This paper reviews related work and state-of-the-ar
t publications for recognizing motor symptoms of
Parkinson's Disease (PD). It presents research effo
rts that were undertaken to inform on how well
traditional machine learning algorithms can handle
this task. In particular, four PD related motor
symptoms are highlighted (i.e. tremor, bradykinesia
, freezing of gait and dyskinesia) and their detail
s
summarized. Thus the primary objective of this rese
arch is to provide a literary foundation for develo
pment
and improvement of algorithms for detecting PD rela
ted motor symptoms.
Wireless healthcare: the next generationJeffrey Funk
These slides use concepts from my (Jeff Funk) course entitled analyzing hi-tech opportunities to analyze how wireless healthcare are becoming economic feasible. Improvements in microprocessor and transceiver ICs, MEMS, photo-sensors, and other electronic components are making wireless healthcare economically feasible. These slides show how improvements in these components are making capsule endoscopy, smart drug delivery, and digital pills economically feasible. Capsule endoscopy involves sending a small device through the body, particularly the digestive system, to take images. Further improvements in electronic components are needed to further reduce the size of these devices. Drugs can be dispensed through smart pills at programmed times or can be triggered by sensors that detect the correct location. Digital pills send signals to mobile phones or other devices when the pills have been taken. The slides conclude by discussing the role of mobile phones in increasing the number of wireless healthcare applications.
Wireless healthcare: the next generationJeffrey Funk
These slides use concepts from my (Jeff Funk) course entitled analyzing hi-tech opportunities to analyze how wireless healthcare are becoming economic feasible. Improvements in microprocessor and transceiver ICs, MEMS, photo-sensors, and other electronic components are making wireless healthcare economically feasible. These slides show how improvements in these components are making capsule endoscopy, smart drug delivery, and digital pills economically feasible. Capsule endoscopy involves sending a small device through the body, particularly the digestive system, to take images. Further improvements in electronic components are needed to further reduce the size of these devices. Drugs can be dispensed through smart pills at programmed times or can be triggered by sensors that detect the correct location. Digital pills send signals to mobile phones or other devices when the pills have been taken. The slides conclude by discussing the role of mobile phones in increasing the number of wireless healthcare applications.
human activity recognization using machine learning with data analysisVenkat Projects
Human activity recognition, or HAR for short, is a broad field of study concerned with identifying the specific movement or action of a person based on sensor data.
The sensor data may be remotely recorded, such as video, radar, or other wireless methods. It contains data generated from accelerometer, gyroscope and other sensors of Smart phone to train supervised predictive models using machine learning techniques like SVM , Random forest and decision tree to generate a model. Which can be used to predict the kind of movement being carried out by the person, which is divided into six categories walking, walking upstairs, walking down-stairs, sitting, standing and laying?
MLM and SVM achieved accuracy of more than 99.2% in the original data set and 98.1% using new feature selection method. Results show that the proposed feature selection approach is a promising alternative to activity recognition on smart phones.
Machine Learning for Disease PredictionMustafa Oğuz
A great application field of machine learning is predicting diseases. This presentation introduces what is preventable diseases and deaths. Then examines three diverse papers to explain what has been done in the field and how the technology works. Finishes with future possibilities and enablers of the disease prediction technology.
Prospects of Deep Learning in Medical ImagingGodswll Egegwu
A SEMINAR Presentation on the Prospects of Deep Learning in Medical Imaging Presented to the Department of Computer Science, Nasarawa State Polytechnic, Lafia.
BY:
EGEGWU, GODSWILL
08166643792
http://facebook.com/godswill.egegwu
http://egegwugodswill.name.ng
Exploring temporal graph data with Python: a study on tensor decomposition o...André Panisson
Tensor decompositions have gained a steadily increasing popularity in data mining applications. Data sources from sensor networks and Internet-of-Things applications promise a wealth of interaction data that can be naturally represented as multidimensional structures such as tensors. For example, time-varying social networks collected from wearable proximity sensors can be represented as 3-way tensors. By representing this data as tensors, we can use tensor decomposition to extract community structures with their structural and temporal signatures.
The current standard framework for working with tensors, however, is Matlab. We will show how tensor decompositions can be carried out using Python, how to obtain latent components and how they can be interpreted, and what are some applications of this technique in the academy and industry. We will see a use case where a Python implementation of tensor decomposition is applied to a dataset that describes social interactions of people, collected using the SocioPatterns platform. This platform was deployed in different settings such as conferences, schools and hospitals, in order to support mathematical modelling and simulation of airborne infectious diseases. Tensor decomposition has been used in these scenarios to solve different types of problems: it can be used for data cleaning, where time-varying graph anomalies can be identified and removed from data; it can also be used to assess the impact of latent components in the spreading of a disease, and to devise intervention strategies that are able to reduce the number of infection cases in a school or hospital. These are just a few examples that show the potential of this technique in data mining and machine learning applications.
Data Mining: Concepts and Techniques (3rd ed.)- Chapter 3 preprocessingSalah Amean
the chapter contains :
Data Preprocessing: An Overview,
Data Quality,
Major Tasks in Data Preprocessing,
Data Cleaning,
Data Integration,
Data Reduction,
Data Transformation and Data Discretization,
Summary.
Use Machine learning to solve classification problems through building binary and multi-class classifiers.
Does your company face business-critical decisions that rely on dynamic transactional data? If you answered “yes,” you need to attend this free event featuring Microsoft analytics tools. We’ll focus on Azure Machine Learning capabilities and explore the following topics: - Introduction of two class classification problems.
- Classification Algorithms (Two Class Classification)
- Available algorithms in Azure ML.
- Real business problems that is solved using two class classification.
When it comes to AI use for prediction, diagnosis and treatment of medical conditions, reality is often replaced with a hype. Limitations should be known. A review of AI failures and challenges in healthcare showing why it is not likely for algorithms to replace physicians in the nearest future.
Performance analysis of automated brain tumor detection from MR imaging and CT scan using basic image processing techniques based on various hard and soft computing has been performed in our work. Moreover, we applied six traditional classifiers to detect brain tumor in the images. Then we applied CNN for brain tumor detection to include deep learning method in our work. We compared the result of the traditional one having the best accuracy (SVM) with the result of CNN. Furthermore, our work presents a generic method of tumor detection and extraction of its various features.
We always prefer an unobtrusive continuous health monitoring system in the home for the purpose of assessing early health changes. Identification followed by assessment of the health issues at early stages of health disorder provides a window of opportunity for curing the issues before they become lethal. This presentation discusses various Artificial Intelligence techniques which can be used in this regard.
it is a smart wheelchair which uses voice and bluetooth commands . Also consists of temperature and heartbeat sensors for continuous monitoring by the doctor.
human activity recognization using machine learning with data analysisVenkat Projects
Human activity recognition, or HAR for short, is a broad field of study concerned with identifying the specific movement or action of a person based on sensor data.
The sensor data may be remotely recorded, such as video, radar, or other wireless methods. It contains data generated from accelerometer, gyroscope and other sensors of Smart phone to train supervised predictive models using machine learning techniques like SVM , Random forest and decision tree to generate a model. Which can be used to predict the kind of movement being carried out by the person, which is divided into six categories walking, walking upstairs, walking down-stairs, sitting, standing and laying?
MLM and SVM achieved accuracy of more than 99.2% in the original data set and 98.1% using new feature selection method. Results show that the proposed feature selection approach is a promising alternative to activity recognition on smart phones.
Machine Learning for Disease PredictionMustafa Oğuz
A great application field of machine learning is predicting diseases. This presentation introduces what is preventable diseases and deaths. Then examines three diverse papers to explain what has been done in the field and how the technology works. Finishes with future possibilities and enablers of the disease prediction technology.
Prospects of Deep Learning in Medical ImagingGodswll Egegwu
A SEMINAR Presentation on the Prospects of Deep Learning in Medical Imaging Presented to the Department of Computer Science, Nasarawa State Polytechnic, Lafia.
BY:
EGEGWU, GODSWILL
08166643792
http://facebook.com/godswill.egegwu
http://egegwugodswill.name.ng
Exploring temporal graph data with Python: a study on tensor decomposition o...André Panisson
Tensor decompositions have gained a steadily increasing popularity in data mining applications. Data sources from sensor networks and Internet-of-Things applications promise a wealth of interaction data that can be naturally represented as multidimensional structures such as tensors. For example, time-varying social networks collected from wearable proximity sensors can be represented as 3-way tensors. By representing this data as tensors, we can use tensor decomposition to extract community structures with their structural and temporal signatures.
The current standard framework for working with tensors, however, is Matlab. We will show how tensor decompositions can be carried out using Python, how to obtain latent components and how they can be interpreted, and what are some applications of this technique in the academy and industry. We will see a use case where a Python implementation of tensor decomposition is applied to a dataset that describes social interactions of people, collected using the SocioPatterns platform. This platform was deployed in different settings such as conferences, schools and hospitals, in order to support mathematical modelling and simulation of airborne infectious diseases. Tensor decomposition has been used in these scenarios to solve different types of problems: it can be used for data cleaning, where time-varying graph anomalies can be identified and removed from data; it can also be used to assess the impact of latent components in the spreading of a disease, and to devise intervention strategies that are able to reduce the number of infection cases in a school or hospital. These are just a few examples that show the potential of this technique in data mining and machine learning applications.
Data Mining: Concepts and Techniques (3rd ed.)- Chapter 3 preprocessingSalah Amean
the chapter contains :
Data Preprocessing: An Overview,
Data Quality,
Major Tasks in Data Preprocessing,
Data Cleaning,
Data Integration,
Data Reduction,
Data Transformation and Data Discretization,
Summary.
Use Machine learning to solve classification problems through building binary and multi-class classifiers.
Does your company face business-critical decisions that rely on dynamic transactional data? If you answered “yes,” you need to attend this free event featuring Microsoft analytics tools. We’ll focus on Azure Machine Learning capabilities and explore the following topics: - Introduction of two class classification problems.
- Classification Algorithms (Two Class Classification)
- Available algorithms in Azure ML.
- Real business problems that is solved using two class classification.
When it comes to AI use for prediction, diagnosis and treatment of medical conditions, reality is often replaced with a hype. Limitations should be known. A review of AI failures and challenges in healthcare showing why it is not likely for algorithms to replace physicians in the nearest future.
Performance analysis of automated brain tumor detection from MR imaging and CT scan using basic image processing techniques based on various hard and soft computing has been performed in our work. Moreover, we applied six traditional classifiers to detect brain tumor in the images. Then we applied CNN for brain tumor detection to include deep learning method in our work. We compared the result of the traditional one having the best accuracy (SVM) with the result of CNN. Furthermore, our work presents a generic method of tumor detection and extraction of its various features.
We always prefer an unobtrusive continuous health monitoring system in the home for the purpose of assessing early health changes. Identification followed by assessment of the health issues at early stages of health disorder provides a window of opportunity for curing the issues before they become lethal. This presentation discusses various Artificial Intelligence techniques which can be used in this regard.
it is a smart wheelchair which uses voice and bluetooth commands . Also consists of temperature and heartbeat sensors for continuous monitoring by the doctor.
Overview of critical factors affecting medical user interfaces in intensive c...hiij
This paper provides a comprehensive overview of cri
tical factors, which affect on-screen user interfac
es of
medical devices in Intensive Care Unit (ICU). A lit
erature survey with relevant research publications
has
led to selection of thirty eight critical factors i
n ICU. The critical factors identified are categori
zed into
various groups based on three major aspects – syste
m evaluation parameters, constituents of patient
management and user interface design. Physicians’ s
urvey, in which five physicians are involved, is us
ed to
categorize the identified critical factors into rel
ated groups. In the process, fourteen critical fact
ors are
mainly selected, which affect on-screen user interf
ace design of medical devices. The applicability of
such
factors is demonstrated with the help of a case stu
dy of head-injury patient admitted in ICU. The crit
ical
factors identified are definitely useful to device
manufacturers, user interface designers, ICU
administrators and physicians for improved device d
esign, ICU resource management and patient care.
