Parkinson's disease motor symptoms in machine learning a reviewhiij
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.
This Powerpoint prsentation contains information about the overview of various successful works performed for Biometric Recognition using Deep Learning. This work is based on an existing survey paper.
Parkinson's disease motor symptoms in machine learning a reviewhiij
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.
This Powerpoint prsentation contains information about the overview of various successful works performed for Biometric Recognition using Deep Learning. This work is based on an existing survey paper.
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.
Disease prediction using machine learningJinishaKG
Github link :
https://github.com/jini-the-coder/Diseaseprediction
Blog link :
http://amigoscreation.blogspot.com/2020/07/disease-prediction-using-machine.html
Youtube link :
https://youtu.be/3YmAbta16yk
Predictive Analytics and Machine Learning for Healthcare - DiabetesDr Purnendu Sekhar Das
Machine Learning on clinical datasets to predict the risk of chronic disease conditions like Type 2 Diabetes mellitus beforehand; as well as predicting outcomes like hospital readmission using EMR RWE data.
Dr. Dina Zamzam presentation about Multiple Sclerosis.nourhan mohsen
What is Multiple sclerosis?
مرض التصلب المتعدد في عشرة اسئلة
طبيعة المرض وكيفية التعايش معه
Ms symptoms
in which age does the multiple sclerosis attack
Ms causes
Ms treatments
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.
Disease prediction using machine learningJinishaKG
Github link :
https://github.com/jini-the-coder/Diseaseprediction
Blog link :
http://amigoscreation.blogspot.com/2020/07/disease-prediction-using-machine.html
Youtube link :
https://youtu.be/3YmAbta16yk
Predictive Analytics and Machine Learning for Healthcare - DiabetesDr Purnendu Sekhar Das
Machine Learning on clinical datasets to predict the risk of chronic disease conditions like Type 2 Diabetes mellitus beforehand; as well as predicting outcomes like hospital readmission using EMR RWE data.
Dr. Dina Zamzam presentation about Multiple Sclerosis.nourhan mohsen
What is Multiple sclerosis?
مرض التصلب المتعدد في عشرة اسئلة
طبيعة المرض وكيفية التعايش معه
Ms symptoms
in which age does the multiple sclerosis attack
Ms causes
Ms treatments
Sepsis is one of the top causes of inpatient mortality and rapid detection presents numerous challenges. In March, 2016, an interdisciplinary team consisting of top clinicians, data scientists and machine learning experts at a large academic medical center (AMC) embarked on an innovation pilot to develop a novel machine learning model to detect sepsis. A computable sepsis definition and deep learning model were developed using a curated dataset capturing over 43,000 inpatient admissions between October 1, 2014 and December 31, 2015. Ten computable sepsis definitions were compared and our clinicians agreed on the following: >= 2 SIRS criteria, blood culture order, and end organ damage. This sepsis phenotype identified patients early in the hospital course: 38% of cases occur an average of 1.3 hours after presentation to the ED and 42% of cases occur an average of 15 hours after hospital admission. At 4 hours prior to sepsis, the best deep learning model generated 1.4 false alarms per true alarm at a sensitivity of 80%, compared to 3.2 false alarms per true alarm for National Early Warning System (NEWS).
Purpose
Sepsis Watch detects sepsis early, guides completion of appropriate treatment, and supports front-line providers with minimal interruption of clinical workflows. Key Performance Indicators include emergency department (ED) length of stay, hospital length of stay, inpatient mortality, intensive care unit requirement, and time to antibiotics for patients who develop sepsis.
Description
The core technology components of Sepsis Watch are web services to extract electronic health record (EHR) data in real-time, a data pipeline to normalize features, a computable sepsis definition, a deep learning sepsis prediction model, a web application (Figure 1), an automated report that calculates KPI performance, and a model input and output monitoring tool. A suite of education, training, communication, and workflow materials were also prepared with nurse educators and are hosted on an intranet training site. After a three-month silent period, Sepsis Watch was deployed in the ED of the 1,000 bed flagship hospital on November 5, 2018.
