Conference PaperPDF Available

Smart Phone Application Development for Monitoring Epilepsy Seizure Detection based on EEG signal Classification

Authors:

Abstract and Figures

Automated epilepsy seizure detection is the solution to the limitation and time consuming of manual epilepsy monitoring and detection using EEG signals. In this paper we developed a technique for epilepsy seizure detection using EEG signals. The signal will be pre-processed and filtered using multiple filters. Then, the filtered signal will be decomposed into sub-bands. Furthermore, feature extraction is applied; we developed a combined feature consists of combining three features into one. Finally, we used well-known classifiers such as Support Vector Machine (SVM), Artificial Neural Network (ANN), and K-Nears Neighbor (KNN) to differentiate between epileptic and non-epileptic signals, and we achieved an accuracy of 97%. Furthermore, we developed an Android-based smartphone application for monitoring epilepsy detection based on the classification results of the EEG signal. A notification will be sent to the patient, doctors, and family members when an epilepsy seizure occurs. Once the EEG signal is classified as epileptic, the App will display a visual notification indicating that Epileptic Seizure has been detected. Moreover, it will trigger an alarm and send a message notification to all associated phone numbers, and the EEG signal will display on the App. Although we are using an EEG signal from a dataset, we have generated both normal and epileptic EEG signals using a waveform generator, and we have displayed those signals on the spectrum analyzer for future real time detection using our Android App.
Content may be subject to copyright.
Smart Phone Application Development for
Monitoring Epilepsy Seizure Detection based on EEG
signal Classification
Zakareya Lasefr, Ramasani Rakesh Reddy, and Khaled Elleithy
Department of Computer Science and Engineering
University of Bridgeport
Bridgeport, CT, USA
zlasefr@my.bridgeport.edu, rramasan@my.bridgeport.edu, elleithy@bridgeport.edu
Abstract
Automated epilepsy seizure detection is the solution to
the limitation and time consuming of manual epilepsy monitoring
and detection using EEG signals. In this paper we developed a
technique for epilepsy seizure detection using EEG signals. The
signal will be pre-processed and filtered using multiple filters.
Then, the filtered signal will be decomposed into sub-bands.
Furthermore, feature extraction is applied; we developed a
combined feature consists of combining three features into one.
Finally, we used well-known classifiers such as Support Vector
Machine (SVM), Artificial Neural Network (ANN), and K-Nears
Neighbor (KNN) to differentiate between epileptic and non-
epileptic signals, and we achieved an accuracy of 97%.
Furthermore, we developed an Android-based smartphone
application for monitoring epilepsy detection based on the
classification results of the EEG signal. A notification will be sent
to the patient, doctors, and family members when an epilepsy
seizure occurs. Once the EEG signal is classified as epileptic, the
App will display a visual notification indicating that Epileptic
Seizure has been detected. Moreover, it will trigger an alarm and
send a message notification to all associated phone numbers, and
the EEG signal will display on the App. Although we are using an
EEG signal from a dataset, we have generated both normal and
epileptic EEG signals using a waveform generator, and we have
displayed those signals on the spectrum analyzer for future real
time detection using our Android App.
KeywordsDigital Ocean Server, NGINX Server, python Flask.
I. I
NTRODUCTION
Epilepsy is a neurological disorder disease that affects the central
nervous system of the human brain that can disrupt the activity of the
nervous cells in the brain which will result in unusual behavior that can
lead to loss of consciousness called epileptic seizure. The seizure can
endanger the life of the patient when driving, swimming, or performing
any other activities [1].
Today the smartphone based applications are being used for
solving human real life problems. And these applications play a
vital role in the health sciences for improving the ways of
treatment by introducing most promising methods. Epilepsy is a
serious and disorder behavior that affects 5% of the world
population, therefore, there is a need for developing an efficient
smart phone application that can monitor the behavior of an
epileptic patient as well as sending an immediate update to the
care takers [2]. The main idea behind developing this android
application is to notify time to time behavior of epilepsy patients
as soon as possible to all the care takers [3]. An immediate
medical assistance is needed when there is a seizure. The main
idea behind our android application is to give immediate
notification of spectrum signal and epilepsy state, and an alarm
is triggered based on it.