Parkinson’s: What Do We Know About the Disease and What Can Be Done About It?asclepiuspdfs
ABSTRACT
In this article, I aim to answer important questions regarding Parkinson’s disease and the associated dementia. While the
disease was identified and described over a century ago, we still have not as yet been able to ferret out its root cause,
notwithstanding the tremendous progress made in recent years. Like for many other diseases, it is believed to involve three
main causal components (inherited genetics, environmental influences, and, to a much lesser extent, lifestyle choices),
which collectively determine if someone will develop the disease. I will survey its signs, symptoms (motor and non-motor),
risks, and stages, distinguishing between the disease’s early- and late-onset. While discriminating between the disease and
its associated dementia, I will localize the latter within the broad spectrum of dementias. I will also describe what happens
to the brain as the disease takes hold and evolves. A number of medical conditions called Parkinsonisms may have one or
more of their signs and symptoms mimicking Parkinson’s. I will discuss them in some detail, including their five proposed
mechanisms (protein aggregation in Lewy bodies, disruption of autophagy, mitophagy, neuroinflammation, and breakdown
of the blood–brain barrier). I will further describe the approach to diagnosis, prediction, prevention, and prognosis. While
there is no cure and treatment for each affected person, motor symptoms are managed with several medications (Levodopa
always combined with a dopa decarboxylase inhibitor and sometimes also with a catechol-O-methyltransferase [COMT]
inhibitor, dopamine agonists, and monoamine oxidase-B [MAOB]-inhibitors) and eventually surgical therapy. Numerous
pharmaceutical agents are also available for individual non-motor symptoms (L-Dopa emulsions, non-ergot dopamine
agonists, cholinesterase inhibitors for dementia, modafinil for daytime sleepiness, and quetiapine for psychosis). Fortunately,
we can track the drug effectiveness with exosomes. Keeping in mind patients and their caregivers/partners, I will outline
available complementary therapies, palliative care, and rehabilitation, measures they can take beyond seeking standard
treatments, and supporting and advocating organizations at their disposal. Finally, I will survey promising new research
vistas in the field.
Statistical, Energy Values And Peak Analysis (SEP) Approach For Detection of ...IJMERJOURNAL
ABSTRACT: In this paper, a technique of statistical, Energy values and peak analysis (SEP) approach is used for detection of neurodegenerative diseases from the signal of force sensitive resistors. In this work within the time series Left Stride Interval, Right Stride Interval, Left Swing Interval, Right Swing Interval, Left Stance Interval, Right Stance Interval and Double support interval are obtained and apply the SEP method. In statistical analysis, energy, standard deviation, mean, variance, co-variance are calculated. Two approximations and two details of energy values are extracted from wavelet decomposition. Average peak interval and peak histogram are calculated using peak analysis. Support Vector Machine (SVM) and Random Forest are used as a classifier. Data sets which include a healthy control (HC), various types of Neuro degenerative Diseases: Parkinson’s Disease (PD), Huntington Disease (HD), Amyotrophic Lateral Sclerosis. For disease diagnostic Force Sensitive resistor signals are used for evaluation. The results show that the proposed technique can successfully detect the NDD pathologies. For NDD detection, the accuracy, the Sensitivity, the Specificity values are 97%, 97% and 97% using Random forest Classifier.
Diagnosis of some diseases in medicine via computerized experts systemijcsit
Nowadays medical application especially diagnosis of some heart diseases has been rapidly increased
because its importance and effectiveness to detect diseases and classify patients. In this research, we
present the design of an expert system that aims to provide the patient with background for suitable
diagnosis and treatment (Especially Angina Pectoris and Myocardial infarction). The proposed
methodology is composed of four stages. The first stage is receiving the symptoms from the patient. The
second stage is requesting from the patient to make some analysis and investigation to help the system to
make a correct decision in the diagnosis. The third stage is doing diagnosis of patient according to
information from patient (symptoms, analysis and investigation). The four stage is determining the name of
appropriate medication or what should be done until the patient recovers (step therapy), so this system is
able to give appropriate diagnosis and treatment for two heart diseases namely; angina pectoris and
infarction. There are several programs used for diagnosis and system analysis, such as CLIPS and
PROLOG. A medical expert system in this search made by Visual Prolog 7.3 is proposed.
Multisensory Environments and the Patient with Alzheimer’s Disease: An Eviden...CrimsonPublishersTNN
Multisensory Environments and the Patient with Alzheimer’s Disease: An Evidence-based Review by Hassan Izzeddin Sarsak in Techniques in Neurosurgery & Neurology
Effect of Transcranial Direct Current Stimulation (Tdcs) on the Consumption o...semualkaira
Migraine is a disorder that has a pulsating, unilateral character, which worsens with physical exertion and hinders the quality of life of people who suffer from it. Even though
there are several comorbidities related to the disease, pain control
is the main objective and neuromodulation can be used in patients
who have some drug intolerance or capacity for non-pharmacological management.
Effect of Transcranial Direct Current Stimulation (Tdcs) on the Consumption o...semualkaira
Migraine is a disorder that has a pulsating, unilateral character, which worsens with physical exertion and hinders the quality of life of people who suffer from it. Even though
there are several comorbidities related to the disease, pain control
is the main objective and neuromodulation can be used in patients
who have some drug intolerance or capacity for non-pharmacological management.
REVIEWpublished 24 June 2015doi 10.3389fnhum.2015.003.docxmalbert5
REVIEW
published: 24 June 2015
doi: 10.3389/fnhum.2015.00359
Pathophysiology of ADHD and
associated problems—starting points
for NF interventions?
Björn Albrecht*, Henrik Uebel-von Sandersleben, Holger Gevensleben and
Aribert Rothenberger
Department of Child and Adolescent Psychiatry, University Medical Center Göttingen, Göttingen, Germany
Edited by:
Martijn Arns,
Research Institute Brainclinics,
Netherlands
Reviewed by:
Roumen Kirov,
Institute of Neurobiology, Bulgarian
Academy of Sciences, Bulgaria
Leon Kenemans,
Utrecht University, Netherlands
*Correspondence:
Björn Albrecht,
Department of Child and Adolescent
Psychiatry, University Medical Center
Göttingen, von Siebold Straße 5,
37075 Göttingen, Germany
[email protected]
Received: 06 October 2014
Accepted: 02 June 2015
Published: 24 June 2015
Citation:
Albrecht B, Uebel-von Sandersleben
H, Gevensleben H and Rothenberger
A (2015) Pathophysiology of ADHD
and associated problems—starting
points for NF interventions?
Front. Hum. Neurosci. 9:359.
doi: 10.3389/fnhum.2015.00359
Attention deficit hyperactivity disorder (ADHD) is characterized by severe and
age-inappropriate levels of hyperactivity, impulsivity and inattention. ADHD is a
heterogeneous disorder, and the majority of patients show comorbid or associated
problems from other psychiatric disorders. Also, ADHD is associated with cognitive and
motivational problems as well as resting-state abnormalities, associated with impaired
brain activity in distinct neuronal networks. This needs to be considered in a multimodal
treatment, of which neurofeedback (NF) may be a promising component. During NF,
specific brain activity is fed-back using visual or auditory signals, allowing the participants
to gain control over these otherwise unaware neuronal processes. NF may be used
to directly improve underlying neuronal deficits, and/or to establish more general self-
regulatory skills that may be used to compensate behavioral difficulties. The current
manuscript describes pathophysiological characteristics of ADHD, heterogeneity of
ADHD subtypes and gender differences, as well as frequently associated behavioral
problems such as oppositional defiant/conduct or tic disorder. It is discussed how NF
may be helpful as a treatment approach within these contexts.
Keywords: Neurofeedback (NF), ADHD, ODD/CD, tic disorder, comorbidity, children, neurobiology
Introduction
Difficulties with Inattention or Hyperactivity and Impulsivity as the core symptoms of Attention
deficit Hyperactivity disorder (ADHD) are a frequent psychosocial burden. With an early onset
during childhood, ADHD is often persisting throughout life. It is a heterogeneous disorder, and a
challenge to treat. In light of this heterogeneity, the most promising treatment approach should
be multimodal in nature (Taylor et al., 2004; Swanson et al., 2008). Pharmacological interventions
particularly with stimulants such as methylphenidate and amphetamine sulfate, as well as non-
s.
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
HEALTH DISPARITIES: DIFFERENCES IN VETERAN AND NON-VETERAN POPULATIONS USING ...hiij
Introduction: This study investigated self-reported health status, health screenings, vision problems, and
vaccination rates among veteran and non-veteran groups to uncover health disparities that are critical for
informed health system planning for veteran populations.
Methods: Using public-use data from the National Health Interview Survey (2015-2018), this study adopts
an ecologic cross-sectional approach to conduct an in-depth analysis and visualization of the data assisted
by Generative AI, specifically ChatGPT-4. This integration of advanced AI tools with traditional
epidemiological principles enables systematic data management, analysis, and visualization, offering a
nuanced understanding of health dynamics across demographic segments and highlighting disparities
essential for veteran health system planning.
Findings: Disparities in self-reports of health outcomes, health screenings, vision problems, and
vaccination rates were identified, emphasizing the need for targeted interventions and policy adjustments.
Conclusion: Insights from this study could inform health system planning, using epidemiological data
assessment to suggest enhancements for veteran healthcare delivery. These findings highlight the value of
integrating Generative AI with epidemiological analysis in shaping public health policy and health
planning.
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
BRIEF COMMENTARY: USING A LOGIC MODEL TO INTEGRATE PUBLIC HEALTH INFORMATICS ...hiij
The COVID-19 pandemic has been a watershed moment in public health surveillance, highlighting the
crucial role of data-driven insights in informing health actions and policies. Revisiting key concepts—
public health, epidemiology in public health practice, public health surveillance, and public health
informatics—lays the foundation for understanding how these elements converge to create a robust public
health surveillance system framework. Especially during the COVID-19 pandemic, this integration was
exemplified by the WHO efforts in data dissemination and the subsequent global response. The role of
public health informatics emerged as instrumental in this context, enhancing data collection, management,
analysis, interpretation, and dissemination processes. A logic model for public health surveillance systems
encapsulates the integration of these concepts. It outlines the inputs and outcomes and emphasizes the
crucial actions and resources for effective system operation, including the imperative of training and
capacity development.
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
AUTOMATIC AND NON-INVASIVE CONTINUOUS GLUCOSE MONITORING IN PAEDIATRIC PATIENTShiij
Glycated haemoglobin does not allow you to highlight the effects that food choices, physical activity and
medications have on your glycaemic control day by day. The best way to monitor and keep track of the
immediate effects that these have on your blood sugar levels is self-monitoring, therefore the use of a
glucometer. Thanks to this tool you have the possibility to promptly receive information that helps you to
intervene in the most appropriate way, bringing or keeping your blood sugar levels as close as possible to
the reference values indicated by your doctor. Currently, blood glucose meters are used to measure and
control blood glucose. Diabetes is a fairly complex disease and it is important for those who suffer from it
to check their blood sugar (blood sugar) periodically throughout the day to prevent dangerous
complications. Many children newly diagnosed with diabetes and their families may face unique challenges
when dealing with the everyday management of diabetes, including treatments, adapting to dietary
changes, and the routine monitoring of blood glucose. Many questions may also arise when selecting a
blood glucose meter for paediatric patients. With current blood glucose meters, even with multiple daily
self-tests, high and low blood glucose levels may not be detected. Key factors that may be considered when
selecting a meter include accuracy of the meter; size of the meter; small sample size required for testing;
ease of use and easy-to-follow testing procedure; ability for alternate testing sites; quick testing time and
availability of results; ease of portability to allow testing at school and during leisure time; easyto- read
numbers on display; memory options; cost of meter and supplies. In this study we will show a new
automatic portable, non-invasive device and painless for the daily continuous monitoring (24 hours a day)
of blood glucose in paediatric patients.
INTEGRATING MACHINE LEARNING IN CLINICAL DECISION SUPPORT SYSTEMShiij
This review article examines the role of machine learning (ML) in enhancing Clinical Decision Support
Systems (CDSSs) within the modern healthcare landscape. Focusing on the integration of various ML
algorithms, such as regression, random forest, and neural networks, the review aims to showcase their
potential in advancing patient care. A rapid review methodology was utilized, involving a survey of recent
articles from PubMed and Google Scholar on ML applications in healthcare. Key findings include the
demonstration of ML's predictive power in patient outcomes, its ability to augment clinician knowledge,
and the effectiveness of ensemble algorithmic approaches. The review highlights specific applications of
diverse ML models, including moment kernel machines in predicting surgical outcomes, k-means clustering
in simplifying disease phenotypes, and extreme gradient boosting in estimating injury risk. Emphasizing
the potential of ML to tackle current healthcare challenges, the article highlights the critical role of ML in
evolving CDSSs for improved clinical decision-making and patient care. This comprehensive review also
addresses the challenges and limitations of integrating ML into healthcare systems, advocating for a
collaborative approach to refine these systems for safety, efficacy, and equity.
BRIEF COMMENTARY: USING A LOGIC MODEL TO INTEGRATE PUBLIC HEALTH INFORMATICS ...hiij
The COVID-19 pandemic has been a watershed moment in public health surveillance, highlighting the
crucial role of data-driven insights in informing health actions and policies. Revisiting key concepts—
public health, epidemiology in public health practice, public health surveillance, and public health
informatics—lays the foundation for understanding how these elements converge to create a robust public
health surveillance system framework. Especially during the COVID-19 pandemic, this integration was
exemplified by the WHO efforts in data dissemination and the subsequent global response. The role of
public health informatics emerged as instrumental in this context, enhancing data collection, management,
analysis, interpretation, and dissemination processes. A logic model for public health surveillance systems
encapsulates the integration of these concepts. It outlines the inputs and outcomes and emphasizes the
crucial actions and resources for effective system operation, including the imperative of training and
capacity development.
