Epileptic Seizure: prediction and prevention

1
Epileptic Seizure prediction and prevention

• Dan CoughlinKevin McCabeBob McCarthySteve
  Moffett

2
Background

• Epilepsy is a brain disease that triggers
  seizures
• Electroencephalograms (EEGs) read electrical
  impulses from the brain

3
Prediction
4
Methods

• Artificial Neural Networks (ANNs)
• Support Vector Machines (SVM)
• Fuzzy Logic

5
ANNs Lyapunovs Constant

• Impending epileptic epoch will lower chaos of
  brain waves
• Lyapunovs exponent can model the amount of
  chaos.
• Exponent (normally positive) will decrease when
  seizure is coming

6
ANNs

• Zandi - correlation between the time intervals
  between positive zero crossings in the signal and
  an oncoming seizure
• Use the probability density function, p(x), to
  model entropy and predict seizure

7
SVMs

• Represent the examples as points in space, mapped
  so the examples of the categories are as wide as
  possible
• Use Caos method to classify each data series

8
Fuzzy Logic

• Creates states in between 0 and 1.
• Eg. very low, low, medium, high, very
  high.
• This creates a better way to classify the risk of
  an epileptic seizure

Ex. Cold 0, 1 Hot 0, 1 Warm can be
interpreted as Cold 0, Hot 0.
9
Fuzzy Logic

• Basic Structure
• A fuzzifier, which converts crisp values (real
  time values) into fuzzy values.
• An interference engine, that applies a fuzzy
  reasoning mechanism to obtain a fuzzy output
• A defuzzifier, which translates this new output
  into crisp values
• A knowledge base which contains both an ensemble
  of fuzzy rules known as rule base and an ensemble
  of membership functions know as database

10
Fuzzy Logic with HDT

• Hierarchical Decision Trees
• Greatly reduce miss-classification
• Removes unnecessary computations from the system

11
Accuracy
Method Best Accuracy
ANNs 84.7
SVMs 100
Fuzzy Logic 97.5
12
Prevention
13
PREVENTION

• Prevention of epileptic seizures through
  medications
• Prevention of epileptic seizures through surgery
  and common practices
• Preventing a seizure with the use of Biosensors
• Vagus Nerve Stimulation
• Purdue University Nanotech Sensor
• Glutamine-Glutamate Transfer
• Electrical Pulses on Rats
• Most Viable Method

14
PREVENTION

• Prevention using medications
• Most Common Medications
• Tegretol or Carbatrol (carbamazepine)
• Zarontin (ethosuximide)
• Valium and similar tranquilizers such as Klonopin
  or Tranxene
• Anti-convulsion Phenytoin also known as
  Dilantin
• Common Side Effects
• Slurred Speech
• Nausea and Vomiting
• Rash
• Depression
• Headache
• Light Headed

15
PREVENTION

• Surgery
• Not 100 effective
• Can be effective if taken with medication
• Common Practices
• Ketogenic Diet
• Low carbohydrate, high-fat diet
• Get plenty of sleep
• Avoid bright, flashing lights and other visual
  stimuli
• Avoid video games, watching TV, drugs and alcohol

16
PREVENTION

• VNS Vagus Nerve Stimulation (passes thru neck
  to brain)
• Designed to prevent seizures by sending regular,
  mild pulses of electrical energy to the brain via
  the vagus nerve
• Pulses supplied by a device similar to a
  pacemaker
• Works for 30 seconds of stimulation followed by
  5 minutes of no stimulation
• Holding magnetic near devices activates it
  outside of its programmed interval
• Stimulation Parameters
• Stimulation amplitude, frequency, pulse width
• Relieves side effects (pain) and controls
  seizure

17
PREVENTION

• Purdue University Nanotech Sensor
• Transmitter and battery implanted in the brain
• Detects the signs of an epileptic seizure before
  it occurs
• Data will be picked up by an external receiver
  not implanted under the scalp
• Collect data specifically related to epileptic
  seizures from one thousand channels or locations
  in the brain
• The more channels, the more parts of the brain to
  look at simultaneously
• The electrodes that will get the data are
  inserted in the brain through holes made in the
  skull and are connected directly to the
  transmitter by the use of wires
• Prevents an epileptic focal seizure
• Researchers are creating a neuroprosthesis that
  dispenses a neurotransmitter called GABA that
  calms the brain once a seizure is detected
• Electrode is coated with engineered neurons and
  once they are stimulated, will release the
  neurotransmitter to inhibit the seizure

18
PREVENTION

• Molecular Imaging Biosensor
• Identifies excess amounts of neurotransmitter
  glutamate build up in brain tissue
• Excess levels thought to be produced by
  dysfunctional glutamate-glutamine shuttle
• Biological sensors being developed to detect
  glutamate levels from shuttle process
• Using FRET (fluorescence resonance energy
  transfer) imaging and electrical signals to
  detect evidence of alterations
• If technology is feasible and shows that
  epileptic seizures occur from this imbalance,
  this will be a potential new therapeutic way to
  control epilepsy