Android framework contains rich application based interface
which helps to build innovative apps which are useful in various
fields especially in health sciences. Android operation system is
built by modifying a Linux kernel, and the open source
methodology allows developers across the world to participate
and develop innovative applications for solving the real-life
problems.
This study has developed an android based application
which can give immediate notifications once a seizure occurs to
an epileptic patient so that s/he can get the proper medical
attention. The app provides the features such as sending an
immediate notification on seizure detection, showing the EEG
Spectrum of the patient state, the Epilepsy detected signal and
trigger an alarm to all registered care takers.
II. RELATED
WORK
There are not many related works using methods involving
smartphone applications, or similar devices for epilepsy seizure
detection. Moreover, most of the existing devices require the
patient to wear some uncomfortable electrodes, and other
recording, or monitoring devices that will limit the movement of
the patient as well as making him/her uncomfortable.
In 2016 Luca Cattani et al. developed an android app for
neonatal colonic seizures detection based on the movement of
some body parts obtained from video and image processing. In
this paper, they have used a laptop in the background of the
processing as well as testing their technique on a colonic seizure
simulator and not on a real patient [4].
A method proposed by DeVaul, Barkalow et al. in 2006
where they have developed an algorithm to carefully monitor the
sleep activity for epilepsy detection [5]. For unusual activity
during sleep, a live video recording is done to store the
information to be analyzed. However, this method lacks
immediate notification. The video based app records the video
for an abnormal event then processed. Processing the stored
videos may result in a delay in detecting the abnormal activity.
978-1-5386-1104-3/17/$31.00 ©2017 IEEE 83
This delay may cause dangerous situations. There are few
studies which suggest the use of Fitbit devices to notify the
abnormal seizure detection based on SMS and GPS alerts. The
studies lack compatibility with all the mobile versions and
operating systems available. Moreover, the cost of these devices
is high.
Another smartphone App development was introduced by
Stefan Madansingh et al. in 2015 for fall detection which was
based on analyzing the movement of the body using the
embedded sensors already available in the smartphone [6].
However, this technique does not provide any type of
notification or location.
Shih-Hau Fang et al. have implemented an android
smartphone app for fall detection based on three components;
sensing the accelerometers data from the phone embedded
sensor, learning the relationship between the fall behavior and
the collected data, and alerting the pre-configured contacts via
messages[7]. They have achieved a sensitivity of 72.22% and a
specificity of 73.78%. [8]have developed an android smartphone
app for fall detection for elderly people [8]. Their algorithm was
implemented based on accelerometers available on the
smartphone to evaluate the performance of the robust
algorithms, rather than thresholding. Furthermore, their
application also provides GPS locations using the available GPS
interface on the phone as well as sending an email and SMS
notification.
We can conclude from the related work that most of the
developed applications are related to the fall detection rather
than the epilepsy seizure detection. Even though there is only a
few app development involving epilepsy seizure detection, they
are based on the body movement rather than EEG signals that is
the main source of epilepsy information.
III. M
ETHODOLOGY
The current application development consists of three main
modules namely a Login/Register, EEG Signal, and Visual and
Alarm Notification. All modules communicate through socket
events between App and server. All spectrum analyzer signals
and Alarm notifications are notified through an http POST
request notification to a Digital Ocean server. The Digital Ocean
Server utilizes the Http POST Response interface to get
notifications from the client device and GET Request between
App and the server.
The Spectrum Analyzer will upload the EEG signal onto a
Digital Ocean Http Web bases NGINX Server interface, that will
send a socket event notification to the app of a new upload. Once
the app receives a new notification, it retrieves the uploaded
EEG signal through a PULL notification and displays the signal.
The Alarm module triggers an alarm to all registered users and
care takers when there is a new notification from POSTMAN if
there is an epilepsy seizure in the signal. The Alarm module uses
Android Email notification interface available in Android APIs
to trigger and notify the alarm. Once the EEG signal has been
processed to detect epilepsy, a visual notification is sent through
a http POST request to determine whether an epilepsy seizure
has occurred.