INTEGRATING MACHINE LEARNING IN CLINICAL DECISION SUPPORT SYSTEMShiij
This review article examines the role of machine learning (ML) in enhancing Clinical Decision Support
Systems (CDSSs) within the modern healthcare landscape. Focusing on the integration of various ML
algorithms, such as regression, random forest, and neural networks, the review aims to showcase their
potential in advancing patient care. A rapid review methodology was utilized, involving a survey of recent
articles from PubMed and Google Scholar on ML applications in healthcare. Key findings include the
demonstration of ML's predictive power in patient outcomes, its ability to augment clinician knowledge,
and the effectiveness of ensemble algorithmic approaches. The review highlights specific applications of
diverse ML models, including moment kernel machines in predicting surgical outcomes, k-means clustering
in simplifying disease phenotypes, and extreme gradient boosting in estimating injury risk. Emphasizing
the potential of ML to tackle current healthcare challenges, the article highlights the critical role of ML in
evolving CDSSs for improved clinical decision-making and patient care. This comprehensive review also
addresses the challenges and limitations of integrating ML into healthcare systems, advocating for a
collaborative approach to refine these systems for safety, efficacy, and equity.
Health Informatics - An International Journal (HIIJ)hiij
Healthcare Informatics: An International Journal is a quarterly open access peer-reviewed journal that Publishes articles which contribute new results in all areas of the health care.
The journal focuses on all of aspect in theory, practices, and applications of Digital Health Records, Knowledge Engineering in Health, E-Health Information, and Information Management in healthcare, Bio-Medical Expert Systems, ICT in health promotion and related topics. Original contributions are solicited on topics covered under the broad areas such as (but not limited to) listed below:
The Proposed Guidelines for Cloud Computing Migration for South African Rural...hiij
It is now overdue for the hospitals in South African rural areas to implement cloud computing technologies in order to access patient data quickly in an emergency. Sometimes medical practitioners take time to attend patients due to the unavailability of kept records, leading to either a loss of time or the reassembling of processes to recapture lost patient files. However, there are few studies that highlight challenges faced by rural hospitals but they do not recommend strategies on how they can migrate to cloud computing. The purpose of this paper was to review recent papers about the critical factors that influence South African hospitals in adopting cloud computing. The contribution of the study is to lay out the importance of cloud computing in the health sectors and to suggest guidelines that South African rural hospitals can follow in order to successfully relocate into cloud computing.The existing literature revealed that Hospitals may enhance their record-keeping procedures and conduct business more effectively with the help of the cloud computing. In conclusion, if hospitals in South African rural areas is to fully benefit from cloud-based records management systems, challenges relating to data storage, privacy, security, and the digital divide must be overcome.
SUPPORTING LARGE-SCALE NUTRITION ANALYSIS BASED ON DIETARY SURVEY DATAhiij
While online survey systems facilitate the collection on copious records on diet, exercise and other healthrelated data, scientists and other public health experts typically must download data from those systems
into external tools for conducting statistical analyses. A more convenient approach would enable
researchers to perform analyses online, without the need to coordinate additional analysis tools. This
paper presents a system illustrating such an approach, using as a testbed the WAVE project, which is a 5-
year childhood obesity prevention initiative being conducted at Oregon State University by health scientists
utilizing a web application called WavePipe. This web application has enabled health scientists to create
studies, enrol subjects, collect physical activity data, and collect nutritional data through online surveys.
This paper presents a new sub-system that enables health scientists to analyse and visualize nutritional
profiles based on large quantities of 24-hour dietary recall records for sub-groups of study subjects over
any desired period of time. In addition, the sub-system enables scientists to enter new food information
from food composition databases to build a comprehensive food profile. Interview feedback from novice
health science researchers using the new functionality indicated that it provided a usable interface and
generated high receptiveness to using the system in practice.
AN EHEALTH ADOPTION FRAMEWORK FOR DEVELOPING COUNTRIES: A SYSTEMATIC REVIEWhiij
#Health #clinic #education #StaySafe #pharmacy #healthylifestyle
call for papers..!
-----------------------------
Health Informatics: An International Journal (HIIJ)
ISSN : 2319 - 2046 (Online); 2319 - 3190 (Print)
Here's where you can reach us : hiij@aircconline.com
visit us on : https://airccse.org/journal/hiij/index.html
**************
published articles..!
AN EHEALTH ADOPTION FRAMEWORK FOR
DEVELOPING COUNTRIES: A SYSTEMATIC REVIEW
https://aircconline.com/hiij/V10N3/10321hiij01.pdf
GENDER DISPARITYOF TUBERCULOSISBURDENIN LOW-AND MIDDLE-INCOME COUNTRIES: A SY...hiij
The tuberculosis burden is higher in the population from low- and middle-income countries (LMICs) and
differently affects gender. This review explored risk factors that determine gender disparity in tuberculosis
in LMICs. The research design was a systematic review. Three databases; Google Scholar, PubMed, and
HINARI provided 69 eligible papers.The synthesized data were coded, grouped and written in a descriptive
narrative style. HIV-TB co-infected women had a higher risk of mortality than TB-HIV-infected men. The
risk of Vitamin-D deficiency-induced tuberculosis was higher in women than in men. Lymph node TB,
breast TB, and cutaneous and abdominal TB occurred commonly in women whereas pleuritis, miliary TB,
meningeal TB, pleural TB and bone and joint TB were common in men. Employed men had higher contact
with tuberculosis patients and an increased chance of getting the disease. Migrant women were more likely
to develop tuberculosis than migrant men. The TB programmers and policymakers should balance the
different gaps of gender in TB-related activities and consider more appropriate approaches to be genderbased and have equal access to every TB-associated healthcare.
BRIEF COMMUNICATIONS DATA HYGIENE: IMPORTANT STEP IN DECISIONMAKING WITH IMPL...hiij
Medical and health data that have been entered into an electronic data system in real-time cannot be
assumed to be accurate and of high quality without verification. The adoption of the electronic health
record (EHR) by many countries to the support care and treatment of patients illustrates the importance of
high quality data that can be shared for efficient patient care and the operation of healthcare systems.
This brief communication provides a high-level overview of an EHR system and practices related to high
data quality and data hygiene that could contribute to the analysis and interpretation of EHR data for use
in patient care and healthcare system administration.
Dev Dives: Train smarter, not harder – active learning and UiPath LLMs for do...UiPathCommunity
💥 Speed, accuracy, and scaling – discover the superpowers of GenAI in action with UiPath Document Understanding and Communications Mining™:
See how to accelerate model training and optimize model performance with active learning
Learn about the latest enhancements to out-of-the-box document processing – with little to no training required
Get an exclusive demo of the new family of UiPath LLMs – GenAI models specialized for processing different types of documents and messages
This is a hands-on session specifically designed for automation developers and AI enthusiasts seeking to enhance their knowledge in leveraging the latest intelligent document processing capabilities offered by UiPath.
Speakers:
👨🏫 Andras Palfi, Senior Product Manager, UiPath
👩🏫 Lenka Dulovicova, Product Program Manager, UiPath
Securing your Kubernetes cluster_ a step-by-step guide to success !KatiaHIMEUR1
Today, after several years of existence, an extremely active community and an ultra-dynamic ecosystem, Kubernetes has established itself as the de facto standard in container orchestration. Thanks to a wide range of managed services, it has never been so easy to set up a ready-to-use Kubernetes cluster.
However, this ease of use means that the subject of security in Kubernetes is often left for later, or even neglected. This exposes companies to significant risks.
In this talk, I'll show you step-by-step how to secure your Kubernetes cluster for greater peace of mind and reliability.
Key Trends Shaping the Future of Infrastructure.pdfCheryl Hung
Keynote at DIGIT West Expo, Glasgow on 29 May 2024.
Cheryl Hung, ochery.com
Sr Director, Infrastructure Ecosystem, Arm.
The key trends across hardware, cloud and open-source; exploring how these areas are likely to mature and develop over the short and long-term, and then considering how organisations can position themselves to adapt and thrive.
Accelerate your Kubernetes clusters with Varnish CachingThijs Feryn
A presentation about the usage and availability of Varnish on Kubernetes. This talk explores the capabilities of Varnish caching and shows how to use the Varnish Helm chart to deploy it to Kubernetes.
This presentation was delivered at K8SUG Singapore. See https://feryn.eu/presentations/accelerate-your-kubernetes-clusters-with-varnish-caching-k8sug-singapore-28-2024 for more details.
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Tobias Schneck
As AI technology is pushing into IT I was wondering myself, as an “infrastructure container kubernetes guy”, how get this fancy AI technology get managed from an infrastructure operational view? Is it possible to apply our lovely cloud native principals as well? What benefit’s both technologies could bring to each other?
Let me take this questions and provide you a short journey through existing deployment models and use cases for AI software. On practical examples, we discuss what cloud/on-premise strategy we may need for applying it to our own infrastructure to get it to work from an enterprise perspective. I want to give an overview about infrastructure requirements and technologies, what could be beneficial or limiting your AI use cases in an enterprise environment. An interactive Demo will give you some insides, what approaches I got already working for real.
Elevating Tactical DDD Patterns Through Object CalisthenicsDorra BARTAGUIZ
After immersing yourself in the blue book and its red counterpart, attending DDD-focused conferences, and applying tactical patterns, you're left with a crucial question: How do I ensure my design is effective? Tactical patterns within Domain-Driven Design (DDD) serve as guiding principles for creating clear and manageable domain models. However, achieving success with these patterns requires additional guidance. Interestingly, we've observed that a set of constraints initially designed for training purposes remarkably aligns with effective pattern implementation, offering a more ‘mechanical’ approach. Let's explore together how Object Calisthenics can elevate the design of your tactical DDD patterns, offering concrete help for those venturing into DDD for the first time!
Neuro-symbolic is not enough, we need neuro-*semantic*Frank van Harmelen
Neuro-symbolic (NeSy) AI is on the rise. However, simply machine learning on just any symbolic structure is not sufficient to really harvest the gains of NeSy. These will only be gained when the symbolic structures have an actual semantics. I give an operational definition of semantics as “predictable inference”.
All of this illustrated with link prediction over knowledge graphs, but the argument is general.
UiPath Test Automation using UiPath Test Suite series, part 4DianaGray10
Welcome to UiPath Test Automation using UiPath Test Suite series part 4. In this session, we will cover Test Manager overview along with SAP heatmap.
The UiPath Test Manager overview with SAP heatmap webinar offers a concise yet comprehensive exploration of the role of a Test Manager within SAP environments, coupled with the utilization of heatmaps for effective testing strategies.
Participants will gain insights into the responsibilities, challenges, and best practices associated with test management in SAP projects. Additionally, the webinar delves into the significance of heatmaps as a visual aid for identifying testing priorities, areas of risk, and resource allocation within SAP landscapes. Through this session, attendees can expect to enhance their understanding of test management principles while learning practical approaches to optimize testing processes in SAP environments using heatmap visualization techniques
What will you get from this session?
1. Insights into SAP testing best practices
2. Heatmap utilization for testing
3. Optimization of testing processes
4. Demo
Topics covered:
Execution from the test manager
Orchestrator execution result
Defect reporting
SAP heatmap example with demo
Speaker:
Deepak Rai, Automation Practice Lead, Boundaryless Group and UiPath MVP
DevOps and Testing slides at DASA ConnectKari Kakkonen
My and Rik Marselis slides at 30.5.2024 DASA Connect conference. We discuss about what is testing, then what is agile testing and finally what is Testing in DevOps. Finally we had lovely workshop with the participants trying to find out different ways to think about quality and testing in different parts of the DevOps infinity loop.
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualityInflectra
In this insightful webinar, Inflectra explores how artificial intelligence (AI) is transforming software development and testing. Discover how AI-powered tools are revolutionizing every stage of the software development lifecycle (SDLC), from design and prototyping to testing, deployment, and monitoring.
Learn about:
• The Future of Testing: How AI is shifting testing towards verification, analysis, and higher-level skills, while reducing repetitive tasks.
• Test Automation: How AI-powered test case generation, optimization, and self-healing tests are making testing more efficient and effective.
• Visual Testing: Explore the emerging capabilities of AI in visual testing and how it's set to revolutionize UI verification.
• Inflectra's AI Solutions: See demonstrations of Inflectra's cutting-edge AI tools like the ChatGPT plugin and Azure Open AI platform, designed to streamline your testing process.
Whether you're a developer, tester, or QA professional, this webinar will give you valuable insights into how AI is shaping the future of software delivery.
GraphRAG is All You need? LLM & Knowledge GraphGuy Korland
Guy Korland, CEO and Co-founder of FalkorDB, will review two articles on the integration of language models with knowledge graphs.
