Brain Interface Design for Asynchronous Control

1
Brain Interface Designfor Asynchronous Control
2
ASIMO made by HONDA
3
What is Neil Squire Society
The Neil Squire Society is the organization in
Canada that for the past 20 years has used
technology, knowledge and passion to empower
Canadians with physical disabilities
Project Background

• The long-term objective of the brain interface
  project is to create a multi-position,
  brain-controlled switch that is activated by
  brain signals measured directly from the scalp of
  an individual.
• They believe that such a switch will allow an
  individual with a severe disability to have
  effective control of devices such as assistive
  appliances, computers

4
Introduction(i)

• Within the context of general brain interface
  (BI) technology research, The Nail Squire Brain
  lab has focused on BI system design specifically
  for self-paced or asynchronous control
  environments.
• This asynchronous mode of device is more natural
  than the more commonly studied synchronized
  control mode whereby the system dictates the
  control of the user

5
Introduction(ii)

• Schemes of Presentation
• Presentation about an overview of asynchronous
  control.
• Summary of Neil Squire Societys effort to
  develop asynchronous BI system.
• Discussion of several major issues that have
  arisen from asynchronous BI development.

6
Which feature is used in BCI systems?

• In typical BCI based on EEG, the operator
  generates a control signal by consciously
  changing his cognitive state when he wants to
  control device.
• The change in cognitive state is measured as
  specific temporal patterns, and signal power
  level in the operators EEG activity

7
AsynchronousVs. Synchronous
8
Synchronous Vs. Asynchronous
9
What is the Synchronous mode?

• Synchronous mode
• In this system, the system initiates the period
  of control, not the user, and user is expected to
  be consciously controlling the interface during
  the control periods likely fig.

10
What is the Synchronous mode?

• Drawbacks of the synchronous
• BI transducer will make a control signal
  regardless of whether the person is actually
  intending control.
• User cannot make control signal until system
  polling period to occur in order to engage BI
  transducer.

11
What is the Asynchronous mode?

• Asynchronous mode
• In this system, which we will call asynchronous
  control applications, are characterized by
  alternating periods of attentive idleness and
  active control as illustrated in Fig.
• In this system, there are applications that
  require constant users attention and irregular,
  user-initiated control. This system is usually
  not communications applications but control
  applications.

12
Asynchronous Control (i)

• Asynchronous control refer to the type of control
  where output signals are changed or commands are
  issued only when control is intended.
• During the no control(NC) state, one would expect
  the system output to remain neutral or unchanged.

13
Asynchronous Control (ii)

• Examples of asynchronous control
• turning on lights, changing television channels,
  and interacting with a computer.
• When you remove your hand from your computer
  mouse, you enter an NC state and the mouse output
  remains stable and unchanged-that is, the mouse
  pointer does not continue to move on the computer
  screen. In other word, The mouse is then
  available for control simply by replacing your
  hand.
• In short, asynchronous control allows the user to
  define when things happen.
• According to these points, asynchronous control
  is more characteristic of most real-world control
  applications than synchronous control.

14
Asynchronous Control (iii)

• System idling
• The neutral or unchanging system output
  response desired during periods of NC
• For effective system, we should use the idling
  system to be realized.
• This system is analogous to car engine.
• Using FP(false positive) rate measures how well
  BI transducers idle for more complete measure of
  asynchronous control.

15
Classification of EEG Device

• Idle support indicates if the interface device
  will support idling
• Available define when the interface device
  allows user control

16
Classification of EEG Device

• Idle support
• If control paradigms not support the idling,
  the system produce an unintended action if and
  when the user enters the NC state. (Midas touch
  problem)

17
Control Paradigm
Availability Idle support Idle support
Availability No idle support Idle support
Periodically Synchronous System-paced
Continuously Constantly engaged Asynchronous

• Constantly engaged mode is impractical operation
  where the user is continuously
• controlling the interface without a break and
  any NC activity will cause an error.

18
EEG-BasedAsynchronous Brain-Switches
19
Obtaining the EEG signal related movement

• Obtain the EEG signals when spinal cord injury
  (SCI) people and able-bodied people move their
  finger.
• The only difference between the people with (SCI)
  and those without are the actual physical
  movements that may or may not occur during their
  attempted finger movement.
• 10-20 System of Electrode Placement

• The international 10-20 system of electrode
  placement is the most widely used method to
  describe the placement of electrodes at specific
  intervals along the head.
• Even numbers refer to the right hemisphere and
  odd numbers refer to the left hemisphere

20
Electrode placement for Experiments
Motor cortex The anatomical region of the
brain known as Area 4 was given the name primary
motor cortex (symbol M1) after Penfield showed
that focal stimulations in this region elicited
highly localized muscle contractions at various
locations in the body.
21
Electrode placement for Experiments
They chose to limit the features to the top six
primarily because these features were the minimal
set that provided uniform coverage of the motor
areas (SMA, MI) of the cortex. The strongest
discriminatory features were found in
autocorrelations within six electrode pairs
F1-FC1, Fz-FCz, F2-FC2, FC1-C1, FCz-Cz, and
FC2-C2 on the 10-20 system for electrode
placement.
22
Asynchronous Brain-Switches

• For asynchronous control application
• Idle support
• Low FP rates

Specific signal processing algorithm
23
Classify NC state and IC state

• Problems in State Classification.
• i) This needs an indication of intent in order
  to process the data.
• ii) It is not case for synchronous
  applications because user intent is assumed
  during the control periods.

24
Classify NC state and IC state

• Method 1.
• Subjects to self-report intent during the data
  recording
• Disadvantage
• i) Self-report complicates the signal analysis
  because one does not know exactly when the
  movement was made.
• ii) It is hard to provide the user with any
  form of feedback during these sessions.

25
Classify NC state and IC state

• What is the Outlier Processing Method?
• - The outlier processing method (OPM) is
  the only BCI technique that has been designed
  specifically to differentiate idle from active
  EEG in an asynchronous control applications
• - They use the outlier processing method
  (OPM) for extracting single-trial voluntary
  movement-related potentials(VMRPs) from EEG
  related to finger movement.

26
Classify NC state and IC state

• Method 2.
• Separate the VMRP state from NC state via OPM in
  the obtained signal.
• They find the feature that VMRPs has higher
  relative power than NC state in 1-4Hz bandwidth
  via time-frequency analysis of EEG pattern.
• Using this feature, they has developed the
  low-frequency asynchronous switch design(LF-ASD)

27
Classify NC state and IC state

• The BI experiments results by SCI people and
  able-bodied people are same. In other words,
  results by actual physical movements and imagery
  of movements are same.
• LF-ASD demonstrated TP rates of 30-78 percent
  during IC states in combination with low FP rates
  of 0.5-2 percent during NC. The result is
  independent subjects.
• When the FP rate is high, the LF-ASD system is
  frustrating to use. Thus, they intentionally
  operated the system in this case.

28
Thank You.
29

• Feature Extraction Methodology
• First, signals prefiltered between 1-4Hz and the
  compound feature described by (1)

30
(No Transcript)
31
(No Transcript)