Figure 1. Architecture diagram
The Architecture shown in Fig 1 involves three main
components namely Epilepsy Spectrum Device (EEG signal
classifier), Cloud SERVER, and Android Application.The cloud
SERVER conists of NGINX Http based Webserver and
Gunicorn Python based WSGI Server interfaced with Mongo
Database.
The deviceside can be a spectrum anlyzer or an EEG signal
classifier. The EEG classifier signals are matlab proceesed
samples of EEG signals. The devic-eside continuously monitors
the patient and updates the server if there is an epilypsy detected.
The Server componant is implemented on a machine running
continously on cloud service provided by Digital Ocean. A
server is implemented on Ubuntu machine running with Nginx
based Http web server. The server module also contains a
Gunicorn WSGI based server to monitor and maintain the data
received via Mongo Database.
The NGINX is based Http webserver that is based on
python, an open-source and high-performance HTTP server is
an IMAP/POP3 proxy server. NGINX is known for its high
performance, stability, rich feature set, simple configuration, and
low resource consumption.
Unlike traditional servers, NGINX doesn’t rely on threads
to handle requests. Instead it uses a much more scalable event-
driven (asynchronous) architecture. This architecture uses small,
but more importantly, predictable amounts of memory under the
load.
The storage engine is the primary component of Mongo DB
that is responsible for managing data. Mongo DB provides a
variety of storage engines, allowing the user to choose one most
suited to your application.
IV. E
PILEPSY
A
PP
The Android application involves a CONNECT phase
followed by a Capture and monitor phase. The client is basically
a care taker or doctor who installs the app, before the registration
process goes through a CONNECT phase between server and
client.
84
Figure 2. Connection between APP and server
The connect establishment as shown in figure 2 is totally
based on socket connection. The Android application
implemented based on Android API 23 is compatible to all the
Android Lower level devices. When a user opens the app for a
socket connection request to the server, the Gunicorn server
responds back with event Connect_ok to the App. Once The App
establishes the connection with the server, it sends a Socket
Data_Request to the Server. The server parses the currnt
available data in csv format. The data is received from the device
that is posted throught the Http POST request in the form of the
jason format. The received jason format data is parsed and
posted to the application.
Figure 3. APP, server, and device connection
The Epilepsy device keeps monitor of the patient, and when
there is an epilepsy signal detected, the device captures the
signal and sends an Http POST Request to the Server with csv
data of the graph in the format of Jason. The received data is
stored in the Mongo Storage and parses data. Once the data is
parsed, it sends a socket update event to the application. The app
which is waiting for Update Event, sends an Http GET Request
to the server to indicate that app is ready to receive the data and
display the graph. See figure 3.
V. A
LARM
M
ODULE
The Alarm modules helps to send an immediate email based
notification to the registered care takers of the patient. The
Alarm module also sends an alarm to itself when a seizure
occurs. The Alarm module uses Android Email notification
interface available in Android APIs to trigger and notify the
alarm. Once the EEG signal has been processed to detect
Epilepsy, a Visual Notification is sent through an Http POST
request to determine whether an epilepsy seizure is occurred.
VI. R
ESULTS
The app requires the user to signup first and then activate the
verification link. This helps to send immediate notification of
epilepsy seizure and alarm to all the registered clients through
the email id notification. Sign up/login screen can be seen in
figure 4.
Figure 4. Login/Sign up screen
The spread spectrum device basically monitors and saves the
EEG spectrum in csv format. Each client or the care taker who
installs the app receives the http response from the server in the
form of Jason data, parses the data to signal, and displays the
same. Android application utilizes the mkphil graph interface
available to convert the csv data to the graph. The converted
graph is as below in figure 5.
85
Figure 5. EEG Signal
During the registration process the app keeps track of the
users who needs to be notified with the alarm when there is a
new epilepsy seizure detected. The alarm is notified as shown in
figure 6.