1. Unifying Large Language Models and Knowledge Graphs: A Roadmap.
https://arxiv.org/abs/2306.08302
2. Microsoft Research's GraphRAG paper and a review paper on various uses of knowledge graphs:
https://www.microsoft.com/en-us/research/blog/graphrag-unlocking-llm-discovery-on-narrative-private-data/
Parkinson's disease motor symptoms in machine learning a review
1. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
PARKINSON'S DISEASE MOTOR SYMPTOMS IN
MACHINE LEARNING: A REVIEW
Claas Ahlrichs and Michael Lawo
Mathematics and Computer Science, University of Bremen,
PO Box 330 440, 28334 ,Bremen, Germany
ABSTRACT
This paper reviews related work and state-of-the-art publications for recognizing motor symptoms of
Parkinson's Disease (PD). It presents research efforts that were undertaken to inform on how well
traditional machine learning algorithms can handle this task. In particular, four PD related motor
symptoms are highlighted (i.e. tremor, bradykinesia, freezing of gait and dyskinesia) and their details
summarized. Thus the primary objective of this research is to provide a literary foundation for development
and improvement of algorithms for detecting PD related motor symptoms.
KEYWORDS
Parkinson's Disease, Machine Learning, Artificial Intelligence, Review, State-of-the-Art
1. INTRODUCTION
This research focuses on algorithms for detecting Parkinson's disease (PD) related symptoms in
time series data. PD is a disorder of the central nervous system resulting in a loss of motor
function, increased slowness and rigidity. Artificial intelligence (AI)-based techniques can be
utilized to detect symptoms such as tremor or bradykinesia while focusing on minimizing false
negatives (i.e. failing to recognize a symptom) and false positives (i.e. detection of a symptom
where none is apparent). Those affected by PD bear a great burden and have to cope with a rather
reduced quality of life. In the authors' eyes, this is an even more pressing issue when considering
leading role of Germany. In 2004, Germany inhabited the largest number of people with
Parkinson's within Europe [3].
Even though it can manifest itself at any age, PD is among other diseases (e.g. Alzheimer's,
dementia, chronic bronchitis) usually attributed to elderly subgroups of the population.
Considering demographic changes of the last decades, the number of cases and burden of PD is
expected to increase [24, p. 36]. The World Health Organisation (WHO) estimates that around 5.2
million people were suffering from PD worldwide in 2004 [40]. Depending on the estimating
organization, Europe inhabited 1.2 [24] - 2.0 [40] million of them in the same year.
PD is typically characterized as a chronic, progressive, neurodegenerative disorder [4], [26], [58],
[20], [27]. The cardinal symptoms are bradykinesia, rigidity, tremor and postural instability [26],
[58], [27], [20], [23], [60], [3], [32]. Among many other symptoms, these symptoms result from a
dopamine deficiency in the substantia nigra. A part of the brain that is located within the basal
ganglia circuit (see Figure 1). Dopamine is a neurotransmitter involved in movement control [69].
Usually by the time of diagnosis, a great number of dopamine-producing neurons have already
DOI: 10.5121/hiij.2013.2401
1
2. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
diminished [58]. Current treatments aim at slowing the progression of the disease, focus on
symptomatic relief and attempt to lift the enormous burden of PD. However, a cure is yet to be
found.
Figure 1. Illustrates structures of the brain related to the basal ganglia circuit and substantia nigra.
Latter is located in the upper end of the brain stem. The image is based on a figure which has
been retrieved from Wikimedia Commons and belongs to the public domain.
PD is a great burden, not just for people suffering from the disease but also for those being
indirectly affected (i.e. relatives and caretakers). In an advanced stage of the disease and without
proper treatment, patients are no longer capable of taking care of themselves. In the Global
Burden of Disease study, the WHO rated PD to be on the same disability level as: amputated arm,
congestive heart failure, deafness, drug dependence and tuberculosis [40, p. 33].
A large number of symptoms have been shown by people with Parkinson's [27], [26]. The most
visible and easily noticeable symptoms are related to motor functions. However, quality of life is
affected by an even greater number of motor and non motor symptoms (e.g. depression, sleep
disorder, cognitive / neurobehavioral abnormalities, autonomic and gastrointestinal dysfunction)
[27], [26], [3]. As the disease progresses, patient's symptoms change and fluctuate (i.e. some
symptoms simply disappear, while others (re-) appear), creating a unique symptomatic history for
each individual patient. Unfortunately, in an advanced stage of Parkinson's further (drug-induced)
symptoms may become apparent. Dyskinesia is one of these symptoms and results from a lengthy
pharmacological treatment (i.e. several years). It manifests itself as an involuntary movement of
entire body parts (e.g. rhythmical moving of upper body).
Tremor at rest (also known as rest tremor or resting tremor) is only present when muscles are at
rest and dissolves during sleep as well as with action (i.e. voluntary movement of affected
extremity) [34]. It manifests itself as an involuntary, unilateral (one-sided) shaking of an
extremity (e.g. hand, foot, etc.). The shaking generally occurs at a frequency between 4-6 Hz
[27].
Bradykinesia refers to slowness of movement [27], [34]. It usually appears in very early stages of
the disease[58] and it is characteristic for basal ganglia disorders [27]. Depending on the severity,
2
3. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
movements may not only be slowed (bradykinesia), but also diminished (hypokinesia) or
completely abrogated (akinesia).
Freezing of gait (FOG) (also known as freezing or motor blocks) is a form of akinesia which
presents itself as an inability to initiate or continue movement [27], [58]. Motor blocks are a
common symptom, experienced by people with Parkinson's (although it does not occur
uniformly) and can affect various extremities (e.g. arms and legs) as well as the face [27]. After
onset of the symptom, it typically lasts for several seconds and disappears afterward. It is a
common cause of falls [27], [58].
Publications reveal a great number of techniques for automatic detection of PD motor symptoms
which employ various AI-based methods such as neural networks (NNs) [30], [15], [5], [9], [55],
[21], [16], [14], hidden markov models (HMMs) [54] and support vector machines (SVMs) [9],
[50], [14]. Depending on the symptom and utilized sensors, various features are calculated (e.g.
entropy [9], [50], [14], [43], spectral or fractal features [64], [61], [66], [5], [57], [44], [47], [29],
[10], [11]). Over time, sensor signals are analyzed and compared or set in relationship to known
samples of each symptom in order to recognize them. No matter whether these AI methods are
continuous or window-based, all of them can be viewed as either data mining techniques and / or
time series (analysis) algorithms. Much literature presents algorithms for detecting a single
symptom (e.g. [63], [66], [54], [5], [7], [45], [39], [21], [16], [14]). Considering the
heterogeneous nature of symptom profiles in PD patients, this is not sufficient. Few publications
focus on detecting of multiple motor symptoms (e.g. [57], [15], [55], [50]), but even those rarely
consider enough symptoms for use in real-world scenarios. In reality, patients are likely to
experience multiple symptoms, thus increasing the chance of false negatives and false positives.
To summarize, the objective of this research is to provide a literary foundation for
development and improvement of algorithms for detecting PD related motor symptoms.
2. IDENTIFYING PARKINSON'S DISEASE AND ITS SYMPTOMS
Much research has been published with a focus on biological, chemical and genetic aspects of
PD. Over the last two decades, an increasing number of publications originate from fields like
computer science or AI, focusing on signifying motor symptoms in people with Parkinson's.
Some of which are dedicated to detecting single motor symptoms [63], [66], [54], [5], [7], [45],
[39], [21], [16], [14] and others on detection of multiple symptoms [57], [15], [55], [50]. These
publications reveal a great number of techniques for automatically indicating the presence of PD
motor symptoms (e.g. NNs [30], [15], [5], [9], [55], [21], [16], [14], HMMs [54] and SVMs [9],
[50], [14]). Depending on symptom and utilized sensors, various features have been proposed and
applied in literature (e.g. entropy [9], [50], [14], [43], spectral or fractal features [64], [61], [66],
[5], [57], [44], [47], [29], [10], [11]) are known to be used in this context. In the course of this
section a strong focus will be on the most common symptoms that are experienced by PD
patients. Preference is given to those publications that do not focus on a single symptom (as
opposed to multiple symptoms) or use synthetic datasets (as opposed to data recorded from
sensors on the subject's body) but rather use unconstrained and unscripted activities of daily
living (ADL).
It should be kept in mind that there are many other publications with a focus on PD symptom
indication of their severity, but do not make use of body-mounted sensors or otherwise do not
resemble closely the previously elaborated criteria. Despite this reservation, a few selected
publications that do not fit these criteria are presented nonetheless.
3
4. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
2.1. TREMOR AT REST
In an early study by Salarian et al. [56], they were able to achieve a specificity of 98% and
sensitivity of 76.6% on a dataset with ten patients and ten control subjects. In total, close to
twenty hours worth of data were captured by the authors. Two tri-axial gyroscopes (i.e. one on
each wrist) were used to record data while participants performed a set of scripted everyday
activities. Spectral analysis was used to filter interesting regions within the specific frequency
range specific to resting tremor(i.e. 3.5Hz-7.5Hz). In a later study [57] based on the same dataset,
the data stream was divided into chunks with a length of three seconds to which the Burg method
[13] was applied. Additionally a meta-analysis was introduced to remove isolated segments that
were classified to exhibit tremor or tremor-like behavior (e.g. a single segment with tremor
surrounded by none-tremor segments). This increased the sensitivity to 99.5% but lowered the
specificity to 94.2%.
In [72], inertial sensor data (i.e. acceleration and angular velocity) were gathered from six patients
and seven control subjects. Zwartjes et al. had captured approximately 1.5-2 hours worth of data
while participants were performing a set of scripted activities in laboratory conditions. A multistaged algorithm is utilized to indicate regions of tremor. At first some preprocessing is applied to
the raw data, which is then used to classify the subject's activity and / or posture. This preclassification is used to highlight regions of interest where tremor is more noticeable (e.g. arms
are hanging still while standing). If activity or posture were usable for detection of tremor (and its
severity) then those portions of the data stream are divided into segments of three seconds length
with a two-thirds overlap. For each segment, the Fourier Transform is used to identify tremor
specific frequencies and thus tremor episodes as well. In the algorithm's last stage, a metaanalysis removes isolated segments of tremor (very similar to the process that was utilized by
Salarian et al. in [57]). Zwartjes et al. achieved an accuracy of about 84.7%. However, in
comparison to studies by Salarian et al. [56],[57], the recorded activities were less constrained.
Rigas et al. [54] achieved an accuracy of 87% in detecting tremor in an accelerometer based
dataset with twenty-three participants (i.e. ten patients and thirteen control subjects). All
participants performed daily activities in laboratory conditions. The data stream is divided into
three second windows with 50% overlap. Having applied standard filtering and analysis
techniques (i.e. finite response (FIR) filters, Fast Fourier Transform, etc.) an HMM is utilized to
detect tremor episodes. This is different from most algorithms for tremor indication. More
common approaches rely on spectral features alone [64], [61], [66], [5], [57], [44], [47], [29],
[10], [11] while other classify based on NNs [30], [15], [5], [9], [55], [21], [16], [14] or SVMs
[9], [50], [14]. Rigas et al. state that HMMs are suitable for tremor indication because "tremor
presents time-dependency" [54]. They consider HMMs as a time sensitivity extension of the naive
Bayes classifier.
Cole et al. [15] were able to detect tremor with a sensitivity of 93% and specificity of 95% in
unconstrained and unscripted activities. The dataset contained about 48 hours worth of
acceleration and electromyogram (EMG) measurements from twelve participants (i.e. eight
patients and four control subjects). Here a dynamic neural network (DNN) is used in combination
with a set of FIR filters to detect tremor. It is stated by the authors that DNNs [67] were utilized
because they are more capable of learning and classifying time-dependent classes (e.g. tremor)
when compared to regular and static neural networks. Cole et al. divide the data stream into
segments of two seconds length for feature extraction. The features are simply passed to the
DNN, where artificial neurons do their work. However, the neurons' outputs are not simply
forwarded to the next layer of neurons. Instead, each neuron has an FIR filter attached to it which
transforms the output before it is passed to subsequent neurons. Their results are mainly
4
5. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
dependent on the choice of training data. Here a handcrafted representative subset of data was
chosen.
A dataset with nineteen patients and four control subjects was used by Roy et al. [55] to signify
tremor. They achieved a sensitivity of 91.2% and a specificity of 93.4% in EMG and acceleration
data. The participants were performing unscripted and unconstrained activities in a home-like
environment for several hours. Here the data stream is also divided into two second windows and
a combination of DNNs with FIR filters is fed with various features that were extracted from the
two second segments.
Niazmand et al. [46] collected data from accelerometers integrated into a pullover. Ten patients
and two healthy control subjects performed standardized PD motor tasks. An average sensitivity
of 80% in indicating postural tremor and resting tremor and a specificity of 98.5% was achieved
by the authors. Their algorithm first determines the relative acceleration among the sensors and
then determines the movement frequency. This is done because sensors are not fixed on the
patient's body but rather in a garment which position can change depending on executed
movements. The raw data is simply filtered, normalized and a noise removal method is applied.
