3. Problem
• Education is the main foundation of life
• 1 of every 88 children is diagnosed with
autism (NIMH)
• The current state of autism education is
separated from its optimum potential by
the lack of incorporation of current
technology
4. Current Solutions
• Children are taught using several educational devices
and programs such as Hatch TeachSmartTM
and Smart
Interactive Whiteboards
• Inability to accurately assess the attention of a child
while using educational material
• Need to aid teachers in evaluating educational
progress of children
• Current methods of progress monitoring are not
optimized for children with cognitive disabilities
5. External Collaborator
• Hatch is a leading
provider of early
childhood educational
technology
• Provides adaptive
teaching technology in all
levels of school to
enhance the education
system
6. External Collaborator
• Hatch has provided:
– Smart Interactive Whiteboard
– JORO Pro Lift Stand
– TeachSmartTM
software
– Lenovo computer
• We will provide to Hatch:
– Preliminary data to determine
applicability of biometric tools to
assist technology based
educational activities
– Device testing, building and
implementation, design, building
and testing of a prototype
7. Specific Aims
• Detect the attention level of a child
• Assist instructors in determining the child’s
level of interest during use of educational
technology
8. Device Specifications
Product Specification Design Specification
Detects EEG Signals
1) EEG headset used to determine attention towards educational
material
2) Output chart of attention and meditation scores for duration of
use
3) Visible real-time indications of attention levels
Aids Teacher Assessment of Children
1) In program feature to allow teachers to attach notes to specific
attention sample
2) Output notes as .txt file for later use
Ease of Use 1) Adjustable or one-size-fits-all design
2) Less than one minute for setup
Unobtrusive to Child During Operation
1) EEG headset is integrated into a child friendly design to reduce
distraction or anxiety
Reliability 1) Requires little to no maintenance between tests
Safety 1) Device must conform to electronics safety standards
9. Gantt Chart
Research and
Initial Testing
All Group
Members
All Group
Members
Emotiv Epoc EEG
Headset
Preliminary
Testing
Nicholas
Helmstetter &
Bishestha
Adhikari
Nicholas
Helmstetter &
Bishestha
Adhikari
Nicholas
Helmstetter &
Bishestha
Adhikari
Arthur Assamoi
& Bishestha
Adhikari
Arthur Assamoi
& Bishestha
Adhikari
Hall Effect
Research and
Preliminary
Testing
Morgan
Auzenne &
Arthur Assamoi
Morgan Auzenne
& Arthur Assamoi
Morgan Auzenne
& Arthur
Assamoi
Eye/Gaze
Tracking Research
Bishestha
Adhikari &
Morgan Auzenne
Bishestha
Adhikari &
Morgan Auzenne
Bishestha
Adhikari &
Morgan
Auzenne
Laser Tracking
Construction and
Preliminary
Testing
Nicholas
Helmstetter
Nicholas
Helmstetter
Cursor Tracking
Research
Nicholas
Helmstetter
Nicholas
Helmstetter
NeurosSky EEG
and LabVIEW
Construction
Nicholas
Helmstetter
Nicholas
Helmstetter
NeuroSky EEG
Headset and
LabVIEW Program
Testing
All Group
Members
Data Analysis
Nicholas
Helmstetter
Final Presentation
All Group
Members
September October November December January February March April May
10. Design Constraints
• Short setup time required when working with children
(ECEC Ruston, LA)
• Child friendly headset required to reduce anxiety (Holy
Angels, Shreveport, LA)
• Near real time visual data reporting (Caddo Parish
School System, LA)
• Easy-to-use interface enabling quick retrieval and
analysis of useful information (Gilmore Foundation, MS)
13. Design Decision
Criteria Weight (%)
Design Alternative
Emotiv Epoc NeuroSky MindWave
Rating Weighted
Rating
Rating Weighted
rating
Raw EEG 5 4 0.2 2 0.1
Data Collection 15 1 0.15 3 0.45
Setup Time 30 1 0.3 4 1.2
User Comfort 10 1 0.1 4 0.4
Cost 10 2 0.2 4 0.4
Reliability 30 4 1.2 3 0.9
Total 100 2.15 3.45
Rating Value
Unsatisfactory 1
Tolerable 2
Good 3
Very Good 4
14. NeuroSky Headset
• Commercially available
headset
• Low Cost
– Available versions range
from $70.00 - $200.00
• Biosensor measures
brainwave impulses from
FP1 and EMG from ear lobe
• Capable of measuring
Alpha, Beta, Gamma, Delta,
and Theta brain waves
15. NeuroSky Headset
• Utilizes ThinkGearTM
chip
to communicate with
LabVIEW programming
via Bluetooth
• Capable of outputting
raw EEG signals,
attention and meditation
scores, and blink
strength detection
16. Determining Attention
• Neural bio-recorder used as input
which measures and interprets brain
activities
• The application of a single electrode
measures the change in field
potential over time arising from
synaptic current and forms the basis
for EEG
• Readings are inferred from
processing beta and alpha waveform
activity
• Provides two 100-scale outputs
operating at 0.5 Hz described by the
ThinkGearTM
chip as “Attention” and
“Meditation”
Robelledo-Mendez, 2009
18. Quantifying Attention