Figure 6. Alarm upon seizure occurrence
The seizure position and the EEG signal status can be viewed
immediately through the app pic process as shown in figure7.
Once the EEG signal has been processed to detect Epilepsy, a
Visual Notification is sent through an http POST request to
determine whether an epilepsy seizure is occurred.
Figure 7. EEG signal
VII. C
OMPARISON
Although there are not many researches that involve
smartphone application in epilepsy seizure detection, we have
compared our proposed work to the most relevant methods that
we found after a comprehensive research. It is clear from table 1
that our proposed method has outperformed the existing
methods because it is based on EEG signals as well as of the
features that it provides.
APP Author Method Features
Android
APP for
neonatal
colonic
seizures
detection
Luca Cattani
et al.
movement
of some
body parts,
video and
image
processing
A laptop
always needed
for processing.
No real patient.
Fitbit
devices
Fang, S.-H.,
et al.
SMS and
GPS alerts
Lack the
compatibility.
High cost.
Android
APP
Stefan
Madansingh
et al.
Analyzing
body
movement
using
existing
embedded
sensors
No
Notification.
Android
APP
Shih-Hau
Fang et al.
fall
detection
Low accuracy
Android
APP
Yavuz, Kocak
et al.
fall
detection
GPS locations
SMS
Proposed
Method
Android
App
Zakareya
Lasefr et al.
EEG Signal
processing
High
Accuracy.
Immediate
notification.
Development
for real time
processing.
Table 1. Comparison table
86
VIII. B
LUETOOTH
D
EVICE
Figure 7 shows the MindWave Mobile EEG Headset which
is available in the market for only $100. This device records the
EEG signal from the brain and sends it to any device via
Bluetooth. In our final App development we are using the EEG
signals collected from the MindWave Mobile EEG Headset and
process it on the modified APP in order to detect epilepsy
seizure. Such an improvement would result in faster, and more
reliable epilepsy seizure detection using only a very light
weight as well as low cost Bluetooth headset, and a smartphone.
We believe that our research would benefit all epileptic patients
as well the researchers in the field of epilepsy.
This device was used and tested in a research by Khald
Aboalayon et al for a development of a sleep stages
classification technique using EEG signal, and has proven to be
very accurate, and effective in their real time experiment
obtaining a high accuracy [9].
Figure 7. MindWave Mobile EEG Headset
IX. C
ONCLUSION
We developed an APP that monitors the behavior of epileptic
patients based on EEG processing and classification. The APP
is connected to the classification procedure through the sever,
and updates the results instantly whenever a new signal is
processed. Upon classification completion, an alarm will be
triggered should an epilepsy seizure detected.
The future work involves replacing the spectrum device with
a Bluetooth device that is attached to the patient, and sends an
EEG signal resulting in more compatible, and generalized
solution for better and faster seizure detection. The
improvements in the APP involves location detection using
google maps, and triggering an alarm as well as sending the
location of the patient to the caretakers. The seizure detection
will be based solely on signals collected from the attached
Bluetooth device, and also on the process of the signal on the
APP itself without the need to any other device. As a result, the
APP would be an efficient time saving epileptic seizure
detection tool.
X. R
EFERENCES
[1] J. F. Wernicke, K. C. Holdridge, L. Jin, T. Edison, S. Zhang, M.
E. Bangs, et al., "Seizure risk in patients with attentiondeficit
hyperactivity disorder treated with atomoxetine," Developmental
Medicine & Child Neurology, vol. 49, pp. 498-502, 2007.
[2] F. Bert, M. Giacometti, M. R. Gualano, and R. Siliquini,
"Smartphones and health promotion: a review of the evidence," J
Med Syst, vol. 38, p. 9995, Jan 2014.
[3] C. Thompson, J. White, B. Dougherty, A. Albright, and D. C.
Schmidt, "Using smartphones to detect car accidents and provide
situational awareness to emergency responders," in International
Conference on Mobile Wireless Middleware, Operating Systems,
and Applications, 2010, pp. 29-42.