For determining the movement frequency, a combination of thresholds and peak counting is
utilized.
2.2. BRADYKINESIA
In [14], Cancela et al. present a motor symptom monitoring and management system. Their work
originates from a European research project called PERFORM (Personal Health Systems for
Monitoring and Point-of-Care Diagnostics-Personalized Monitoring). Here a set of classification
algorithms (e.g. SVM, k-nearest neighbors (KNN), NN, decision tree (DT), etc.) was evaluated.
The highest accuracy of 86% was achieved by the SVM. The corresponding dataset consists of
acceleration data from twenty patients performing a set of ADL (within the limits of a scripted
protocol). A standard analysis procedure is used by Cancela et al. At first a Butterworth filter is
applied to raw sensor data then the data stream is epoched in five second segments with a 50%
overlap. A set of features (i.e. sample entropy, root mean square, cross correlation, etc.) is
calculated for each segment and passed to the classification algorithms. Here, the algorithms
classify presence and severity of bradykinesia. Interestingly, the severity is not derived from
standard motor tasks but instead from ADL.
Cancela was also involved in a publication by Pastorino et al. [49]. Here a slightly modified
version of Cancela's algorithm is utilized (as in [14]). A dataset from twenty-four patients
performing unconstrained and unscripted activities at their home was gathered for a week. Twice
a day, a clinician came to visit the patient and performed a short protocolized session which was
later used to test the previously developed algorithm. Pastorino et al. show that an additional
meta-analysis can improve classification results. Instead of using the generated outputs from the
SVM directly, they can be further filtered / smoothed to ignore impossible and unrealistic
scenarios. Using a patient independent algorithm an accuracy of 68.3% ± 8.9% was achieved and
a 74.4% ± 14.9% accuracy was achieved with the additional meta-analysis. They indicate that a
patient specific training of the algorithms would likely lead to improved results.
Salarian et al. were not only involved in detecting tremor in time series data, they were also using
gyroscopes on the wrists to indicate the presence of bradykinesia. In [56], ten patients and ten
healthy control subjects participated in the collection of twenty hours worth of data. All
participants were performing scripted ADL. Salarian et al. showed that the features rotation of
hand (RH) and mobility of hand (MH) correlate well with the clinician's ground truth (r=-0.84 and
r=-0.83 for MH and RH respectively and p<0.00001). Several years later, Salarian et al. were able
5
6. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
to reproduce their results in [57]. However in latter publication, window sizes of five minutes and
above were used. Even though their work does not produce results in real-time, it does give hope
that not many sensors are required for a decent accuracy in bradykinesia detection.
The authors Zwartjes et al. [72] were able to identify bradykinesia related parameters that
correlate well with the patient's unified Parkinson's Disease rating scale (UPDRS) scores. Here a
dataset based on accelerometers and gyroscopes from six patients and seven healthy control
subjects was analyzed. All subjects performed a mixture of standardized motor tasks and ADLs in
a random predefined order. An activity / posture classifier is used to identify a set of elementary
activities (i.e. walking, standing up) and postures (i.e. standing and sitting). For upper extremities,
an average arm acceleration is calculated while various gait-related features (i.e. step length, step
velocity, etc.) are determined from a tri-axial gyroscope and a tri-axial accelerometer that are
placed on a foot. These features provide the basis for bradykinesia (slowness of movement) and
hypokinesia (poverty of movement) quantification. The authors' results indicate that a significant
correlation is present in almost all bradykinesia-related parameters while "none of the
hypokinesia-related parameters were significantly correlated" [72].
2.3. AKINESIA / FREEZING OF GAIT
In [21], Djurić-Jovičić et al. employed a neural network and a simple thresholding technique to
classify walking patterns in PD patients. A set of six inertial measurement units, each containing a
tri-axial accelerometer and a tri-axial gyroscope, were attached to the subjects' legs (i.e. thigh and
shin) as well as their feet. The kinematics of four patients (as they were following a
predetermined path) were gathered, annotated and used to train a neural network. In total, about
30 minutes of data were collected. The path itself included several (potential) hurdles which have
been designed to provoke FOG (e.g. start hesitation, destination hesitation, narrow path or turn
hesitation, etc.). A combination of heuristically determined thresholds and a NN were utilized to
differentiate between "normal" (i.e. standing and regular steps) and pathological (i.e. festination,
akinesia, shuffling and small steps) walking patterns. The authors of [21] achieved an error rate as
high as 16% due to the choice of thresholds (i.e. thresholds were independent of patients, etc.).
On the contrary, the algorithm was working in real-time (i.e. about 0.5 seconds delay).
A similar technique was developed by Cole et al. in [16]. However, instead of a regular (static)
neural network a dynamic neural network [67] was utilized. Their indicator algorithm showed a
sensitivity of 82.9% and specificity of 97.3% in a dataset containing unconstrained and unscripted
activities. Ten patients and two healthy control subjects contributed and helped to gather about
two hours worth of data from several accelerometers (i.e. forearm, thigh and shin) and an EMG
sensor (i.e. shin). The authors employed a multi-staged algorithm [16]. In the first stage a simple
linear classifier determines whether the subject is in an upright position (i.e. standing and not
sitting or lying). If this is the case then a DNN determines episodes of akinesia. The idea was to
identify periods in the data stream were episodes of akinesia are more likely to be apparent (both
visually and in data stream). In contrast to a static NN, Cole et al. [16] have decided to use a DNN
because they are able to better capture time-varying weights that are present in FOG episodes.
An accelerometer based smart garment called MiMed-Pants [48] has been used by Niazmand et
al. [47] to extract and analyze gait-related features. In this case, the measurement device has been
successfully integrated in an item that is "suitable for daily use" [47]. The pair of pants can be
washed like a regular textile. A sensitivity of 88.3% and specificity of 85.3% has been achieved
with this setup. Five accelerometers (i.e. each shin, each thigh and belly button) provided
kinematics on six patients while they were performing standard activities [71] (i.e. walking
course, including narrow spaces, gait initialization and reaching destination, etc.). In total about
one hour worth of data was collected by Niazmand et al. No use of advanced artificial intelligence
6
7. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
methods was made in [47], but instead a linear classification was applied to features that were
extracted from the sensors. The algorithms provided feedback with a delay of about two seconds.
In 2009, Bächlin et al. published their work on a wearable and context-aware system for real-time
detection of FOG events [11]. The system provides acoustic feedback within a two second
window. A set of accelerometers and gyroscopes was utilized (e.g. on shank, thigh and waist).
Over eight hours worth of data were gathered from ten patients performing ADL, as well as
walking in a straight line and a random walk. Bächlin et al. claim to have built the first contextaware and wearable system to assist PD patients in detecting FOG events. An overall sensitivity
of 73.1% and 81.6% specificity were achieved in [11]. The results were mainly due to different
walking styles (that were used by the subjects in their dataset), choice of features and use of
patient independent thresholding. They state that a personalized training and choice of threshold
might have produced better results.
In [62], Stamatakis et al. were able to show differences in walking patterns between a PD patient
and a healthy control subject. The authors identified a set of features that may be used for
differentiating between PD patients and healthy control subjects as well as for detecting FOG
events and their duration. Even though no results in terms of accuracy or significance were
presented in [62], their presented features may prove to be beneficial.
2.4. DYSKINESIA
Keijsers et al. [30] were able to achieve an accuracy of 96.8% in detecting dyskinesia. Thirteen
participants were enrolled in their study. Each subject contributed about 2.5 hours of acceleration
data while they were performing a set of scripted activities (approximately 35) in a controlled
environment. In total six tri-axial acceleration sensors were attached to the subject's body (i.e. one
on each thigh, one on each shoulder, one on trunk and one on wrist) during their recording
session. The algorithm, used by Keijsers et al., classified fifteen minute segments using a regular
neural network. The output of the NN indicated the presence (or absence) of dyskinesia within the
segment. Several segment sizes were empirically evaluated (e.g. fifteen and one minute
segments). The best accuracy was achieved with the fifteen minutes segments (i.e. 96.8%).
However when using one minute segments the accuracy drops to about 80% on the same dataset.
In contrast, Tsipouras et al. [65] were able to achieve similar results but on smaller segments.
While Keijsers et al. [30] used fifteen minute segments, here two second intervals with 75%
overlap are utilized. Tsipouras et al. state to have achieved a 93.7% accuracy using their dataset.
This contains inertial sensor data (i.e. acceleration and angular rate) of four patients and six
control subjects. All participants were performing a set of scripted activities. In total two
gyroscopes (i.e. one on trunk and one on waist) and six accelerometers (i.e. one next to each
gyroscope, one on each arm and one on each leg) were used during recording sessions. Here five
classification methods were evaluated (i.e. naive Bayes, KNN, fuzzy lattice reasoning, DTs and
random forests (RFs)). The RFs performed best with 93.7%. C4.5 is close behind with about
93.5% while all other remaining classification algorithms achieved an accuracy around 85%.
Cole et al. [15] used a DNN to better capture the time-based variables. The authors were able to
achieve a 91% sensitivity and a 93% specificity in detecting dyskinesia. Here a similar procedure
to their work in detecting tremor [15] and bradykinesia [16] was utilized. Their dataset contained
several hours' worth of acceleration data and EMG measurements from eight patients as well as
four control subjects. Participants were performing unscripted and unconstrained activities during
their recording session. A set of features is extracted from a two second sliding window and fed
to the DNN. Additionally, outputs of each artificial neuron (i.e. node within the neural network)
are filtered using a five point FIR filter.
7
8. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
Similarly Roy et al. [55] combine DNNs with a rule-based reasoning method. They were able to
achieve a sensitivity of 90% and specificity of 93.4% in a dataset containing acceleration data and
EMG measurements from nineteen patients and four control subjects. One hybrid sensor
(containing a tri-axial accelerometer and an EMG sensor) was located on each arm and leg. All
participants were performing unconstrained and unscripted activities in a home-like environment.
In total about 30 hours worth of data were gathered and used by Roy et al. [55]. Again the
extracted features originate from two second segments. Their algorithm uses those features to
feed two DNNs (i.e. one for mobility states and one for motor states). These DNNs provide
preliminary results on patient's mobility state (i.e. sitting walking, standing, etc.) and motor
symptoms. They are used in combination with a framework called IPUS (Integrated Processing
and Understanding of Signals), which in turn activates different DNNs to maximize symptom
recognition rates (e.g. based on the fact that the subject is walking, sitting, etc.).
Patel et al. [51] utilized clustering techniques (and expectation maximization (EM)) to distinguish
various levels of severity in PD patients while they were performing standardized motor tasks. A
similar approach was used by Sherrill et al. [59]. Here six patients provided acceleration data to
which clustering was applied in order to detect dyskinesia.
2.5 DYSARTHIA AND DYSPHAGIA
In [53], Revett et al. employ a rough sets approach for distinguishing healthy subjects and people
with PD based on vocal data. Their dataset is based on thirty-one participants (i.e. twenty-three
patients and eight healthy controls) performing a phonation task. Little et al. [36] originally
constructed this dataset and donated it to the University of California Irvine (UCI) Machine
Learning Repository [1]. On average each participant performed six phonations of the vowel [a].
Thus resulting in close to 200 samples of which a set of twenty-three features (including spectral
features, shimmer, jitter, presence of PD, etc.) has been documented. Revett et al. report to have
achieved a 100% accuracy when using all available features in their rough sets approach. This
holds if the classification category is binary (i.e. healthy subject or PD patient). In this case
several hundred rules are generated to identify a data sample's category. When trying to reduce
the number of rules, the accuracy drops but stays well above 90% with about 100 rules. However,
the authors did not attempt to perform classification based on the patient's duration of the disease,
severity or UPDRS scores.
In a publication by Bakar et al. [8], they present a speech-based assessment tool for identifying
PD. Here the same dataset (as originally constructed by Little et al. [36]) has been utilized. The
authors performed several tests in which they compared testing accuracy, training accuracy,
average mean square error (MSE) as well as average number of iterations of two training /
learning algorithms for NNs (Levenberg-Marquardt (LM) and Scaled Conjugate Gradient (SCG)).
Their results indicate that LM outperforms the SCG algorithm. Generally speaking, the LMapproach resulted in better testing and training accuracy as well as a lower MSE while SCG
performed better in terms of "number of iterations". The best result was achieved with a 97.9%
accuracy in training and 93.0% accuracy in testing.
Asgari and Shafran [6] utilized a similar set of features for classifying speech and phonation data.
They performed a prediction of UPDRS motor scores based on these features. In more than one
hundred recording sessions, twenty-one control subjects and sixty-one patients were asked to
perform three tasks: sustained phonation (i.e. phonation of vowel [a]), diadochokinetic test (i.e.
repetition of syllables [pa], [ta] and [ka]) and a reading task. The actual recordings were done
with a dedicated device that was designed to be an at-home testing tool. The dataset itself was
analyzed with 100 frames per second using Hamming windows (each of 25ms length). A rather
large set of features is extracted from each frame and were generated for both voiced and
8
9. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
unvoiced segments. About 15K features were generated based on their recordings of phonation
and speech data (including pitch, frequency, harmony, etc.). A SVM was employed to translate
these features into UPDRS motor scores. Depending on the used features (or subset of features) a
mean absolute error between 6.1 and 5.7 (UPDRS) points was achieved.
