted

squared error approach as a preliminary outcome

measure. Each stroke of the movement and reversal

in a trial will be normalised according to the

procedure listed in the 3.2 and further centralised

with respect to the origin of the fit. Subsequently a

squared error will be calculated between each data

point of the velocity/position data and the closest

point demarked by radius from the fitted circle. The

median value will be extracted for each trial and

condition for each participant.

Those values will be compared across patients

and healthy elderly controls.

Figure 2: Illustration of how squared error measure is

calculated based on the normalised phase plot data and

fitted circle based on Taubin method. Black arrow depicts

the 2-D distance between the fitted circle radius point and

velocity/position data (red point). For each stroke of the

movement and reversal, velocity/position data is

centralised with respect to the origin of the fitted circular

shape. The dashed red line represents a circular fit

modelled to the positional data.

4 PRELIMINARY RESULTS

Preliminary data analysis provides an optimistic

outlook for the method. So far 9 patient data were

analysed along with the data from 16 healthy elderly

subjects (10 tested on the right hand and 6 on the left

hand). Graphical representation of the sample data

are visualised in Figure 3. We have observed

increased variability in terms of velocity/positional

data in the group of apraxic patients in comparison

to healthy elderly.

As illustrated on the Figure 3, preliminary data

shows that in patients showing features of apraxia,

we observe a disrupted pattern of harmonicity of the

movement, when represented as phase plots. In

addition, a difference in the surface fitting will be

taken as a mean difference between the centralised

movement cycle and fitted circle. In the preliminary

data inspection other measures also revealed

differences between patients and age matched

controls, such as movement frequency, amplitude

and movement path ratio (x/y/z to xyz).

BIOSIGNALS2014-InternationalConferenceonBio-inspiredSystemsandSignalProcessing

298

Figure 3: An example of normalised phase planes showing

relation between the velocity of the movement against the

position on the main movement axis (respectively z, y, x)

from preliminary data analysis. Each row illustrates

performance in the pantomime mode on the tasks:

hammering, sawing and tooth brushing. Left panel depicts

movement organisation of the apraxic individual (CVA in

August 2012, 12 months prior to data collection). In the

right side panel, control data for age and sex matched

volunteer.

5 CONCLUSIONS

The work on this line of CogWatch project is in

progress and requires detailed analysis to identify

differences between selected CVA patients that

show apraxic behaviour and healthy elderly

performance. On the basis of the data collected in

this study, a new measure might emerge that will

feed into the development of the rehabilitation

approach for those patients. This parsimonious

approach to kinematic analysis might provide a

novel insight into understanding the kinematic

consequences of apraxia. The long term goal is to

use harmonicity of movement as the biomarker for

non-motor execution related disruptions in the

performance of ADL that require cyclical

movements. Authors are not aware of any kinematic

biomarkers specific to apraxia being identified in a

body of research. In this paper, we have argued that

using the harmonicity measure might allow one to

encapsulate many features of apraxic behaviour on

the spatiotemporal dimension such as: movement

amplitude, movement path, frequency of the

movement, speed and acceleration profiles. The

purpose of application of harmonicity measure in

CogWatch is to compare how different

interventions, based on supplementary sensory

information, influence motor behaviour in patients

with apraxia and monitor the progress of recovery.

ACKNOWLEDGEMENTS

This work was funded by the EU STREP Project

CogWatch (FP7-ICT- 288912). Authors would like

thank to the Klinikum Bogenhausen patients and

staff members for participation in the research and

student assistants: Johannes Pflüger, Andrea

Schlegel, Anna Voitl and Saskia Steinl for the help

with running the experimental sessions.

REFERENCES

Bickerton, W., Riddoch, J., Samson, D., Balani,A., Mistry,

B., & Humphreys, G., (2012). Systematic assessment

of apraxia and functional predictions from the

Birmingham Cognitive Screen, Journal of neurology,

neurosurgery & psychiatry with practical neurology,

83 (5), 513-52.

Bieńkiewicz, M., Goldenberg, G., Cogollor, J., Ferre, M.,

Hughes, C., Hermsdörfer, J. (2013).Use of Biological

Motion based Cues and Ecological Sounds in the

Neurorehabilitation of Apraxia. HEALTHINF 2013.

Bootsma, R., Fernandez, L., & Mottet, D. (2004). Behind

Fitts’ law: kinematic patterns in goal-directed

movements, International Journal of Human-

Computer Studies, 61 (6), 811-821.

Clark, M.A., Merians, A.S., Kothari, A., Poizner, H.,

Macauley, B., Rothi, L.J.G. and Heilman, K.M. (1994)

Spatial planning deﬁcits in limb apraxia. Brain, 117:

1093–1106.

Cooper, R. P.; Schwartz, M.; Yule, P. & Shallice, T.

(2005). The simulation of action disorganisation in

complex activities of daily living. Cognitive

Neuropsychology 22 (8) 959-1004.

De Renzi, E., Faglioni, P., & Sorgato, P. (1982) Modality

speciﬁc and supramodal mechanisms of apraxia.

Fisk, J., & Goodale, M. A. (1998), The effects of

unilateral brain damage on visually guided reaching:

hemispheric differences in the nature of the

deficit. Experimental Brain Research,72, 425–35.

Goldenberg, G. (2003) Apraxia and beyond: life and work

of Hugo Liepmann. Cortex, 39(3): 509–524.

Goldenberg, G., Hermsdörfer, J., Glindemann, R., Rorden,

C., & Karnath, H. O. (2007). Pantomime of tool use

depends on integrity of left inferior frontal cortex.

