Matteo Pastorino - Remote daily activity of parkinson’s disease patients the akinesia assessment

Remote daily activity of Parkinson’s disease patients: the
Akinesia assessment .

Matteo Pastorino
Technical University of Madrid
WTHS_2011
Valencia, December 2011

WTHS_2011 | Valencia| December 2011 1

Parkinson’s Disease (PD)

Parkinson’s Disease (PD) is a degenerative,
progressive disorder that affects nerve cells in deep
parts of the brain called the basal ganglia and the
substantia nigra.

Nerve cells in the substantia nigra produce the
neurotransmitter dopamine and are responsible for
relaying messages that plan and control body movement.

Causes: Genetic, environmental factors …


Parkinson’s Disease (PD) - Symptoms

 Movement disorders
 Bradykinesia
 Akinesia
 Rigidity
 Tremor
 Dyskinesia
 Freezing of Gait
 Cognitive and behavioural disorders
 Dementia
 Depression
 Hallucination
 Sensory, sleep and emotional problems


Parkinson’s Disease (PD) - Akinesia
Akinesia (α a, "absence", κίνησις kinēsis, “movement")
represents the most promising motor progression marker of the disease.

Characteristics: defined as absence of movement. This is a condition in which any automatic
movement or action, including gestures, blinking or swallowing actions are limited and their
frequency decreases, although the elemental motor functions are maintained and can be
performed voluntarily. Various aspects appear to contribute to the self-initiation of movements:

Causes: reduced dopaminergic input to the striatum.
Such changes also cause bradykinesia, rigidity, tremor and postural instability, although the underlying
mechanisms leading to these symptoms are still not understood.

Treatment: dopamine precursor levodopa is the most efficient treatment for the improvement of
Parkinson´s disease signs and symptoms.
However, abnormal involuntary movements (dyskinesia) are motor fluctuations that occur in the majority of
PD’s patients undergoing this treatment.


Parkinson’s Disease (PD) – ONOFF

Specific Chronic Neurodegenerative Diseases
 Progressive loss of motion ability (due to muscle weakening)
 Appearance of new motion symptoms (new muscles affected)
 Inability to move (at later stages)

Parkinson’s Disease:
 Progression is restricted with treatment
 Daily motion status is fluctuating due to
treatment
 Dyskinesia


Monitoring & Assessment TODAY

Every 4-6 months or
As instructed

Patient visits Clinic

Treatment
Adjustment

Clinician tries to Clinician PERFORMs Made from visits
Reconstruct the patient status UPDRS or other tests to identify observations & subjective
Throughout day and night Current patient status assessment


Parkinson Disease & p-Health solutions

24h Every day
monitoring
Patient at home

ALERTS!
Treatment
Adjustment
Test Devices
Based on objective
observation

Other Info 24 h objective status assessment

Immediate Response


PD & p-Health solutions: OBJECTIVES

Short Term
• 24h objective assessment of patient status
• Detection of dosage wearing-off
• Adjustment of therapy according to personal characteristics and reaction
• Medication schedule/dosage
• Food Intake
• Detection of changes in patient reaction to therapy
• All patient info at-a-glance and detailed info one-click away
Long Term
• Objective therapy assessment
• Analysis of symptoms progression in time
• Recognition of changes in therapeutic response


PERFORM Project
 This work is part of the PERFORM project, partially funded by the European Commission

under the 7th framework programme www.perform-project.eu

 Consortium:


PERFORM Architecture
24h
monitoring

Other Info

Test Devices Monitor Detect & Quantify
Patient Symptoms & Gait Build
New Treatment Patient Specific
Regime
disease profile
Suggest Assess
Treatment Changes Disease Progress


PERFORM Architecture
Clinician

Administrator
Professional GUI
Central Hospital Unit

Local Base Unit Central Hospital Unit Login Manager User Database Account Manager
 Exploits the recorded patient information in order to build a
Alert Manager
 Processes Patient GUI
the patient signals patient symptom profile.
acquired; Patient List Index of Processed Info
Local Base Unit the targeted patient
 Detects For each symptom
symptoms (e.g.: tremor,  produces a patient profile which describes the patient’s
Clinical Decision Support Systems

Central Unit Information Handler
levodopa induced dyskinesia,
Device Controller Scheduler common symptom features.