In [41], Mekyska, Rektorova and Smekal evaluate a set of features for automatic analysis of
speech disorders in PD. The authors provide an extensive summary on various speech-related
parameters and highlight a few common problems in automatic speech analysis. Their dataset is
composed of forty-two male control subjects and twelve male PD patients. Each participant was
asked to pronounce all vowels (i.e. [a], [e], [i], [o], [u]) once in a natural speed and once slowly.
The inter-intra class distance ratio method (IICDR) and minimum redundancy maximum
relevance (MRMR) method were used by Mekyska et al. in order to sort out the top 20
parameters for each method (after having started with 510 parameters in total). In a second step,
the Jarque-Bera test was utilized to see which features show a normal probability distribution.
Those with a normal distribution were used in a multi-factor analysis of variance (ANOVA).
Their results show up to three features (i.e. “mean B-F1“ for p ≤ 5% as well as “mean F0” and
“mean NHR” for p ≤ 10%) that can be used to distinguish healthy control subjects from those
afflicted with PD. Mekyska et al. add that there are also several parameters which do not show a
coherent tendency in published literature. The authors point out several papers in which a
particular feature has been shown to be significant, non-significant and indifferent in terms of
separating patients from healthy subjects. However, they also comment that these conflicting
publications usually used rather small datasets. Thus they are more prone to random variations
and clusters.
Xiuming et al. [70] describe a diagnostic approach to PD based on principle component analysis
(PCA) and Sugeno integral. The authors employ a dataset by Little et al. [36] from UCI Machine
Learning Repository [1]. Thus data of thirty-one participants (i.e. eight patients and twenty-three
healthy control subjects) was analyzed. Xiuming et al. show five principle components which
account for 86.5% of the information within the signal. In order to propose a diagnosis, the
Sugeno measure and Sugeno integral are then determined for the top five most relevant features.
The authors report a classification accuracy of 81.0%.
2.6 OTHER
In 2000, Hamilton et al. [25] published their work on outcome prediction in pallidotomy in PD
patients. It was their goal to build a reliable tool which estimates an operation's outcome based on
intra-operational recordings of neural activity. A standard NN was employed and trained with a
set of features (e.g. signal power, entropy or fractal dimensions). This system could provide
supplementary data and aid surgeons in minimizing risks (e.g. blindness, difficulties in speaking
or swallowing, etc.) and maximizing effectiveness. Their results indicate that all evaluated NN
performed similarly in terms of overall outcome prediction. Hamilton and colleagues found that
their NNs handled exceptional cases well.
In terms of diagnosis Kupryjanow et al. [35] came up with an alternative measurement technique
for determining UPDRS sub-scores related to motor tests (i.e. finger tapping and rapid alternating
movement of hands). Instead of relying on arguably subjective assessments from neurologists,
they present a device called Virtual-Touchpad (VTP). Here a webcam is used to capture
movements of hands and translate them into machine readable features. In comparison to other
methods, this approach does not require equipment to be attached to the patient (or mounted on
the patient). A SVM recognizes hand gestures and / or postures. The succession of those postures
is used to extract the mentioned features and determine UPDRS scores. The authors did not
describe a user study in [35].
9
10. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
In [18], Cunningham et al. presented their work on a computerized assessment tool. The work is
intended to identify movement difficulties found in people with PD and similar movement
disorders. Here the participants' ability to point and click on targets on the computer screen is
compared among those afflicted with PD and healthy control subjects. A benefit of this approach
is that patients are not expected to wear "unusual" or specialized hardware. However, on the other
hand, it requires the patient to sit in front of a computer and cannot be mobile (as it would
considerably alter their ability to point and click). Their results show a difference in control
subjects and PD patients. The control group was generally more accurate (i.e. clicked closer to the
target's center and made less accidental clicks) and faster (i.e. required less time to click once the
target has been reached). This holds for both computer literates and computer illiterates. Those
being computer illiterates and suffering from PD showed a higher variance in terms of accuracy
of clicking the target center. In [19], Cunningham et al. present their tool's abilities in indicating
akinesia, bradykinesia, dyskinesia, rigidity and tremor.
In a preceding publication by Cunningham and colleagues [17], another study has been performed
with ten PD patients. Here participants were asked to use a home-based assessment tool twice a
day (i.e. once in ON state and once in OFF state) for a period of four days. It is their goal to
differentiate between a participant's ON state and OFF state based on their test performance. As
in other studies by Cunningham et al. [18], the subjects clicked on targets while their speed, time,
distance and location of click were recorded. Regarding the time, a statistical significance was
found when comparing performances in ON and OFF states (p = 0.017). The authors also note
that a few subjects showed an increased variance in ON-OFF state which indicates that they
might not have been in a clearly defined ON-OFF state at the time of testing. Nonetheless their
results appear to be promising and larger follow-up studies are to be seen.
Wang et al. [68] developed a new method for quantitative evaluation of symptoms in people with
PD. Their proposed method is based on free spiral drawings with a digitizing table. Spiral
drawings itself have been used a number of times to quantify motor dysfunctions (in particular
[38], [37], [42] were highlighted by Wang et al.). However, these methods were usually
employing some sort of guidance or template. In contrast, Wang et al. utilized free spiral
drawings. A group of ten participants enrolled in their study at a hospital in Japan (Kaizuka). All
of them were asked to draw a spiral which was then used to extract several features (e.g. number
of turns, mean of radius, maximum radius, etc.). Their results indicate that healthy control
subjects can be clearly separated from the remaining eight patients regarding the number of
extreme points in radius curve. Most PD subjects were not able to "rapidly enlarge the circle as
spiral" [68]. The authors remark that the features mean value of radius and slope of radius curve
demonstrated stiffness and could be also used to distinguish between healthy and pathological
subjects.
Pradhan et al. [52] considered a similar methodology, but instead of drawing spirals thirty PD
patients were tracking waves by applying force to sensors. Here two force sensors (i.e. one for
index finger and one for thumb) were "squeezed" in order to track a wave (i.e. simple sine wave
and complex wave with multiple frequency components) with and without mental distraction.
Their goal was to provide an assessment tool for clinical progression of PD patients. Pradhan and
colleagues state that similar studies have been performed but "which may not be effective in
documenting subtle changes in motor control" [52]. When comparing their task of wave tracking
(involving precision control), other studies were usually employed to quantify surgical results or
treatment progression. Three features were considered: spectral density, root mean square (RMS)
error and lag. Although some of the features correlated significantly with UPDRS scores, there
have been no significant improvements in prediction. Nonetheless, the authors suggest that their
test may add an extra objective measure that other tests fail to capture.
10
11. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
In a publication by Brewer et al. [12], a similar approach has been used to predict UPDRS scores.
Here twenty-six participants (all PD patients) were exhibiting pressure on force and torque
sensors while they were performing wave tracking tasks. The authors used the same parameters to
summarize the participants ability to properly track waves (i.e. spectral density, RMS error and
lag). These features were evaluated in terms of their ability to predict UPDRS scores. The authors
present four approaches: PCA, least squares linear regression, lasso regression and ridge
regression. Their results indicate that ridge regression works best with an absolute error of 3.5
UPDRS points. This is followed by lasso regression (i.e. 4.5 UPDRS points) and PCA (i.e. 7
UPDRS points).
Similarly, Kondraske et al. [31] utilize ordinary computer hardware for specialized PD tests. The
authors present an initial evaluation of three objective, self-administered and web-based tests (i.e.
alternating movement quality, simple visual-based response speed and upper extremity
neuromotor channel capacity). Each test has an equivalent version in the real-world based on a
testing device called "BEP 1". Twenty-one subjects (i.e. eight healthy controls and thirteen PD
patients) enrolled in their evaluation where both lab-based and web-based tests were performed.
The results indicate an encouraging well correlation by lab-based and web-based "rapid
alternating movement" and "`neuromotor channel capacity" tests. The correlation for the "simple
visual" test did not show expected results. The authors envision a three-tiered approach that first
involves digital, web-based tests then lab-based tests and finally screening by an expert. As
suggested by its nature, web-based tests are easily accessible to a broad population. They provide
objective measurements within an uncontrolled environment and may provide an initial
assessment on whether any signs of PD are apparent. The second tier can then be used for a
complementary assessment in a controllable environment. Afterwards a proper clinical screening
can be performed by a neurologist if previous results suggested parkinsonian behavior.
An automatic evaluation approach for early detection of PD is presented by Jobbágy et al. [28].
The authors propose and evaluate a set of tests that were specifically designed to highlight
features of PD symptoms. They employ a motion tracking system, called precision motion
analysis system (PRIMAS), for recording movements patterns. The system uses a combination of
infrared (IR) light, passive markers (i.e. small, lightweight reflective disks mounted on body) and
cameras in order to track the participants' movements of their fingers and hands. Jobbágy and
colleagues aim at providing tests and / or measures to indicate the presence of early to moderate
PD and subtle changes in its progression. Twenty-nine participants took part in their study (i.e.
thirteen young healthy subjects, ten elderly healthy subjects and six subjects afflicted with PD).
Three tasks were performed: tapping task, twiddling task as well as a pinching and circling task.
The authors describe their analysis of raw movement data from their tracking system and
highlight their chosen features (e.g. frequency, symmetry, dexterity, amplitude, etc.). Based on
these parameters a score (between zero and one) is proposed in which people with PD achieve
higher score-values (as in UPDRS). Their empirical results indicate that their scale does indeed
separate PD patients from healthy subjects.
3.CONCLUSIONS
It is apparent that indication of PD motor symptoms in time series data is clearly not an unwritten
page. Alone in the past decade a great number of publications with a focus on this very topic have
been seen. Some of the mentioned authors have published their work on several symptoms (e.g.
Cole et al. [15],[16], Salarian et al. [56],[57] and Zwartjes et al. [72]).
In recent years, accuracy of symptom indication and severity indication have reached percentages
well above 90%. However it should be noted that datasets vary greatly in quality and quantity
(e.g. from a few minutes to several hours or days of data). The accuracy increases and decreases
11
12. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
with the used datasets and employed algorithms. Authors with small datasets or even synthetic
datasets tend to achieve higher accuracies than those that utilize medium-sized or large datasets
from real people. Another aspect of quality is the task / activity which have been performed
during recording sessions (i.e. scripted vs. unscripted, constrained vs. unconstrained, etc.). Here,
preference has been given to those publications that were not using standardized motor tasks to
identify symptoms (and their severity). Sensitivities in the range of 90%-95% (sometimes even
greater) were achieved with today’s methods, but usually at the cost of a lower specificity.
Table 2 summarizes the papers that were presented in this paper. Despite the reservation of
highlighting publications that enable indication of PD motor symptoms and / or assessing their
severity while being mobile, a set of publications that do not fit these criteria was presented.
Table 1 points to several noteworthy publications with a similar focus, but do not necessarily
intend to identify cardinal symptoms or make use of body-mounted sensors. As a consequence,
they do not necessarily present the state-of-the-art. Nonetheless the interested reader is
encouraged to read through them.
FUTURE WORK
Despite the fact that fairly high accuracies has already been reached, the presented results still
allow for some improvements. The authors would like to employ a rather untraditional set of
algorithms for indicating the presence of the mentioned PD motor symptoms in time series data
(e.g. StreamKM++[2], ClusTree[33] and LogLog algorithm [22]). It is their intention to evaluate
whether these approaches can perform on a similar level of accuracy or maybe even outperform
the mentioned publications. A suitable framework, called MOSIS, for evaluating these
approaches is actually under development (publication is pending; download available on
mloss.org). Furthermore, the author's are likely to review more publications.
ACKNOWLEDGEMENTS
This work has been motivated and partly funded by the European Commission through AAL JP
Call 1 project HELP and ICT FP 7 project REMPARK (287677). The authors acknowledge the
support by the Commission and the HELP and REMPARK partners for their fruitful work and
contribution in the research.
Table 1. Lists several related publications. The author(s) and title of their publication are highlighted. This
list is intended to supplement state-of-the-art publications (shown in Table 2) with additional noteworthy
and relevant papers.
Author(s)
Note
Brewer et al. [12]
Application of Modified Regression Techniques to a Quantitative Assessment
for the Motor Signs of Parkinson's Disease
Cunningham et al. [18]
Identifying fine movement difficulties in Parkinson's disease using a computer
assessment tool
Cunningham et al. [17]
Home-Based Monitoring and Assessment of Parkinson's Disease
Hamilton et al. [25]
Neural networks trained with simulation data for outcome prediction in
pallidotomy for Parkinson's disease
Kondraske et al. [31]
Web-based evaluation of Parkinson's Disease subjects: Objective performance
capacity measurements and subjective characterization profiles
Wang et al. [68]
A new quantitative evaluation method of Parkinson's disease based on free
spiral drawing
Jobbágy et al. [28]
Early detection of Parkinson's disease through automatic movement evaluation
12
13. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
Table 2. Summarization of state-of-the-art publications on PD symptom indication algorithms. For each
symptom (T: tremor, B: bradykinesia, F: FOG, D: dyskinesia) and reference, the employed classification
techniques and utilized sensors (A: accelerometer, G: gyroscope, E: EMG sensor) are highlighted.