Conclusions
Sepsis Watch is the first deployment of deep learning model in real-time to detect sepsis integrated with an EHR. The tool is used by Rapid Response Team (RRT) nurses to provide proactive support to ED providers to identify and manage sepsis. A six-month clinical trial will be completed in May 2019 to rigorously assess the clinical and operational impact of the program.
A Study On Clinical Profile Of Sepsis Patients In Intensive Care Unit Of A Te...dbpublications
Background : Sepsis is life-threatening organ dysfunction caused by a dysregulated host response to infection which is one of the most important cause of mortality & morbidity in critically ill patients. In this study clinical profiles of the sepsis patients admitted in ICU in this part of India have been evaluated. Methods & Materials: This prospective hospital based observational study was undertaken in the department of Emergency Medicine ICU of Gauhati Medical College & Hospital, over a period of one year from August 2014 to July 2015 after obtaining institutional ethical committee clearance.
RESULTS: Clinical profiles of 50sepsis patients, with male preponderance (56%) & mortality rate 36% were studied. Mean age was 48.36 years (SD ±17.16). fever & tachycardia were present in all patients. 30 patients (60%) required ventilatory support, 28 patients (56%) required inotropic support, 10 patients (20%) required dialysis. Gram negative bacteria were found to be the predominant pathogens associated with sepsis(73.4%) where most common organism responsible was Klebsiella (36.8%). Conclusion : assessment of clinical signs & initial serological & radiological investigations are of utmost importance to detect more critically ill patients as early as possible to intervene earlier for saving the life of the sepsis patients.
Thank You for referencing this work, if you find it useful!
Citation of a related scientific paper:
Manea, V., Wac, K., (2018). mQoL: Mobile Quality of Life Lab: From Behavior Change to QoL, Mobile Human Contributions: Opportunities and Challenges (MHC) Workshop in conjunction with ACM UBICOMP, Singapore, October 2018.
Katarzyna Wac, From Quantified Self to Quality of Life, Book Chapter in "Digital Health", Health Informatics, Springer Nature, p. 83-108, Dordrecht, The Netherlands, 2018.
The talk details:
Katarzyna Wac, “Quality of Life Technologies: From Cure to Care”, Société Suisse des Pharmaciens Hospitaliers (GSASA), November 2018, Switzerland
class GERONTOLOGICAL NURSINGJournal Article Summary AssignmentT.pdflanuszickefoosebr429
class GERONTOLOGICAL NURSING
Journal Article Summary Assignment:
The purpose of the journal article summary assignment is for students to improve their
knowledge of evidence-based geriatric nursing practice and evidence-based protocols.
Students will review evidence-based literature and reflect on how the literature impacts
their professional nursing practice. Students will summarize two articles published in a
peer-reviewed journal within the last 10 years. The journal articles must address the
geriatric population. Topics will be chosen from the provided list (unless prior approval is
given), and the topic may not be repeated on the two journal article summaries. The
student should summarize each article and discuss how the findings are significant to
clinical practice. Article summaries should use APA format (double spaced, but no cover
page) and should be no more than 3 pages. The two article summaries are 10% of the total
class grade (2 x 5%).
Journal summaries should use the following format:
• Purpose: Describe the purpose of the article/study.
• Strength of Evidence: Identify the type of evidence used to support the findings, and
the strength of the evidence. If the article is based on research, describe the study
design, setting, subjects, and sample size.
• Results: Summarize the findings of the study.
• Limitations: Identify study limitations that may weaken evidence or limit
generalizability.
• Significance: Describe how the findings are significant to geriatric nursing practice.
Do the findings represent a change in practice and how do the findings inform your
nursing practice (what did you learn)?