Central Unit Communicator
Akinesia,..).  compared with thePatient Modelling
patient symptom profile. Patient Management
Information Handler

For Processorsymptom a dedicated
Test each Gait Early Wearing Off
submodule : Action Tremor If significant differences are found, it might be due to two reasons:
 Processes the relevant signals;  temporarily patient behaviour abnormalityMedication Change
On – Off
 Detects the symptomFreezing of Gait
episode; or Stability-Worsening
LID
 the symptom episode:  a change in the patient profile.
Gait
 according to the Unified Tremor
Parkinson’s Disease
Tremor
Bradykinesia

Monitoring
Wearable
Sensors
Logger Rating Scale
or
LID Frequent falls

System  Other features Activity as
such Checks whether a substantial number of similar situations are
User-
Hardware
Interface
duration, frequency, identified for the last time period for the specific patient and if
Repository
erergy and Bradykinesia
amplitude that occurs, it creates an alert.
Fall Detector Interoperability Manager
might also be Akinesia
provided
Alert
for further clinician
PERFORM
Manager review and
Communicator system Repository
evaluation.
External Resources


PERFORM Monitoring System

Gyroscope/
Accelerometer

Accelerometer
Day Monitoring Accelerometer
wearable

Accelerometer / Control
Unit

Accelerometer Accelerometer


PERFORM: Patient Interface

Interface easy to use
Look and feel of the phone dialling pad
Drag and Drop functions
Used to declare subjective estimation of Patient status
Used to receive instructions on life-style interventions
(medication/food intake)

PERFORM Technological Innovation

 Continuous Patient Monitoring & Assessment

 Detection of all symptoms using a single and low cost
sensor setting

 Early recognition of disease progression and patient
reaction changes

 Assistance in patient management with expert
knowledge based systems

 Prognosis of disease evolution according to patient
characteristics


PERFORM: Pilots description

Phase Characteristics # of patients Objective Pilot Sites

Data Collection 8 healthy + Design the
1 with SHIMMERS 8 PD
Madrid and Pamplona(Spain)
algorithm

Data Collection in Design the Pamplona(Spain) and Ioannina
2 a supervised 20 algorithm and train (Greece)
environment the classifier

Data Collected in Pamplona(Spain) and Ioannina
3 a unsupervised 24 Data Collection (Greece)
environment
Data Collected in
Test and Validation Modena (Italy)
4 a unsupervised 22
of algorithms
environment


Akinesia Algorithm and design
Different modules were created in order to detect and quantify different symptoms

AKINESIA module assesses the amount of movement of the patient in space for any
given period of time.


Akinesia: Algorithm design

Output

• Pre-processing:
• Resultant Computation: eliminate position dependence
of the sensors given by:
•
x2 + y2 + z2
• Filtering: Akinesia is related with the slowness of the
movement ,therefore we are interested in the low
frequencies of the signal:
• Band-Pass IIR Butterworth 4th order filter [1÷3] Hz.


Akinesia: Algorithm design

Output

The signal is split in 5 minutes length epochs to evaluate a considerable portion of signal. There is 50%
overlapping in epochs to study the whole signal.
For each epoch computes Total Amount of Energy for each working sensor.

OUTPUT:
The resulting energies of each sensor are combined by using a
weighted sum in order to take into account all the possible
combination of sensors.

The module is able to recognize automatically the sensor’s setting


Akinesia: ON-OFF Evaluation
→ clear relationship between ON-OFF phases and the akinesia levels.

→ Strong correlation between the lack of movement and OFF status.

→Using the akinesia is possible to discriminate ON and OFF periods in PD patients.

ON OFF ON


Akinesia: Results
For the analysis of the results, two different scenarios are considered.

Akinesia – NO WALKING periods Akinesia – WALKING periods
 mean value of the computed akinesia during the  mean value of the computed akinesia during
periods when the patient is not walking. walking periods.

The global evaluation of both scenarios demonstrates that it is possible to discriminate ON and OFF periods computing
the lack of movement combining the information provided by different modules of the PERFORM system, in this case
the activity recognizer and the Akinesia module.

Conclusions

Useful and Objective tool for the analysis of the akinesia in
PD’s patients.
Suitable for clinical practice
Support health professionals in the diagnosis and follow-
up of PD patients
PD patients’ quality of life improvement
Discriminating parameter for the ON – OFF condition

DATA:
 Recording of one patient during 4h and are focused only in the akinesia results as
discriminating parameter for the ON – OFF

FUTURE WORK:
 More exhaustive analysis using all the recordings collected during the pilot phases;
 Combining the results of all PERFORM classifier outputs
 Create a complete profile of patients


Thank you!

Matteo Pastorino
Universidad Politécnica de Madrid
Life Supporting Technologies
mpastorino@lst.tfo.upm.es
Skype id: matteo_pasto


Matteo Pastorino - Remote daily activity of parkinson’s disease patients the akinesia assessment

Recommended

Recommended

More Related Content

Viewers also liked

Viewers also liked (20)

Similar to Matteo Pastorino - Remote daily activity of parkinson’s disease patients the akinesia assessment

Similar to Matteo Pastorino - Remote daily activity of parkinson’s disease patients the akinesia assessment (20)

More from WTHS

More from WTHS (20)

Recently uploaded

Recently uploaded (20)

Matteo Pastorino - Remote daily activity of parkinson’s disease patients the akinesia assessment