Furthermore their results are indicated. It should be kept in mind that the results among these papers are
not directly comparable due to employment of different
13
14. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
REFERENCES
[1] A. Asuncion, D.J.N.: UCI machine learning repository (2007), http://www.ics.uci.edu/~mlearn/
[2] Ackermann, M.R., Lammersen, C., Märtens, M., Raupach, C., Sohler, C., Swierkot, K.:
Streamkm++: A clustering algorithms for data streams. In: Blelloch, G.E., Halperin, D. (eds.)
ALENEX. pp. 173– 187. SIAM (2010)
[3] Andlin-Sobocki, P., Jnsson, B., Wittchen, H.U., Olesen, J.: Cost of disorders of the brain in
Europe. European Journal of Neurology 12 Suppl 1 (Jun 2005), http://dx.doi.org/10.1111/j.14681331.2005.01202.x
[4] Armstrong, R.A.: Visual signs and symptoms of parkinson's disease. Clinical and Experimental
Optometry 91(2), 129– 138 (2008), http://dx.doi.org/10.1111/j.1444-0938.2007.00211.x
[5] Arvind, R., Karthik, B., Sriraam, N., Kannan, J.K.: Automated detection of pd resting tremor using
psd with recurrent neural network classier. In: Advances in Recent Technologies in
Communication and Computing (ARTCom), 2010 International Conference on. pp. 414 – 417 (oct
2010)
[6] Asgari, M., Shafran, I.: Predicting severity of parkinson's disease from speech. In: Engineering in
Medicine and Biology Society (EMBC), 2010 Annual International Conference of the IEEE. pp.
5201 – 5204 (31 2010-sept 4 2010)
[7] B and, A.P.L., Poignet, P., Geny, C.: Pathological tremor and voluntary motion modeling and
online estimation for active compensation. Neural Systems and Rehabilitation Engineering, IEEE
Transactions on 19(2), 177 – 185 (april 2011)
[8] Bakar, Z.A., Tahir, N.M., Yassin, I.M.: Classification of parkinson's disease based on multilayer
perceptrons neural network. In: Signal Processing and Its Applications (CSPA), 2010 6th
International Colloquium on. pp. 1 – 4 (may 2010)
[9] Bakstein, E., Warwick, K., Burgess, J., Stavdahl, O., Aziz, T.: Features for detection of
parkinson's disease tremor from local field potentials of the subthalamic nucleus. In: Cybernetic
Intelligent Systems (CIS), 2010 IEEE 9th International Conference on. pp. 1 – 6 (sept 2010)
[10] Bächlin, M., Plotnik, M., Roggen, D., Maidan, I., Hausdor, J.M., Giladi, N., Troster, G.: Wearable
assistant for parkinson's disease patients with the freezing of gait symptom. Information
Technology in Biomedicine, IEEE Transactions on 14(2), 436 – 446 (march 2010)
[11] Bächlin, M., Roggen, D., Troster, G., Plotnik, M., Inbar, N., Meidan, I., Herman, T., Brozgol, M.,
Shaviv, E., Giladi, N., Hausdor, J.M.: Potentials of enhanced context awareness in wearable
assistants for parkinson's disease patients with the freezing of gait syndrome. In: Wearable
Computers, 2009. ISWC '09. International Symposium on. pp. 123 – 130 (sept 2009)
[12] Brewer, B.R., Pradhan, S., Carvell, G., Delitto, A.: Application of modied regression techniques to
a quantitative assessment for the motor signs of parkinson's disease. Neural Systems and
Rehabilitation Engineering, IEEE Transactions on 17(6), 568 – 575 (dec 2009)
[13] Burg, J.P.: Maximum Entropy Spectral Analysis. Ph.D. thesis, Stanford University (1975)
[14] Cancela, J., Pansera, M., Arredondo, M.T., Estrada, J.J., Pastorino, M., Pastor-Sanz, L., Villalar,
J.L.: A comprehensive motor symptom monitoring and management system: The bradykinesia
case. In: Engineering in Medicine and Biology Society (EMBC), 2010 Annual International
Conference of the IEEE. pp. 1008 – 1011 (31 2010-sept 4 2010)
[15] Cole, B.T., Roy, S.H., De Luca, C.J., Nawab, S.H.: Dynamic neural network detection of tremor
and dyskinesia from wearable sensor data. In: Engineering in Medicine and Biology Society
(EMBC), 2010 Annual International Conference of the IEEE. pp. 6062 – 6065 (31 2010-sept 4
2010)
[16] Cole, B.T., Roy, S.H., Nawab, S.H.: Detecting freezing-of-gait during unscripted and
unconstrained activity. In: Engineering in Medicine and Biology Society, EMBC, 2011 Annual
International Conference of the IEEE. pp. 5649 – 5652 (30 2011-sept 3 2011) 11
[17] Cunningham, L., Mason, S., Nugent, C., Moore, G., Finlay, D., Craig, D.: Homebased monitoring
and assessment of parkinson's disease. Information Technology in Biomedicine, IEEE
Transactions on 15(1), 47 – 53 (jan 2011)
[18] Cunningham, L., Nugent, C., Moore, G., Finlay, D., Craig, D.: Identifying fine movement
difficulties in parkinson's disease using a computer assessment tool. In: Information Technology
and Applications in Biomedicine, 2009. ITAB 2009. 9th International Conference on. pp. 1 – 4
(nov 2009)
[19] Cunningham, L., Nugent, C., Moore, G., Finlay, D., Craig, D.: Computer-based assessment of
bradykinesia, akinesia and rigidity in parkinsons disease. In: Mokhtari, M., Khalil, I., Bauchet, J.,
14
15. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
Zhang, D., Nugent, C. (eds.) Ambient Assistive Health and Wellness Management in the Heart of
the City, Lecture Notes in Computer Science, vol. 5597, pp. 1– 8. Springer Berlin Heidelberg
(2009), http://dx.doi.org/10.1007/978-3-642-02868-7_1
Davie, C.A.: A review of parkinson's disease. British Medical Bulletin 86(1), 109– 127 (2008),
http://bmb.oxfordjournals.org/content/86/1/109.abstract
Djurić-Jovičić, M., Jovičić, N.S., Milovanovic, I., Radovanovic, S., Kresojevic, N., Popovic,
M.B.: Classification of walking patterns in parkinson's disease patients based on inertial sensor
data. In: Neural Network Applications in Electrical Engineering (NEUREL), 2010 10th
Symposium on. pp. 3 – 6 (sept 2010)
Durand, M., Flajolet, P.: Loglog counting of large cardinalities. In: Battista, G., Zwick, U. (eds.)
Algorithms - ESA 2003, Lecture Notes in Computer Science, vol. 2832, pp. 605– 617. Springer
Berlin Heidelberg (2003), http://dx.doi.org/10.1007/978-3-540-39658-1_55
Fahn, S.: Concept and classification of dystonia. Adv Neurol 50 (1988)
Gustavsson, A., Svensson, M., Jacobi, F., Allgulander, C., Alonso, J., Beghi, E., Dodel, R.,
Ekman, M., Faravelli, C., Fratiglioni, L., Gannon, B., Jones, D.H., Jennum, P., Jordanova, A.,
Jnsson, L., Karampampa, K., Knapp, M., Kobelt, G., Kurth, T., Lieb, R., Linde, M., Ljungcrantz,
C., Maercker, A., Melin, B., Moscarelli, M., Musayev, A., Norwood, F., Preisig, M., Pugliatti, M.,
Rehm, J., Salvador-Carulla, L., Schlehofer, B., Simon, R., Steinhausen, H.C., Stovner, L.J., Vallat,
J.M., den Bergh, P.V., van Os, J., Vos, P., Xu, W., Wittchen, H.U., Jnsson, B., Olesen, J.: Cost of
disorders of the brain in europe 2010. European Neuropsychopharmacology 21(10), 718 – 779
(2011), http://www.sciencedirect.com/science/article/pii/S0924977X1100215X
Hamilton, J.L., Micheli-Tzanakou, E., Lehman, R.M.: Neural networks trained with simulation
data for outcome prediction in pallidotomy for parkinson's disease. In: Engineering in Medicine
and Biology Society, 2000. Proceedings of the 22nd Annual International Conference of the IEEE.
vol. 1, pp. 1 – 4 vol.1 (2000)
Hou, J.G.G., Lai, E.C.: Non-motor symptoms of parkinson's disease. International Journal of
Gerontology
1(2),
53
–
64
(2007),
http://www.sciencedirect.com/science/article/pii/S1873959808700243
Jankovic, J.: Parkinson's disease: clinical features and diagnosis. Journal of Neurology,
Neurosurgery & Psychiatry 79(4), 368– 376 (2008), http://jnnp.bmj.com/content/79/4/368.abstract
Jobbágy, A., Furnee, E., Harcos, P., Tarczy, M.: Early detection of parkinson's disease through
automatic movement evaluation. Engineering in Medicine and Biology Magazine, IEEE 17(2), 81
– 88 (mar/apr 1998)
Jovanov, E., Wang, E., Verhagen, L., Fredrickson, M., Fratangelo, R.: defog—a real time system
for detection and unfreezing of gait of parkinson's patients. In: Engineering in Medicine and
Biology Society, 2009. EMBC 2009. Annual International Conference of the IEEE. pp. 5151 –
5154 (sept 2009) 12
Keijsers, N.L.W., Horstink, M.W.I.M., Gielen, S.C.A.M.: Automatic assessment of levodopainduced dyskinesias in daily life by neural networks. Movement Disorders 18(1), 70– 80 (2003),
http://dx.doi.org/10.1002/mds.10310
Kondraske, G.V., Stewart, R.M.: Web-based evaluation of parkinson's disease subjects: Objective
performance capacity measurements and subjective characterization profiles. In: Engineering in
Medicine and Biology Society, 2008. EMBS 2008. 30th Annual International Conference of the
IEEE. pp. 799 – 802 (aug 2008)
Korczyn, A.D.: Parkinson's disease. In: in Chief: Kris Heggenhougen, E. (ed.) International
Encyclopedia of Public Health, pp. 10 –
17. Academic Press, Oxford (2008),
http://www.sciencedirect.com/science/article/pii/B9780123739605000289
Kranen, P., Assent, I., Baldauf, C., Seidl, T.: Self-adaptive anytime stream clustering. Data
Mining, IEEE International Conference on 0, 249– 258 (2009)
Krenz, A.: The pathological role of synphilin-1 and the therapeutic potential of Hsp70 in models
of Parkinson's disease using viral vectors. PhD. thesis, Universität Tübingen, Wilhelmstr. 32,
72074 Tbingen (2010), http://tobias-lib.uni-tuebingen.de/volltexte/2010/4620
Kupryjanow, A., Kunka, B., Kostek, B.: Updrs tests for diagnosis of parkinson's disease
employing virtual-touchpad. In: Database and Expert Systems Applications (DEXA), 2010
Workshop on. pp. 132 – 136 (30 2010-sept 3 2010)
Little, M.A., McSharry, P.E., Hunter, E.J., Spielman, J., Ramig, L.O.: Suitability of dysphonia
measurements for telemonitoring of parkinson's disease. Biomedical Engineering, IEEE
Transactions on 56(4), 1015 – 1022 (april 2009)
15
16. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
[37] Liu, X., Carroll, C.B., Wang, S.Y., Zajicek, J., Bain, P.G.: Quantifying druginduced dyskinesias in
the arms using digitised spiral-drawing tasks. Journal of Neuroscience Methods 144(1), 47 – 52
(2005), http://www.sciencedirect.com/science/article/pii/S0165027004003723
[38] Longsta, M.G., Mahant, P.R., Stacy, M.A., Van Gemmert, A.W.A., Leis, B.C., Stelmach, G.E.:
Discrete and dynamic scaling of the size of continuous graphic movements of parkinsonian
patients and elderly controls. Journal of Neurology, Neurosurgery & Psychiatry 74(3), 299– 304
(2003), http://jnnp.bmj.com/content/74/3/299.abstract
[39] Marchis, C.D., Schmid, M., Conforto, S.: An optimized method for tremor detection and temporal
tracking through repeated second order moment calculations on the surface emg signal. Medical
Engineering
&
Physics
34(9),
1268
–
1277
(2012),
http://www.sciencedirect.com/science/article/pii/S1350453311003444
[40] Mathers, C., Fat, D.M., Boerma, J.T., Organization, W.H.: The global burden of disease : 2004
update.