Topics for journal article summaries and class presentations choose one of those topic and APA
styles
• Pain
• Heart Failure
• Stroke
• Substance/Alcohol Abuse
• Urinary Incontinence
• Sexuality Issues
• Frailty/Fall Risk
• Iatrogenesis
• Sleep Disturbances
• Nutrition
• Hydration
• Orthostatic Hypotension
• Dementia
• Vision
• Hearing
• Polypharmacy
• Cultural Considerations
• Elder Communities
Solution
Ques-1: Purpose:
The purpose of the article is to examine the evidence-based practice of geriatric patients who
have been suffering from “urinary tract infection induced- urinary incontinence”. So that
catheterization of urinary tract have reduce urinary incontinence in geriatric patients thereby it is
essential implement to procedures to reduce urinary infection induced incontinence using
catheters for overflow incontinence
Ques-2:
Catheterization regimen:
\"RCT\" (simple randomized control design) and randomized trial: These two methods used
synonymously. However, it has illustrated that RCT is pertaining to trail design that include
control groups. In this design, patient groups who are receiving experimental treatment compared
with control groups (placebo groups).
In the above design it has clearly can be seen a randomized RCT was performed in two intensive
units of respiratory care of total 2990 bedded tertiary referral medical ce.
Journal Club - Discussion of Heriot et al. Criteria for identifying patients ...Salpy Kelian
Discussing, using many Bobs, how a monte carlo simulation works for a Journal Club paper regarding the modality used for detection of infectious endocarditis.
Making health data work for Patients and PopulationsBedirhan Ustun
This presentation contains the slide deck that Professor Iain Buchan - of Manchester University delivered in Koc University about what can be achieved if clinical health information were captured in a digital format.
Prospective Study of Acute Appendicitis with its Clinical, Radiological Profi...semualkaira
Acute appendicitis is the most common condition encountered in general surgical practice. Alvarado and Modified Alvarado Scores (MASS) are the commonly used scoring
systems for its diagnosis, but its performance has been found to
be poor in certain populations. Hence, we compared the RIPASA
score with MASS, to find out which is a better diagnostic tool for
acute appendicitis in the Indian population.
Austin Aging Research is an open access, peer reviewed, scholarly journal dedicated to publish articles covering all areas of Aging Research.
The journal aims to promote research communications and provide a forum for doctors, researchers, physicians and healthcare professionals to find most recent advances in all areas of Aging Research. Austin Aging Research accepts original research articles, reviews, mini reviews, case reports and rapid communication covering all aspects of Aging Research.
Austin Aging Research strongly supports the scientific up gradation and fortification in related scientific research community by enhancing access to peer reviewed scientific literary works. Austin Publishing Group brings universally peer reviewed journals under one roof thereby promoting knowledge sharing, mutual promotion of multidisciplinary science.
Slides from my talk at
Machine Learning Meet-up 2017, Medellin,
Demos available at
https://github.com/MLmeetup/Memories/tree/master/Abril/ML_cognitiveStates
We have compiled the most important slides from each speaker's presentation. This year’s compilation, available for free, captures the key insights and contributions shared during the DfMAy 2024 conference.
Low power architecture of logic gates using adiabatic techniquesnooriasukmaningtyas
The growing significance of portable systems to limit power consumption in ultra-large-scale-integration chips of very high density, has recently led to rapid and inventive progresses in low-power design. The most effective technique is adiabatic logic circuit design in energy-efficient hardware. This paper presents two adiabatic approaches for the design of low power circuits, modified positive feedback adiabatic logic (modified PFAL) and the other is direct current diode based positive feedback adiabatic logic (DC-DB PFAL). Logic gates are the preliminary components in any digital circuit design. By improving the performance of basic gates, one can improvise the whole system performance. In this paper proposed circuit design of the low power architecture of OR/NOR, AND/NAND, and XOR/XNOR gates are presented using the said approaches and their results are analyzed for powerdissipation, delay, power-delay-product and rise time and compared with the other adiabatic techniques along with the conventional complementary metal oxide semiconductor (CMOS) designs reported in the literature. It has been found that the designs with DC-DB PFAL technique outperform with the percentage improvement of 65% for NOR gate and 7% for NAND gate and 34% for XNOR gate over the modified PFAL techniques at 10 MHz respectively.
