22-06-2012, 12:45 PM
A Microcontroller-Based Ambulatory Brain-Computer Interface Training System
INTRODUCTION
The goal of brain-computer interface (BCI) technology is to
augment an individual's ability to control a device without
the need for the individual to physically move. The majority
of BCI applications are for those that are severely disabled
and are unable to reliably control their movement.
Additionally, applications such as aiding the control of a
prosthetic device or providing military personnel with
augmented control capabilities are being investigated.
Although much BCI research has been performed, this
burgeoning field is still in its infancy. Numerous groups and
specialists from various fields are working together to
improve the current state of the art on many fronts. For
instance, some are studying how brainwaves can best be
processed and classified, while others are investigating
implanted electrodes or improved training methods.
DISCUSSION
Prototype data acquisition unit
The graphical user interface allows the operator to enter the
equation coefficients to be used in the classification process.
It also displays a horizontal feedback bar used in the training
game in which the user attempts to mentally alter the level
according to the given task. Additional indicators are
provided to show the number of trials, the number of times
the user succeeded and failed to achieve the given task, and
the overall success rate.
RESULTS
In designing the ABCI, the goal was to have it perform
equivalent to a desktop BCI system. Therefore, the ABCI
was compared to the BCI system developed in our lab,
where the desktop BCI was considered to be the standard.
The comparison was done by simultaneously inputting a
sinusoid wave into both the BCI and the ABCI units and
comparing their respective RMS power outputs. The stages
the waveform passed through include A/D conversion, a
bandpass filter, power calculation, and outputting the
resultant value to the monitor or PDA screen. The input
waveform was swept over a range of 1 to 30 Hz and the
results indicated a difference of less than 1%. Additionally,
the time delay between a change in the input and output was
indistinguishable on both systems and is therefore less than
the quarter of a second refresh rate of the power calculation.
Overall, evaluation of the ABCI shows that the accuracy and
response time are equivalent to the desktop BCI system.
The total cost of the ABCI training system is about $350,
where the data acquisition unit cost is approximately $50
and the PDA cost is $300. Additional equipment necessary
to amplify the EEG signals would increase the cost of the
system.
The ABCI training system was designed to act as a
portable version of a deskrop BCI system. Based on the test
results, the ABCI performs equivalent to a desktop BCI.
This system can therefore be used chronically and away
from any training site, thus removing barriers inherent in
traditional BCI systems. Studies can now take advantage of
training in a more physiologic manner, where subjects have
continuous access to train and improve their cognitive
control capabilities. This brings up some immediate
questions and potential studies such as 1) the development
of optimal training protocols and providing training
motivation and 2) whether training while performing
everyday tasks is feasible.
CONCLUSION
We believe that ambulatory training is the beginning of a
new paradigm in BCI research. An ambulatory BCI that
allows training to be performed away from the traditional
training site has been presented. This should allow more
advanced training protocols to be developed and
implemented in order to more effectively study the ability of
the brain to adapt and control augmentative devices.
INTRODUCTION
The goal of brain-computer interface (BCI) technology is to
augment an individual's ability to control a device without
the need for the individual to physically move. The majority
of BCI applications are for those that are severely disabled
and are unable to reliably control their movement.
Additionally, applications such as aiding the control of a
prosthetic device or providing military personnel with
augmented control capabilities are being investigated.
Although much BCI research has been performed, this
burgeoning field is still in its infancy. Numerous groups and
specialists from various fields are working together to
improve the current state of the art on many fronts. For
instance, some are studying how brainwaves can best be
processed and classified, while others are investigating
implanted electrodes or improved training methods.
DISCUSSION
Prototype data acquisition unit
The graphical user interface allows the operator to enter the
equation coefficients to be used in the classification process.
It also displays a horizontal feedback bar used in the training
game in which the user attempts to mentally alter the level
according to the given task. Additional indicators are
provided to show the number of trials, the number of times
the user succeeded and failed to achieve the given task, and
the overall success rate.
RESULTS
In designing the ABCI, the goal was to have it perform
equivalent to a desktop BCI system. Therefore, the ABCI
was compared to the BCI system developed in our lab,
where the desktop BCI was considered to be the standard.
The comparison was done by simultaneously inputting a
sinusoid wave into both the BCI and the ABCI units and
comparing their respective RMS power outputs. The stages
the waveform passed through include A/D conversion, a
bandpass filter, power calculation, and outputting the
resultant value to the monitor or PDA screen. The input
waveform was swept over a range of 1 to 30 Hz and the
results indicated a difference of less than 1%. Additionally,
the time delay between a change in the input and output was
indistinguishable on both systems and is therefore less than
the quarter of a second refresh rate of the power calculation.
Overall, evaluation of the ABCI shows that the accuracy and
response time are equivalent to the desktop BCI system.
The total cost of the ABCI training system is about $350,
where the data acquisition unit cost is approximately $50
and the PDA cost is $300. Additional equipment necessary
to amplify the EEG signals would increase the cost of the
system.
The ABCI training system was designed to act as a
portable version of a deskrop BCI system. Based on the test
results, the ABCI performs equivalent to a desktop BCI.
This system can therefore be used chronically and away
from any training site, thus removing barriers inherent in
traditional BCI systems. Studies can now take advantage of
training in a more physiologic manner, where subjects have
continuous access to train and improve their cognitive
control capabilities. This brings up some immediate
questions and potential studies such as 1) the development
of optimal training protocols and providing training
motivation and 2) whether training while performing
everyday tasks is feasible.
CONCLUSION
We believe that ambulatory training is the beginning of a
new paradigm in BCI research. An ambulatory BCI that
allows training to be performed away from the traditional
training site has been presented. This should allow more
advanced training protocols to be developed and
implemented in order to more effectively study the ability of
the brain to adapt and control augmentative devices.