Cerebral Cortex, 17, 2769-2776.

HAMMERING

SAWING

TOOTHBRUSHING

CVA_MALE (AGE 54) CONTROL_MALE (AGE 55)

HarmonicityoftheMovementasaMeasureofApraxicBehaviourinStrokeSurvivors

299

Goldenberg, G. and Hagmann, S. (1997) The meaning of

meaningless gestures: a study of visuo-imitative

apraxia. Neuropsychologia, 35(3): 333–341.

Goldenberg, G., & Hagmann, S. (1998) Therapy of

activities of daily living in patients with apraxia.

Neuropsychological Rehabilitation, 8(2), 123–41.

Goldenberg, G., Hermsdörfer, J., & Spatt, J. (1996).

Ideomotor apraxia and cerebral dominance for motor

control. Cognitive Brain Research, 3, 95-100.

Graessel, E., Viegas, R., Stemmer, R., Küchly, B.,

Kornhuber, J. & Donath, C., (2009). The Erlangen

Test of Activities of Daily Living: first results on

reliability and validity of a short performance test to

measure fundamental activities of daily living in

dementia patients. International Psychogeriatrics, 21,

103-112.

Haaland, K.Y., Harrington, D.L., (1994). Limb-

sequencing deficits after left but not right hemisphere

damage. Brain and Cognition, 24, 104-22.

Hanna-Pladdy, B., Heilman, K.M., & Foundas, A.L.

(2003) Ecological implications of ideomotor apraxia:

evidence from physical activities of daily living.

Neurology, 60, 487–490.

Hermsdörfer, J., Mai, N., Spatt, J., Marquardt, C.,

Veltkamp, R., & Goldenberg, G. (1996) Kinematic

analysis of movement imitation in apraxia, Brain, 119,

1575-1586.

Hermsdörfer, J., Ullrich, S., Marquardt, C., Goldenberg,

G., & Mai, N. (1999). Prehension with the ipsilesional

hand after unilateral brain damage. Cortex, 35, 139-

161.

Hermsdörfer, J., Blankenfeld, H., & Goldenberg, G.

(2003). The dependence of ipsilesional aiming deficits

on task demands, lesioned hemisphere, and apraxia.

Neuropsychologia, 41, 1628-1643.

Hermsdörfer, J., Hentze, S., & Goldenberg, G. (2006)

Spatial and kinematic features of apraxic movement

depend on the mode of execution. Neuropsychologia,

44, 1642–1652.

Hermsdörfer, J., Li, Y., Randerath, J., Roby-Brami, A., &

Goldenberg, G. (2012). Tool use kinematics across

different modes of execution. Implications for action

representation and apraxia. Cortex, in press.

Hermsdörfer, J., Bieńkiewicz, M., Cogollor, J., Russel,

M., Baptiste, E., Parekh, M., Wing, A., Ferre, M.,

Hughes, C., (2013). CogWatch - Automated

Assistance and Rehabilitation of Stroke-Induced

Action Disorders in the Home Environment. HCI (17):

343-350 .

Hughes, C., Baber, C., Bieńkiewicz, M., Hermsdörfer, J.

(2013). Application of Human Error Identification

(HEI) Techniques to Cognitive Rehabilitation in

Stroke Patients with Limb Apraxia. HCI (8) , 463-471.

Jason, G. W. (1983). Hemispheric asymmetries in motor

function: Left hemisphere specialization for memory

but not performance. Neuropsychologia, 21 (1), 35-45.

Laimgruber, K., Goldenberg, G., & Hermsdörfer, J. (2005)

Manual and hemispheric asymmetries in the execution

of actual and pantomimed prehension,

Neuropsychologia, 43(5), 682-692.

MATLAB and Statistics Toolbox Release 2012b, The

MathWorks, Inc., Natick, Massachusetts, United

States.

Petreska, B., Adriani, M., Blanke, O. & Billard,

A. (2007) Apraxia: a review. In C. von Hofsten (Ed.).

From Action to Cognition. Progress in Brain Research.

Elsevier. Amsterdam. Vol. 164, pp. 61-83.

Poizner, H., Clark, M.A., Merians, A.S., Macauley, B.,

Rothi, L.J.G. and Heilman, K.M. (1995) Joint

,coordination deﬁcits in limb apraxia. Brain, 118: 227–

242.

Schwartz, M. F., Montgomery, M. W., Fitzpatrick-

desalme, E. J., Ochipa, C., Coslett, H. B., & Mayer, N.

H. (1995). Analysis of a disorder of everyday action.

Cognitive Neuropsychology, 12, 863-892.

Taubin, G. (1991). Estimation Of Planar Curves, Surfaces

And Nonplanar Space Curves Defined By Implicit

Equations, With Applications To Edge And Range

Image Segmentation. IEEE Trans. PAMI, 13, 1115-

1138.

Tretriluxana, J., Gordon, J., Fisher, B. E., & Winstein, C.

J. (2009). Hemisphere specific impairments in reach-

to-grasp control after stroke: effects of object size.

Neurorehabilitation and Neural Repair, 23, 679-691.

Vanbellingen, T., Kersten, B., Van de Winckel, A.,

Bellion, M., et al. (2011). A new bedside test of

gestures in stroke: the apraxia screen of TULIA

(AST). Journal of Neurology, Neurosurgery and

Psychiatry 82(4), 389-92.

BIOSIGNALS2014-InternationalConferenceonBio-inspiredSystemsandSignalProcessing

300