KuberTENes Birthday Bash Guadalajara - K8sGPT first impressionsVictor Morales
K8sGPT is a tool that analyzes and diagnoses Kubernetes clusters. This presentation was used to share the requirements and dependencies to deploy K8sGPT in a local environment.
Online aptitude test management system project report.pdfKamal Acharya
The purpose of on-line aptitude test system is to take online test in an efficient manner and no time wasting for checking the paper. The main objective of on-line aptitude test system is to efficiently evaluate the candidate thoroughly through a fully automated system that not only saves lot of time but also gives fast results. For students they give papers according to their convenience and time and there is no need of using extra thing like paper, pen etc. This can be used in educational institutions as well as in corporate world. Can be used anywhere any time as it is a web based application (user Location doesn’t matter). No restriction that examiner has to be present when the candidate takes the test.
Every time when lecturers/professors need to conduct examinations they have to sit down think about the questions and then create a whole new set of questions for each and every exam. In some cases the professor may want to give an open book online exam that is the student can take the exam any time anywhere, but the student might have to answer the questions in a limited time period. The professor may want to change the sequence of questions for every student. The problem that a student has is whenever a date for the exam is declared the student has to take it and there is no way he can take it at some other time. This project will create an interface for the examiner to create and store questions in a repository. It will also create an interface for the student to take examinations at his convenience and the questions and/or exams may be timed. Thereby creating an application which can be used by examiners and examinee’s simultaneously.
Examination System is very useful for Teachers/Professors. As in the teaching profession, you are responsible for writing question papers. In the conventional method, you write the question paper on paper, keep question papers separate from answers and all this information you have to keep in a locker to avoid unauthorized access. Using the Examination System you can create a question paper and everything will be written to a single exam file in encrypted format. You can set the General and Administrator password to avoid unauthorized access to your question paper. Every time you start the examination, the program shuffles all the questions and selects them randomly from the database, which reduces the chances of memorizing the questions.
Multimodal Assessment of Parkinson’s Disease: A Deep Learning Approach
1. Multimodal Assessment of Parkinson’s Disease:
A Deep Learning Approach
Camilo Vásquez
Pattern Recognition Lab, Friedrich-Alexander University of Erlangen-Nürnberg
July 18, 2018
2. Progress for summer PRS-2018
• Lab member since 01.05.2018
• First PRS
• Publications (subset)
• Journal papers
• J. C. Vasquez-Correa, J. R. Orozco-Arroyave, T. Bocklet, E. Nöth. «Towards an Automatic Eval-
uation of the Dysarthria Level of Patients with Parkinson’s Disease». Journal of Communication
Disorders (to appear), 2018.
• J. C. Vasquez-Correa, T. Arias-Vergara, J. R. Orozco-Arroyave, B. Eskofier, J. Klucken, and E.
Nöth. « Multimodal assessment of Parkinson’s disease: a deep learning approach » IEEE Journal
of Biomedical and Health Informatics (to appear), 2018.
• T. Arias-Vergara, J. C. Vasquez-Correa, J. R. Orozco-Arroyave, and E. Nöth. «Speaker models
for monitoring Parkinson’s disease progression considering different communication channels and
acoustic conditions» Speech Communication, Vol 101, pp 11-25, 2018.
• Conference proceedings
• J. C. Vasquez-Correa, T. Arias-Vergara, J. R. Orozco-Arroyave, and E. Nöth. «A Multitask Learning
Approach to Assess the Dysarthria Severity in Patients with Parkinson’s Disease» Proceedings of
INTERSPEECH, 2018.