[4] L. Cattani, H. P. Saini, G. Ferrari, F. Pisani, and R. Raheli,
"SmartCED: An Android application for neonatal seizures
detection," in 2016 IEEE International Symposium on Medical
Measurements and Applications (MeMeA), 2016, pp. 1-6.
[5] R. DeVaul, D. Barkalow, J. Carlton-Foss, and C. Elledge,
"Method and system for fall detection and motion analysis," ed:
Google Patents, 2006.
[6] S. Madansingh, T. A. Thrasher, C. S. Layne, and B. C. Lee,
"Smartphone based fall detection system," in 2015 15th
International Conference on Control, Automation and Systems
(ICCAS), 2015, pp. 370-374.
[7] S. H. Fang, Y. C. Liang, and K. M. Chiu, "Developing a mobile
phone-based fall detection system on Android platform," in 2012
Computing, Communications and Applications Conference,
2012, pp. 143-146.
[8] G. Yavuz, M. Kocak, G. Ergun, H. O. Alemdar, H. Yalcin, O. D.
Incel, et al., "A smartphone based fall detector with online
location support," in Proceedings of the International Workshop
on Sensing for App Phones, Zurich, Switzerland, 2010, pp. 31-35.
[9] K. A. I. Aboalayon, M. Faezipour, W. S. Almuhammadi, and S.
Moslehpour, "Sleep stage classification using EEG signal
analysis: a comprehensive survey and new investigation,"
Entropy, vol. 18, p. 272, 2016.
87
... Lasefr et al. [10] had developed a smartphone application that monitored EEG signals for epileptic seizures which runs on cloud servers. However, this approach is unsafe as it leaves a backdoor for potential privacy breach of users' health data. ...
... The data was further rescaled to values between 0 and 1 by applying minmax scaler. The data from each channel was then segmented into various window lengths of 1, 2, 4, 8 and 16 seconds in this study [10]. For each window length, it was stored as a 1-dimensional window of w × f s time-steps, where w is the length of the window in seconds and f s is the sampling frequency in Hz. ...
... ProtoNN models can be deployed on devices with scarce storages and constrained computational capacity. This kNN-based algorithm handles the trade-off between prediction accuracy and model size -a solution not proposed in earlier literatures [10]. ProtoNN implements this proposed solution by employing 3 key methods: ...
Preprint
Neonatal seizures are a commonly encountered neurological condition. They are the first clinical signs of a serious neurological disorder. Thus, rapid recognition and treatment are necessary to prevent serious fatalities. The use of electroencephalography (EEG) in the field of neurology allows precise diagnosis of several medical conditions. However, interpreting EEG signals needs the attention of highly specialized staff since the infant brain is developmentally immature during the neonatal period. Detecting seizures on time could potentially prevent the negative effects on the neurocognitive development of the infants. In recent years, neonatal seizure detection using machine learning algorithms have been gaining traction. Since there is a need for the classification of bio-signals to be computationally inexpensive in the case of seizure detection, this research presents a machine learning (ML) based architecture that operates with comparable predictive performance as previous models but with minimum level configuration. The proposed classifier was trained and tested on a public dataset of NICU seizures recorded at the Helsinki University Hospital. Our architecture achieved a best sensitivity of 87%, which is 6% more than that of the standard ML model chosen in this study. The model size of the ML classifier is optimized to just 4.84 KB with minimum prediction time of 182.61 milliseconds, thus enabling it to be deployed on wearable ultra-edge devices for quick and accurate response and obviating the need for cloud-based and other such exhaustive computational methods.
... In the literature there are numerous epileptic seizure prediction systems, which follow different approaches to identify the pre-ictal periods. Although the use of EEG signals is the most common approach [20][21][22][23][24][25][26][27][28], electrocardiograph (ECG) [29,30], electromyograph (EMG) [30], heart rate [30,31], and vibration [31,32] signals have also been used. These systems achieve high accuracy (>80%) on predicting the epileptic seizures. ...