World
Health
Organization,
Geneva,
Switzerland:
(2008),
http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf
[41] Mekyska, J., Rektorova, I., Smekal, Z.: Selection of optimal parameters for automatic analysis of
speech disorders in parkinson's disease. In: Telecommunications and Signal Processing (TSP),
2011 34th International Conference on. pp. 408 – 412 (aug 2011)
[42] Miralles, F., Tarongi, S., Espino, A.: Quantification of the drawing of an archimedes spiral
through the analysis of its digitized picture. Journal of Neuroscience Methods 152(1 - 2), 18 – 31
(2006), http://www.sciencedirect.com/science/article/pii/S0165027005002955
[43] Myers, L.J., MacKinnon, C.D.: Quanti cation of movement regularity during internally generated
and externally cued repetitive movements in patients with parkinson's disease. In: Neural
Engineering, 2005. Conference Proceedings. 2nd International IEEE EMBS Conference on. pp.
281 – 284 (march 2005)
[44] Niazmand, K., Kalaras, A., Dai, H., Lueth, T.C.: Comparison of methods for tremor frequency
analysis for patients with parkinson's disease. In: Biomedical Engineering and Informatics
(BMEI), 2011 4th International Conference on. Vol. 2, pp. 693 – 697 (oct 2011)
[45] Niazmand, K., Tonn, K., Kalaras, A., Fietzek, U.M., Mehrkens, J.H., Lueth, T.C.: Quantitative
evaluation of parkinson's disease using sensor based smart glove. In: Computer-Based Medical
Systems (CBMS), 2011 24th International Symposium on. pp. 1 – 8 (june 2011)
[46] Niazmand, K., Tonn, K., Kalaras, A., Kammermeier, S., Boetzel, K., Mehrkens, J.H., Lueth, T.C.:
A measurement device for motion analysis of patients with parkinson's disease using sensor based
smart clothes. In: Pervasive Computing Technologies for Healthcare (PervasiveHealth), 2011 5th
International Conference on. pp. 9 – 16 (may 2011)
[47] Niazmand, K., Tonn, K., Zhao, Y., Fietzek, U.M., Schroeteler, F., Ziegler, K., Ceballos-Baumann,
A.O., Lueth, T.C.: Freezing of gait detection in parkinson's disease using accelerometer based
smart clothes. In: Biomedical Circuits and Systems Conference (BioCAS), 2011 IEEE. pp. 201 –
204 (nov 2011)
[48] Niazmand, K., Somlai, I., Louizi, S., Lueth, T.C.: Proof of the accuracy of measuring pants to
evaluate the activity of the hip and legs in everyday life. In: Lin, J.C., Nikita, K.S., Akan, O.,
Bellavista, P., Cao, J., Dressler, F., Ferrari, D., Gerla, M., Kobayashi, H., Palazzo, S., Sahni, S.,
Shen, X.S., Stan, M., Xiaohua, J., Zomaya, A., Coulson, G. (eds.) Wireless Mobile
Communication and Healthcare, Lecture Notes of the Institute for Computer Sciences, Social
Informatics and Telecommunications Engineering, vol. 55, pp. 235– 244. Springer Berlin
Heidelberg (2011), http://dx.doi.org/10.1007/978-3-642-20865-2_30,10.1007/978-3-642-20865-2
30
[49] Pastorino, M., Cancela, J., Arredondo, M.T., Pansera, M., Pastor-Sanz, L., Villagra, F., Pastor,
M.A., Martn, J.A.: Assessment of bradykinesia in parkinson's disease patients through a multiparametric system. In: Engineering in Medicine and Biology Society, EMBC, 2011 Annual
International Conference of the IEEE. pp. 1810 – 1813 (30 2011-sept 3 2011)
[50] Patel, S., Lorincz, K., Hughes, R., Huggins, N., Growdon, J., Standaert, D., Akay, M., Dy, J.,
Welsh, M., Bonato, P.: Monitoring motor fluctuations in patients with parkinson's disease using
wearable sensors. Information Technology in Biomedicine, IEEE Transactions on 13(6), 864 –
873 (nov 2009)
[51] Patel, S., Sherrill, D., Hughes, R., Hester, T., Huggins, N., Lie-Nemeth, T., Standaert, D., Bonato,
P.: Analysis of the severity of dyskinesia in patients with parkinson's disease via wearable sensors.
In: Wearable and Implantable Body Sensor Networks, 2006. BSN 2006. International Workshop
on. pp. 4 pp. – 126 (april 2006)
16
17. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
[52] Pradhan, S.D., Brewer, B.R., Carvell, G.E., Sparto, P.J., Delitto, A., Matsuoka, Y.: Relation
between ability to track force during dual tasking and function in individuals with parkinson's
disease. In: Rehabilitation Robotics, 2009. ICORR 2009. IEEE International Conference on. pp.
885 – 892 (june 2009)
[53] Revett, K., Gorunescu, F., Salem, A.B.M.: Feature selection in parkinson's disease: A rough sets
approach. In: Computer Science and Information Technology, 2009. IMCSIT '09. International
Multiconference on. pp. 425 – 428 (oct 2009)
[54] Rigas, G., Tzallas, A., Tsipouras, M., Bougia, P., Tripoliti, E., Baga, D., Fotiadis, D., Tsouli, S.,
Konitsiotis, S.: Assessment of tremor activity in the parkinson's disease using a set of wearable
sensors. Information Technology in Biomedicine, IEEE Transactions on PP(99), 1 (2012)
[55] Roy, S.H., Cole, B.T., Gilmore, L.D., De Luca, C.J., Nawab, S.H.: Resolving signal complexities
for ambulatory monitoring of motor function in parkinson's disease. In: Engineering in Medicine
and Biology Society, EMBC, 2011 Annual International Conference of the IEEE. pp. 4832 – 4835
(30 2011-sept 3 2011)
[56] Salarian, A., Russmann, H., Vingerhoets, F.J.G., Burkhard, P.R., Blanc, Y., Dehollain, C.,
Aminian, K.: An ambulatory system to quantify bradykinesia and tremor in parkinson's disease.
In: Information Technology Applications in Biomedicine, 2003. 4th International IEEE EMBS
Special Topic Conference on. pp. 35 – 38 (april 2003)
[57] Salarian, A., Russmann, H., Wider, C., Burkhard, P.R., Vingerhoets, F.J.G., Aminian, K.:
Quantification of tremor and bradykinesia in parkinson's disease using a novel ambulatory
monitoring system. Biomedical Engineering, IEEE Transactions on 54(2), 313 – 322 (feb 2007)
[58] Samii, A., Nutt, J.G., Ransom, B.R.: Parkinson's disease. The Lancet 363(9423), 1783 – 1793
(2004), http://www.sciencedirect.com/science/article/pii/S0140673604163058
[59] Sherrill, D.M., Hughes, R., Salles, S.S., Lie-Nemeth, T., Akay, M., Standaert, D.G., Bonato, P.:
Advanced analysis of wearable sensor data to adjust medication intake in patients with parkinson's
disease. In: Neural Engineering, 2005. Conference Proceedings. 2nd International IEEE EMBS
Conference on. pp. 202 – 205 (march 2005)
[60] Sian, J., Gerlach, M., Youdim, M.B.H., Riederer, P.: Parkinson's disease: a major hypokinetic
basal ganglia disorder. Journal of Neural Transmission 106, 443– 476 (1999),
http://dx.doi.org/10.1007/s007020050171, 10.1007/s007020050171
[61] Smith, S.L., Shannon, K.: Vector-based analysis of motor activities in patients with parkinson's
disease. In: EUROMICRO 97. 'New Frontiers of Information Technology'. Short Contributions.,
Proceedings of the 23rd Euromicro Conference. pp. 50 – 55 (sep 1997)
[62] Stamatakis, J., Crmers, J., Maquet, D., Macq, B., Garraux, G.: Gait feature extraction in
parkinson's disease using low-cost accelerometers. In: Engineering in Medicine and Biology
Society, EMBC, 2011 Annual International Conference of the IEEE. pp. 7900 – 7903 (30 2011sept 3 2011)
[63] Su, Y., Allen, C.R., Geng, D., Burn, D., Brechany, U., Bell, G.D., Rowland, R.: 3-d motion system
("data-gloves"): application for parkinson's disease. Instrumentation and Measurement, IEEE
Transactions on 52(3), 662 – 674 (june 2003)
[64] Sugiura, T., Sugiura, N., Sugiyama, K., Yokoyama, T.: Chaotic approach to the quantitative
analysis of parkinson's disease. In: Engineering in Medicine and Biology Society, 1998.
Proceedings of the 20th Annual International Conference of the IEEE. vol. 3, pp. 1583 – 1586
vol.3 (oct-1 nov 1998)
[65] Tsipouras, M.G., Tzallas, A.T., Rigas, G., Bougia, P., Fotiadis, D.I., Konitsiotis, S.: Automated
levodopa-induced dyskinesia assessment. In: Engineering in Medicine and Biology Society
(EMBC), 2010 Annual International Conference of the IEEE. pp. 2411 – 2414 (31 2010-sept 4
2010)
[66] Šprdlík, O., Hurak, Z., Hoskovcova, M., Ru zi cka, E.: Tremor analysis by decomposition of
acceleration into gravity and inertial acceleration using inertial measurement unit. In: Information
Technology and Applications in Biomedicine, 2009. ITAB 2009. 9th International Conference on.
pp. 1 – 4 (nov 2009)
[67] Wan, E.A.: Discrete time neural networks. Applied Intelligence 3, 91– 105 (1993),
http://dx.doi.org/10.1007/BF00871724, 10.1007/BF00871724
[68] Wang, M., Wang, B., Zou, J., Chen, L., Shima, F., Nakamura, M.: A new quantitative evaluation
method of parkinson's disease based on free spiral drawing. In: Biomedical Engineering and
Informatics (BMEI), 2010 3rd International Conference on. vol. 2, pp. 694 – 698 (oct 2010)
ᄉ ᄉ
17
18. Health Informatics- An International Journal (HIIJ) Vol.2,No.4,November 2013
[69] Watson, C., Kirkcaldie, M., Paxinos, G.: The brain: an introduction to functional neuroanatomy.
Elsevier/Academic, Amsterdam, 1st ed edn. (2010)
[70] Xiuming, C., Jinjie, S., Caipo, Z.: Diagnose model of parkinson's disease based on principal
component analysis and sugeno integral. In: Control Conference (CCC),2011 30th Chinese. pp.
2830 – 2834 (july 2011)
[71] Ziegler, K., Schroeteler, F., Ceballos-Baumann, A.O., Fietzek, U.M.: A new rating instrument to
assess festination and freezing gait in parkinsonian patients. Movement Disorders 25(8), 1012–
1018 (2010), http://dx.doi.org/10.1002/mds.22993
[72] Zwartjes, D.G.M., Heida, T., van Vugt, J.P.P., Geelen, J.A.G., Veltink, P.H.: Ambulatory
monitoring of activities and motor symptoms in parkinson's disease. Biomedical Engineering,
IEEE Transactions on 57(11), 2778 – 2786 (nov 2010)
Authors
Claas Ahlrichs is a PhD. candidate at Universitaet Bremen, where he studied computer
science with a focus on artificial intelligence and wearable computing. He graduated with his
thesis “Development and Evaluation of an Abstract User Interface for Performing Maintenance
Scenarios with Wearable Computers” in 2011. Since 2008, Ahlrichs is involved at the Center
for Computing and Communication Technologies (TZI) Technologies in the field of wearable
computing. Furthermore, he has been involved in several regional and international research projects (ITASSIT, HELP, REMPARK) related to health care, human computer interaction and wearable computing.
Currently, he works as a software developer at neusta mobile solutions GmbH in Bremen.
Prof. Dr. Michael Lawo is since 2004 at TZI (www.tzi.de) of the Universitaet Bremen, and
since 2009 one of the two Managing Directors of neusta mobile solutions GmbH. He is
professor for applied computer science at Universitaet Bremen, member of the steering board
of Logdynamics (www.logdynamics.de) and involved in numerous projects of logistics,
wearable computing and artificial intelligence. He had been the CEO of a group of SME in the
IT domain since 1999 with a focus on the development and marketing of virtual reality simulators for
surgeons; from 1996 to 2000 he was CEO of an IT consulting firm and from 1991 to 1995 top manager
information systems with the Bremer Vulkan group. Michael Lawo was consultant before joining the
nuclear research centre in Karlsruhe from 1987 to 1991 as head of the industrial robotics department. He is
a 1975 graduate of structural engineering of Ruhr Universität Bochum, received his PhD from Universität
Essen in 1981 and became professor in structural optimisation there in 1992. In 2000 he was appointed as
professor of honour of the Harbin/China College of Administration & Management. He is author, co-author
and co-publisher of eight books and more than 150 scientific papers on numerical methods and computer
applications also in logistics, healthcare, optimization, IT-security, sensorial materials and wearable
computing.
18