• N. Garcia-Ospina, J. C. Vásquez-Correa, J. R. Orozco-Arroyave, and E. Nöth. «Multimodal i-
vectors to Detect and Evaluate Parkinson’s Disease» Proceedings of INTERSPEECH, 2018.
• T. Arias-Vergara, J. C. Vásquez-Correa, J. R. Orozco-Arroyave, P. Klumpp, and E. Nöth.«Unontrusive
Monitoring of Speech Impairments of Parkinson’s Disease Patients Through Mobile Devices» Pro-
ceedings of ICASSP, 2018.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 0
3. Progress for summer PRS-2018
• Lab member since 01.05.2018
• First PRS
• Publications (subset)
• Journal papers
• J. C. Vasquez-Correa, J. R. Orozco-Arroyave, T. Bocklet, E. Nöth. «Towards an Automatic Eval-
uation of the Dysarthria Level of Patients with Parkinson’s Disease». Journal of Communication
Disorders (to appear), 2018.
• J. C. Vasquez-Correa, T. Arias-Vergara, J. R. Orozco-Arroyave, B. Eskofier, J. Klucken, and E.
Nöth. « Multimodal assessment of Parkinson’s disease: a deep learning approach » IEEE Journal
of Biomedical and Health Informatics (to appear), 2018.
• T. Arias-Vergara, J. C. Vasquez-Correa, J. R. Orozco-Arroyave, and E. Nöth. «Speaker models
for monitoring Parkinson’s disease progression considering different communication channels and
acoustic conditions» Speech Communication, Vol 101, pp 11-25, 2018.
• Conference proceedings
• J. C. Vasquez-Correa, T. Arias-Vergara, J. R. Orozco-Arroyave, and E. Nöth. «A Multitask Learning
Approach to Assess the Dysarthria Severity in Patients with Parkinson’s Disease» Proceedings of
INTERSPEECH, 2018.
• N. Garcia-Ospina, J. C. Vásquez-Correa, J. R. Orozco-Arroyave, and E. Nöth. «Multimodal i-
vectors to Detect and Evaluate Parkinson’s Disease» Proceedings of INTERSPEECH, 2018.
• T. Arias-Vergara, J. C. Vásquez-Correa, J. R. Orozco-Arroyave, P. Klumpp, and E. Nöth.«Unontrusive
Monitoring of Speech Impairments of Parkinson’s Disease Patients Through Mobile Devices» Pro-
ceedings of ICASSP, 2018.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 0
5. Introduction: Parkinson’s Disease
• Second most prevalent neuro-
logical disorder worldwide.
• Patients develop several mo-
tor and non-motor impairments.
(Hornykiewicz 1998).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 1
6. Introduction: Parkinson’s Disease
• Second most prevalent neuro-
logical disorder worldwide.
• Patients develop several mo-
tor and non-motor impairments.
(Hornykiewicz 1998).
Motor impairments
• Bradykinesia
• Rigidity
• Resting tremor
• Micrographia
• Dysartrhia
Evaluated by neurologist experts
according to the MDS-UPDRS-III
scale (Goetz et al. 2008).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 1
7. Introduction: Justification
• If the motor evaluation is performed with bio–signals such as speech, hand-
writing and gait, the treatment could be followed in a more objective way
Hypothesis
PD patients have difficulties to begin and to stop the movement of limbs, and such
difficulties can be observed on signals by modeling the transitions between stop
and start movements1
1
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 2
9. Methods: Transitions detection – Speech
Onset transition Offset transition
• Onset and offset are detected according to the presence of the fundamental
frequency of the speech.
• The border is detected, and 80 ms are taken to the left and to the right, forming
chunks of 160 ms.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 3
10. Methods: Transitions detection – Speech
25 50 75 100 125 150
Time (ms)
0
500
1000
1500
2000
2500
3000
3500
4000
Frequency(Hz)
25 50 75 100 125 150
Time (ms)
6
4
2
0
2
4
Energy(dB)
STFT for the speech onset for a PD patient (left) and a HC speaker (right) when
they pronounce the syllable /ka/2
.