... These systems achieve high accuracy (>80%) on predicting the epileptic seizures. Some of these implementations are edge computing (on-sensor computing) wearable devices [15,[21][22][23], whereas others, to reduce the local power consumption, combine simple data acquisition devices with remote computing [20,[24][25][26][27][28][29][30][31][32] (smartphone or cloud computing). In either case, the low power efficiency of these devices limits their capabilities and usability. ...
Article
Full-text available
This paper presents a new analog front-end classification system that serves as a wake-up engine for digital back-ends, targeting embedded devices for epileptic seizure prediction. Predicting epileptic seizures is of major importance for the patient’s quality of life as they can lead to paralyzation or even prove fatal. Existing solutions rely on power hungry embedded digital inference engines that typically consume several μW or even mW. To increase the embedded device’s autonomy, a new approach is presented combining an analog feature extractor with an analog Gaussian mixture model-based binary classifier. The proposed classification system provides an initial, power-efficient prediction with high sensitivity to switch on the digital engine for the accurate evaluation. The classifier’s circuit is chip-area efficient, operating with minimal power consumption (180 nW) at low supply voltage (0.6 V), allowing long-term continuous operation. Based on a real-world dataset, the proposed system achieves 100% sensitivity to guarantee that all seizures are predicted and good specificity (69%)
... In [4], it is adopted Discrete Wavelet Transform (DWT) to acquire inputs parameters for Artificial Neural Network (ANN) training. In [5], the authors developed a smartphone application for epilepsy detection using signal filtering * This study was financed by the Brazilian agencie CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) to decompose the signal in sub-bands, then the classification is made using Support Vector Machine (SVM), Artificial Neural Network (ANN), and K-Nears Neighbor (KNN) algorithms. Authors in [6] proposed a combination of backpropagation (BP), firefly algorithm and simulated annealing, using the last one to determine optimal learning rate and number of unit hidden on BP process. ...
Conference Paper
Full-text available
Epilepsy is a chronic neural disease that suffers around 50 million people in the world. Since epileptic seizures are spontaneous and some cases cannot be controlled, the approach to detect or even anticipate them is highly required. This paper shows the development of an algorithm based on Artificial Neural Network (ANN), in this case, a feedforward neural network with a training method of backpropagation, with the objective to detect an epileptic event. Discrete Wavelet Transform (DWT) and Fast Fourier Transform (FFT) are applied to extract input parameters to train the neural network, which classify the electroencephalogram (EEG) recording in two output classes, normal and epileptic state. Then, the network classifies the EEG signal of 20 patients, regardless of seizure type, where the performance to detect epilepsy seizures was a 98.3% accuracy, 93% sensibility, and 96.1% specificity, besides was achieved a 100% accuracy and specificity in free seizure patients.
... Another application in smart health is seizure detection. Lasefr et al. have presented an android phone-based application for seizure detection in [15]. The system consists of three modules; android phone for acquiring data and sending it to the server, a cloud web server and an epilepsy spectrum device where EEG signals are processed and classified. ...
Conference Paper
Smart health systems have become the need of the hour taking into account the rising number of patients and the limited facilities available to treat them. This paper has profited by the advantages of two major fields of research in the recent times; deep learning and edge computing. A deep learning method for epileptic seizure detection using mobile edge computing is proposed. EEG signals are used for seizure detection. A novel deep Convolutional Neural Network (CNN) with six convolutional layer, two fully connected layer with softmax activation is used to classify the EEG signal into either seizure or non-seizure category. A three tier architecture with an edge layer between the end device and the cloud is used. Out of all the advantages the edge layer can enhance privacy, this paper is mainly focused on using the edge to reduce the size of the data being sent to the cloud. A dataset with 500 samples from healthy and seizure activities is used. The raw data is represented in both time and frequency domain. The best model proposed can achieve an accuracy of 92 % in time domain, 99.22 % in frequency domain.