19
PREVENTION

• Electric pulses used on rats
• Supported by the Canadian Institutes of Health
  Research (CIHR) and The Natural Sciences and
  Engineering Council of Canada
• Electrical stimuli are applied to the neurons and
  in the Mossy Fibers of the rat
• Early results show that this technique can
  prevent the upcoming electrical event
• Successful suppression of these events is
  achieved using an extra cellular field
  stimulating electrode

20
PREVENTION

• Most Viable Method
• VNS Vagus Nerve Stimulation
• Most reviews from patients say it has stopped
  their seizures all together
• Those that have not stopped have experienced
  seizures less frequent and are mild compared to
  before implantation of the device
• Some patients experience side effects such as a
  hoarse voice or speech impediments
• In little cases, the device has not worked at all

21
Market and Products
22
Overview of Market/Products

• Most epilepsy detection/prevention devices are
  not yet products, and lack marketable features
  such as mobility, and battery life.
• Most effective devices are for stationary
  patients hooked up while in a bed or a lab
• Tradeoff in available products, processing power
  versus battery life.
• Prevention devices require implantable sensors,
  whereas detection devices can be found out of
  body.

23
Types of Products

• Open Loop Implantable devices
• Closed Loop Implantable devices
• Seizure Detection while sleeping
• Electrodermal Activity Sensor
• Audio sensors

24
Open Loop vs Closed Loop Devices

• Open Loop constantly provides electrical stimulus
  (usually to a particular nerve or brain region)
  to stave off seizures.
• Open Loop are more mobile with less hardware
  required, less power requirements.
• Closed Loop read in data from the body, and react
  with appropriate feedback response.
• Closed Loop require more processing
  power/hardware, so primarily used at stationary
  locations.

25
Available Prevention/Detection Products

• Emfit detects shaking movements/hyperventilation
  typical of seizures while sleeping
• Placed on bed underneath sleeper, triggers alarm

http//www.tunstall.co.uk/assets/Literature/477-Ep
ilepsy_product_datasheet.pdf
26
Available Prevention/Detection Products

• Bed Sensor with Microphone to detect audible
  sounds sometimes associated with seizures

http//www.medpage-ltd.com/MEDPAGE20MANUAL20MP2
20REV-01-01.04-09.pdf
27
Available Prevention/Detection Products

• Electrodermal Activity Sensor measures skin
  conductance.
• Electrodes sense change when seizure occurs .

http//affect.media.mit.edu/pdfs/10.Poh-etal-EMBC2
010.pdf
28
Bibliography

• Bezobrazova, S. Golovko, V. , "Comparative
  Analysis of Forecasting Neural Networks in the
  Application for Epilepsy Detection," Intelligent
  Data Acquisition and Advanced Computing Systems
  Technology and Applications, 2007. IDAACS 2007.
  4th IEEE Workshop on , vol., no., pp.202-206, 6-8
  Sept. 2007
• Zandi, A.S. Dumont, G.A. Javidan, M. Tafreshi,
  R. , "An entropy-based approach to predict
  seizures in temporal lobe epilepsy using scalp
  EEG," Engineering in Medicine and Biology
  Society, 2009. EMBC 2009. Annual International
  Conference of the IEEE , vol., no., pp.228-231,
  3-6 Sept. 2009

29
Bibliography

• Ye Yuan , "Detection of epileptic seizure based
  on EEG signals," Image and Signal Processing
  (CISP), 2010 3rd International Congress on ,
  vol.9, no., pp.4209-4211, 16-18 Oct. 2010
• Sukanesh, R. Harikumar, R. , "Fuzzy techniques
  and hierarchical aggregation functions decision
  trees for the classification of epilepsy risk
  levels from EEG signals," TENCON 2008 - 2008 IEEE
  Region 10 Conference , vol., no., pp.1-6, 19-21
  Nov. 2008

30
Bibliography

• "Vagus Nerve Stimulation Epilepsy.com."
  Epilepsy and Seizure Information for Patients and
  Health Professionals Epilepsy.com. Ed. Steven
  C. Schachter. 15 Dec. 2006. Web. 03 Apr. 2011.
  lthttp//www.epilepsy.com/epilepsy/vnsgt.
• Reimer, Richard J. "Real-time Imaging of the
  Glutamine-Glutamate Shuttle in Epilepsy - Dana
  Foundation." Brain and Brain Research Information
  - Dana Foundation. Dec. 2006. Web. 03 Apr. 2011.
  lthttp//www.dana.org/grants/imaging/detail.aspx?id
  11182gt.

31
Bibliography

• Dalton, Anthony "Detecting Epileptic Seizure
  Using Wearable Sensor Technologies"
• http//ama-ieee.embs.org/wp- content/themes/ieee/
  papers/March202320- 20PM/Dalton20Abstract2085
  .pdf
• Hively, L.M.,Kruse, K.L., Munro, N.B.,
  Protopopescu, V.A. "Epilepsy Forewarning Using a
  Hand Held Device" Febrary 2005,
  http//www.ornl.gov/webworks/cppr/y2005/rpt/12281
  9.pdf