2
Vásquez-Correa, J. C., Orozco-Arroyave, J. R. and Nöth, E. (2017). “Convolutional Neural Network to Model Articulation Impairments in Patients with
Parkinson’s Disease”. In: 18th International Conference of the Speech and Communication Association (INTERSPEECH).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 4
11. Methods: Transitions detection – Handwriting
Tablet captures six signals: horizontal position, vertical position, pressure of the
pen, in air movement, azimuth angle, altitude angle3
.
3
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 5
12. Methods: Transitions detection – Handwriting
A. B. C. D.
0.0 0.5 1.0
Time (s)
0
500
1000
1500
Amplitude
0.0 0.5 1.0
Time (s)
0.0 0.5 1.0
Time (s)
0.0 0.5 1.0
Time (s)
x y z azimuth altitude pressure
Handwriting onset produced by: A) HC subject; B) PD patient in low state; C) PD
patient in intermediate state; and D) PD patient in severe state.4
4
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 6
13. Methods: Transitions detection – Gait
• Gait signals were captured with the eGaIT system5
• Three-axis accelerometer and gyroscope attached to the lateral heel of the
shoe.
5
Embedded Gait analysis using Intelligent Technology, http://www.egait.de/
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 7
14. Methods: Transitions detection – Gait
• Onset and offset are segmented based on the presence of the fundamental
frequency and an energy threshold.
• The border is detected, and 3 s are taken to the left and to the right.
2
1
0
1
2
3
acceleration(g)
0 1 2 3 4 5 6
Time (s)
10
20
30
40
50
Frequency(Hz)
0 1 2 3 4 5 6
Time (s)
12
10
8
6
4
2
Energy(dB)
Gait onset for a PD patient (left) and a HC subject (right)
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 8
15. Methods: Modeling Transitions with CNNs
Fully connected
layerConvolutional layer 1 Conv. layer 2Pooling layer 1 Pooling layer 2
Input
(c channels)
ො𝑦
d hidden
units 1
d hidden
units 2
ො𝑦
Raw Input
(c channels)
Conv.1
d1 feature
maps
ReLU
Pooling
Dropout Conv.2
d2 feature
maps
ReLU
Pooling
Dropout
Concatenate
Table: Number of inputs of the CNNs for speech, gait, and handwriting signals. c: Number of channels.
Input signal Convolution Input size c Num. inputs
Speech 2D 40 ×65 1 2600
Gait 2D 60 ×65 12 46800
Handwriting 1D 180 ×1 16 2880
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 9
16. Methods: Fusion of modalities
CNN
speech
CNN
handwriting
CNN gait
Feature maps from
last hidden layer
Avg. per
subject
TransitionTransitionTransition
Concatenate
rbf-SVM
6
6
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 10
17. Methods: Validation
Hyper-parameter Values
Filter size convolutional layers {3,5,7}
Depth of convolutional layers {4,8,16,32,64}
Hidden units in fully connected layers {16,32,64,128}
Learning rate {0.0001,0.0005,0.001}
Probability of dropout {0.1···0.9}
• A Bayesian optimization (BO) approach.
• BO use previous observations of a loss function f, to determine the next
(optimal) point to sample f.
• The loss function f is described as a Gaussian Process.
• Five-fold cross-validation.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 11
19. Data
• Speech, handwriting and gait from:
• 39 HC subjects.
• 45 PD patients.
• Balanced in age and gender.
• Several exercises are performed by the participants in each modality.
• Speech: rapid repetition of syllables: /pa-ta-ka/, /pa-ka-ta/
• Handwriting: name, signature, sentence, fixed drawings
• Gait: 40 meters walk in straight line with stops each 10 meters
• PD patients were labeled according to the MDS-UPDRS scale.