Chapter
In this paper, the authors are classifying time series electroencephalogram (EEG) data as epilepsy or not. Epilepsy is a type neurological disorder which affects around one percentage of the people around the world. Epilepsy is diagnosed by electroencephalographic (EEG) reading of the brain. EEG monitors the electrical activity of the brain. Studying EEG is a time-consuming and tedious task and requires lots of patience and domain knowledge. Deep learning techniques can help us reduce and automate this task and make its detection easier for doctors which in turn can help doctors to give better treatment to the patient. In this paper, the authors propose a sequential one-dimensional convolutional neural network (CNN) and state-of-the-art architectures like one-dimensional inception module and one-dimensional ResNet module which will classify time series EEG data as epilepsy or not. Using these approaches, sequential model was able to outperform the state-of-the-art models and achieved a ROC-AUC of 0.98. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Article
Full-text available
Sleep specialists often conduct manual sleep stage scoring by visually inspecting the patient's neurophysiological signals collected at sleep labs. This is, generally, a very difficult, tedious and time-consuming task. The limitations of manual sleep stage scoring have escalated the demand for developing Automatic Sleep Stage Classification (ASSC) systems. Sleep stage classification refers to identifying the various stages of sleep and is a critical step in an effort to assist physicians in the diagnosis and treatment of related sleep disorders. The aim of this paper is to survey the progress and challenges in various existing Electroencephalogram (EEG) signal-based methods used for sleep stage identification at each phase; including pre-processing, feature extraction and classification; in an attempt to find the research gaps and possibly introduce a reasonable solution. Many of the prior and current related studies use multiple EEG channels, and are based on 30 s or 20 s epoch lengths which affect the feasibility and speed of ASSC for real-time applications. Thus, in this paper, we also present a novel and efficient technique that can be implemented in an embedded hardware device to identify sleep stages using new statistical features applied to 10 s epochs of single-channel EEG signals. In this study, the PhysioNet Sleep European Data Format (EDF) Database was used. The proposed methodology achieves an average classification sensitivity, specificity and accuracy of 89.06%, 98.61% and 93.13%, respectively, when the decision tree classifier is applied. Finally, our new method is compared with those in recently published studies, which reiterates the high classification accuracy performance.
Conference Paper
Full-text available
In this paper, we present Smartphone-based Contactless Epilepsy Detector (SmartCED): an Android monitoring application able to diagnose neonatal clonic seizures and warn about their possible occurrences in realtime. SmartCED has, however, wider applicability so that it could also be used on adult patients. The main goal is to implement a wire-free and low-cost epilepsy diagnostic system, executing all the necessary processing directly on the smartphone. Seizures' recognition is based on a well-known statistical criterion, namely Maximum Likelihood (ML). As clonic seizures are characterized by quasi-periodic movements of some body parts, it is possible to detect the presence of a seizure by evaluating this periodicity from the video stream of the smartphone's camera. The heavy computational processing is carried out in the native code (C language) to enhance the performance. SmartCED presents a user-friendly interface in order to extend its use even to unskilled staff. In fact, although it integrates complex software from the technical point of view, the user has just to: start the App, “frame the patient”, and start monitoring with a simple touch.
Conference Paper
Full-text available
Falls are identified as a major health risk not only for the elderly but also for people with neurodegenerative diseases, such as epilepsy, and are considered as a major obstacle to independent living. Fast detection of falls would not only decrease the health risks by enabling quick medical response; but also make independent living a safe option for the elderly. In this paper, we propose a fall detector that uses the accelerometers available in smartphones and incorporates different algorithms for robust fall detection such as thresholding and wavelet transforms. We implemented our fall detector on a smart phone running the Android 2 operating system. We performed an extensive set of experiments for evaluating the performance of the implemented fall detector. To the best of our knowledge, although using smartphones for fall detection have been recently studied, evaluating the performance of robust algorithms, rather than thresholding, has not been explored before. Our experimental results show that compared to a simple thresholding algorithm, using wavelet transforms achieve better true positive performance while decreasing the rate of false positives drastically. Besides the fall detection capability, our implementation also provides location information using Google Maps about the person experienced the fall, using the available GPS interface on the smartphone and a warning about the fall and the location information are transmitted to the users, such as the caregivers, via SMS, email and Twitter messages. 1.