• The active and reference electrodes in the EEG headset measure electrical potential
• Electrical potential is supplied directly to the embedded chipset for filtering and separation
• The relative power of the alpha and beta waves in relation to the total EEG signal can be used to
determine the cognitive state of person
• The equations used for analysis are as follows:
• N is the number of electrodes (one in this case), Pαk is the power in the alpha band for signal k
and αi is the total power in the alpha band for all N signals at time window i
• These variables are similar for the beta band
• The power of the beta wave is multiplied by five because beta waves are usually smaller than
alpha waves by a factor of five
• If αi > βi, then the state is relaxed
• Otherwise, the state is attentive
Gomez, 2002
22. Testing Strategy
• Purpose was to correlate attention scores to
various activities
• Verify NeuroSky/ThinkGearTM
algorithms for
quantifying attention can be reproducibly
correlated to mental states
• Utilized several activities requiring varying
degrees of mental activity
• Recorded attention and meditation scores at 0.5
Hz
24. Pearson’s Correlation
Coefficient Analysis
Self-reported Attention Score vs. Acquired Attention Score
Subject # 1
Activity Comparison Rank Average Attention Score
Game 1 4 40.230
Article 1 3 40.230
Article 2 1 55.426
Game 2 2 41.590
Image 0 77.311
Pearson's r = -0.879
Subject # 2
Activity Comparison Rank Average Attention Score
Game 1 4 54.164
Article 1 3 47.098
Article 2 1 50.049
Game 2 2 45.459
Image 0 63.902
Pearson's r = -0.482
Subject # 3
Activity Comparison Rank Average Attention Score
Game 1 3 68.339
Article 1 2 54.935
Article2 1 32.090
Game 2 4 44.629
Image 0 67.426
Pearson's r = -0.096
25. Histogram Analysis
Percent of Scores Over 50 – 98.36% Percent of Scores Over 50 – 67.21%
Percent of Scores Over 50 – 86.89% Percent of Scores Over 50 – 57.38%
27. Conclusions
• Attention and meditation scores not highly
correlated
• Self-reported scores not highly correlated
to acquired scores
– Based on Pearson’s r correlation coefficient
• Visible trend of attention for activities on
individual basis
– Based on histogram trend analysis
28. Recap
• The device meets most of the design
specifications
• Allows teachers to attach notes to specific
samples
• Setup time takes less than one minute
• Child friendly design reduces distraction or
anxiety
• Little maintenance required between tests
• Device meets all safety standards
29. Future Work
• The specification that will need to be improved is
Detecting EEG Signals
• The LabVIEW program is able to detect and display the
Raw EEG signals and the ThinkGearTM
scores for
attention and meditation
• Obtaining individual EEG band data, particularly for
alpha and beta waves, is the next step in data
acquisition
• A more advanced EEG headset and/or additional
programming may be required
• Create data analysis procedure for determining attention
from alpha and beta waveforms
30. Future Work
• The next phase of testing is planned to include children
– IRB approval will be necessary for this phase of testing
– Incorporate more Hatch educational material into testing
procedure
• Utilize Hatch educational software as primary interactive
material for testing
• Incorporate alternative methods of interaction between the
Smart board system and the child
– Increase accessibility for children with physical disabilities
• Reexamine tracking technologies such as Eye/Gaze tracking
which may assist in quantifying the engagement of the child
31. Acknowledgments
• Neurobotix would like to thank:
– Dr.McManis and Hatch for providing
necessary equipment and support regarding
early childhood education
– Dr. Iasemidis and Dr. Vlachos for research
and testing support
– Austin Hoggatt for assistance in research,
development, testing, data analysis and
troubleshooting
– Dr. O’Neal for guidance throughout the project
32. References
• Bremner, F. J., F. Moritz, and V. Benignus. "EEG Correlates of Attention In
Humans."Neuropsychologia. 10 (1972): 307-12.
• Gomez, Pablo. Power Analysis of Alpha and Beta Waves in EEG Signals to
Determine the Most Likely State of a Subject. Tech. Miami: Florida International
University, 2002.
• Insel, Thomas. "Autism Prevalence: More Affected or More Detected?" NIMH.
National Institutes of Health, 29 Mar. 2012. Web. Oct. 2012.
• Niedermeyer, E. "The Normal EEG of the Waking
Adult." Electroencephalography: Basic Principles, Clinical Applications and Related
Fields. Baltimore: Lippincott Williams & Wilkins, 1999. 149-173.
• Robelledo-Mendez, Genaro, et al. "Assessing NeuroSky’s Usability to Detect
Attention Levels in an Assessment Exercise." Human-Computer Interaction 56.10
(2009): 149-58.
• Tatum, William, IV, and et al. Handbook of EEG Interpretation. N.p.: Demos Medical
LLC, 2008.
• Images taken from:
– www.neurosky.com
– www.hatchearlylearing.com
– www.emotiv.com