0 20 40 60 80 100
MDS-UPDRS-III score
0
2
4
6
0 10 20
MDS-UPDRS-III s
0
2
4
20Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 12
21. Classification of PD patients and HC subjects7
Modality ACC
Speech 92.3±12.3
Handwriting 66.5± 5.5
Gait 80.3±10.3
Fusion 97.6± 2.9
• State-of-art results for the fusion
of modalities.
• High accuracies are obtained
from speech and gait.
• External factors may affect re-
sults for handwrting.
0.0 0.2 0.4 0.6 0.8 1.0
False positive rate
0.0
0.2
0.4
0.6
0.8
1.0
Truepositiverate
Speech AUC=0.925
Gait AUC=0.869
Handwriting AUC=0.699
Fusion AUC=0.954
7
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 13
22. Classification of PD patients in several stages of the disease8
HC PD1 PD2 PD3
HC 95.4% 2.3% 2.3% 0.0%
PD1 28.6% 21.4% 50.0% 0.0%
PD2 15.3% 3.9% 76.9% 3.9%
PD3 0.0% 0.0% 81.8% 18.2%
• Confusion Matrix classifying HC subjects and three groups of PD patients.
• HC are high accurate classified.
• PD patients in intermediate and severe state of the disease are commonly
miss-classified.
8
J. C. Vásquez-Correa et al. (2018). “Multimodal assessment of Parkinson’s disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 14
24. Conclusion
• A multimodal analysis of PD is proposed considering CNNs such that integrate
information from speech, handwriting and gait signals.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 15
25. Conclusion
• A multimodal analysis of PD is proposed considering CNNs such that integrate
information from speech, handwriting and gait signals.
• The proposed method aims to model the difficulty of patients to start/stop the
movement of muscles in lower and upper limbs, and in speech.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 16
26. Conclusion
• A multimodal analysis of PD is proposed considering CNNs such that integrate
information from speech, handwriting and gait signals.
• The proposed method aims to model the difficulty of patients to start/stop the
movement of muscles in lower and upper limbs, and in speech.
• The proposed approach is highly accurate to classify PD patients and HC
subjects.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 17
27. Conclusion
• A multimodal analysis of PD is proposed considering CNNs such that integrate
information from speech, handwriting and gait signals.
• The proposed method aims to model the difficulty of patients to start/stop the
movement of muscles in lower and upper limbs, and in speech.
• The proposed approach is highly accurate to classify PD patients and HC
subjects.
• Classification of PD patients in several stages of the disease is promising with
the proposed fusion strategy.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 18
28. Ongoing work
• Prediction of the speech deficits of the patients according to a specific scale
to assess dysarthria9
.
• A multi-task learning approach to predict several speech deficits of the pa-
tients, e.g., lips, palate, larynx, and tongue movements10
.
• Unobtrusive monitoring of the speech impairments of the patients trough mo-
bile devices11
.
• Tracking the disease progression of the patients in a longitudinal study.
• Time-delay neural networks to track the temporal dynamics of the transitions
for speech, handwriting, and gait.
9
Vásquez-Correa, J. C., Orozco-Arroyave, J. R., Bocklet, T. et. al. (2018). “Towards an Automatic Evaluation of the Dysarthria Level of Patients with Parkinson’s
Disease”. In: Journal of Communication Disorders (to appear).
10
Vásquez-Correa, J. C., et. al. (2018). “A Multitask Learning Approach to Assess the Dysarthria Severity in Patients with Parkinson’s Disease”. In: 19th
International Conference of the Speech and Communication Association (INTERSPEECH).
11
Arias-Vergara, T., Vásquez-Correa, J. C., et. al. (2018). “Unontrusive Monitoring of Speech Impairments of Parkinson’s Disease Patients Through Mobile
Devices”. In: IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 19
29. Thanks for listening.
Any questions?
juan.vasquez@fau.de
www5.cs.fau.de/en/our-team/vasquez-camilo
Training Network on Automatic Processing of PAthological Speech (TAPAS)
Horizon 2020 Marie Sklodowska-Curie Actions Initial Training Network European Training Network
(MSCA-ITN-ETN) project.