Article
Full-text available
Communication via mobile phones has become an essential tool for health professionals. The latest generation of smartphones is comparable to computers, allowing the development of new applications in health field. This paper aims to describe the use of smartphones by health professionals and patients in the field of health promotion. We conducted a bibliographic search through Pubmed. Then, research results were analyzed critically in order to select the best experiences available. All searches were carried out on November 2012 and were not limited by date. Each item from the initial search was reviewed independently by members of the project team. Initial search returned 472 items with PubMed. After the removal of duplicates, 406 items were reviewed by all the members of the project team and 21 articles were identified as specifically centered on health promotion. In the nutrition field there are applications that allow to count calories and keep a food diary or more specific platforms for people with food allergies, while about physical activity many applications suggest exercises with measurement of sports statistics. Some applications deal with lifestyles suggestions and tips. Finally, some positive experiences are reported in the prevention of falls in elderly and of sexually-transmitted diseases. Smartphones are transforming the ways of communication but the lack of monitoring of contents, the digital divide, the confidentiality of data, the exclusion of the health professional from the management of patient, are the main risks related to their use.
Article
Since todays smartphones are programmable and embed various sensors, these phones have the potential to change the way how healthcare is delivered. Fall detection is definitely one of the possibilities. Injuries due to falls are dangerous, especially for elderly people, diminishing the quality of life or even resulting in death. This study presents the implementation of a fall detection prototype for the Android-based platform. The proposed system has three components: sensing the accelerometer data from the mobile embedded sensors, learning the relationship between the fall behavior and the collected data, and alerting preconfigured contacts through message while detecting fall. We adopt different fall detection algorithms and conduct various experiments to evaluate performance. The results show that the proposed system can recognize the fall from human activities, such as sitting, walking and standing, with 72.22% sensitivity and 73.78% specificity. The experiment also investigates the impact of different locations where the phone attached. In addition, this study further analyzes the trade-off between sensitivity and specificity and discusses the additional power consumption of the devices.
Conference Paper
Accident detection systems help reduce fatalities stemming from car accidents by decreasing the response time of emergency responders. Smartphones and their onboard sensors (such as GPS receivers and accelerometers) are promising platforms for constructing such systems. This paper provides three contributions to the study of using smartphone-based accident detection systems. First, we describe solutions to key issues associated with detecting traffic accidents, such as preventing false positives by utilizing mobile context information and polling onboard sensors to detect large accelerations. Second, we present the architecture of our prototype smartphone-based accident detection system and empirically analyze its ability to resist false positives as well as its capabilities for accident reconstruction. Third, we discuss how smartphone-based accident detection can reduce overall traffic congestion and increase the preparedness of emergency responders.
Article
The comorbidity of seizures, epilepsy, and attention-deficit-hyperactivity disorder (ADHD) prompted the examination of whether atomoxetine use for ADHD is associated with an increased risk of seizures. Seizures and seizure-related symptoms were reviewed from two independent Eli Lilly and Company databases: the atomoxetine clinical trials database and the atomoxetine postmarketing spontaneous adverse event database. Review of clinical trial data indicated that the crude incidence rates of seizure adverse events were between 0.1 and 0.2%, and were not significantly different between atomoxetine, placebo, and methylphenidate. Only 2% of the postmarketing spontaneous reports of seizure events were classified as having no clear contributing or confounding factors, and the reporting rate (8 per 100 000 patients exposed) was within the expected range of population-based incidence. Although children with ADHD are increasingly recognized as being at an elevated risk for seizures, treatment of ADHD symptoms with atomoxetine does not appear to elevate this risk further. The shared vulnerability between ADHD and seizure activity should be taken into account when making treatment decisions for populations of children with epilepsy and children with ADHD.
Method and system for fall detection and motion analysis
  • R Devaul
  • D Barkalow
  • J Carlton-Foss
  • C Elledge
R. DeVaul, D. Barkalow, J. Carlton-Foss, and C. Elledge, "Method and system for fall detection and motion analysis," ed: Google Patents, 2006.
  • Computing
Computing, Communications and Applications Conference, 2012, pp. 143-146.