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 19
30. References I
Hornykiewicz, O. (1998). “Biochemical aspects of Parkinson’s disease”. In:
Neurology 51.2 Suppl 2, S2–S9.
Goetz, C. G. et al. (2008). “Movement Disorder Society-sponsored revision of the
Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): Scale presentation
and clinimetric testing results”. In: Movement disorders 23.15, pp. 2129–2170.
Vásquez-Correa, J. C. et al. (2018). “Multimodal assessment of Parkinson’s
disease: a deep learning approach”. In: Journal of Biomedical and Health
Informatics (to appear).
Vásquez-Correa, J. C., Orozco-Arroyave, J. R., Bocklet, T. et. al. (2018). “Towards
an Automatic Evaluation of the Dysarthria Level of Patients with Parkinson’s
Disease”. In: Journal of Communication Disorders (to appear).
Vásquez-Correa, J. C., Orozco-Arroyave, J. R. and Nöth, E. (2017). “Convolutional
Neural Network to Model Articulation Impairments in Patients with Parkinson’s
Disease”. In: 18th International Conference of the Speech and Communication
Association (INTERSPEECH).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 19
31. References II
Vásquez-Correa, et. al. (2018). “Phonological posteriors and GRU recurrent units
to assess speech impairments of patients with ParkinsonâC™s disease”. In:
21st International Conference on Text, Speech and Dialogue (TSD).
Vásquez-Correa, J. C., et. al. (2018). “A Multitask Learning Approach to Assess the
Dysarthria Severity in Patients with Parkinson’s Disease”. In: 19th International
Conference of the Speech and Communication Association (INTERSPEECH).
Arias-Vergara, T., Vásquez-Correa, J. C., et. al. (2018). “Unontrusive Monitoring of
Speech Impairments of Parkinson’s Disease Patients Through Mobile Devices”.
In: IEEE International Conference on Acoustics, Speech and Signal Processing
(ICASSP).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 20
32. Multimodal Assessment of Parkinson’s Disease:
A Deep Learning Approach
Camilo Vásquez
Pattern Recognition Lab, Friedrich-Alexander University of Erlangen-Nürnberg
July 18, 2018
33. Full results
Table: Multimodal classification of PD patients and HC subjects. Acc. Test: accuracy in the test set, Acc. Dev.:
accuracy in the development set, AUC: area under the ROC curve, N.: Number of parameters in the CNN.
Bio-signal Acc. Test Acc. Dev. AUC N.
Speech baseline 74.5±1.7 77.0±2.4 0.841
Speech CNN 92.3±12.3 99.4±0.7 0.925 140055
Gait baseline 63.0±8.9 66.0±3.1 0.725
Gait CNN 80.3±10.3 83.3±8.9 0.878 326977
Handwriting baseline 67.1±4.2 67.7±1.7 0.725
Handwriting CNN 66.5±5.5 98.1±1.7 0.699 255560
Fusion baseline 89.0±7.8 87.8±3.1 0.944
Fusion CNN 97.6±2.9 98.8±0.6 0.988 609549
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 21
34. Appendix 212
Table: Parkinson’s disease classification using gated-recurrent units (GRU) neural networks and phonological
features.
Features Classification Num. ACC. AUC
Task Classes
Phonological PD vs HC 2 76.0±5.8 0.78
MFCC PD vs HC 2 65.0±4.7 0.66
Phonological+MFCC PD vs HC 2 64.0±5.6 0.69
12
Vásquez-Correa, et. al. (2018). “Phonological posteriors and GRU recurrent units to assess speech impairments of patients with ParkinsonâC™s disease”. In:
21st International Conference on Text, Speech and Dialogue (TSD).
Camilo Vásquez | Multimodal Assessment of Parkinson’s disease July 18, 2018 22