Patent No. 6339721 Brain wave data processing device and storage medium
Patent No. 6339721
Brain wave data processing device and storage medium (Yamazaki, et al., Jan 15, 2002)
Abstract
A wavelet conversion means (13) for subjecting brain wave data in a single trial to wavelet conversion, a wavelet coefficient window means (15) for extracting a predetermined area by a wavelet coefficient window from a wavelet coefficient surface obtained by the wavelet conversion, and a brain wave data discriminating means (17) for discriminating brain wave data by judging whether or not the predetermined area is extracted are provided to automate the work to extract distinguishing patterns from individual (in a single trial) brain wave data and to obtain averaging of the brain wave data. By considering whole of wavelet conversion coefficients, averaging and extraction of a vertex latency of distinguishing patterns can be executed without significantly losing the information of original waveform data.
Notes:
SUMMARY
OF THE INVENTION
When the averaging of the brain wave data is determined by the averaging method
and the evoked potential, such as the event-related potential (ERP), is observed,
in the past, an enormous period of time was required in order to extract distinguishing
patterns from individual brain wave data to obtain the averaging of the brain
wave data. The reason for it is that the work of extracting the patterns was
all performed by the inspection of an experimenter (or a decipherer of the brain
wave).
It is an object of the present invention to provide a method for analyzing brain
wave data which can automate the work to extract distinguishing patterns from
brain wave to reduce the load of the experimenter and improve the quality and
reliability of the brain wave data obtained as well as the efficiency of the
work for analyzing the brain wave data.
In addition, as described above, in the past, a concrete application of the
wavelet conversion for the brain wave data was not made clear, however, it is
also another object of the present invention to exhibit concrete applications.
Accordingly, the present invention exhibits the averaging of the brain wave
data as a concrete application of the wavelet conversion, and also it is an
object of the present invention to provide a device which performs the averaging
without significantly losing the information of original waveform data by considering
not only the waveform data itself as with the prior art but also whole of values
of the wavelet conversion parameters, when the results of the wavelet conversion
are compared and examined.
The brain wave data processing device according to the present invention comprises,
in a brain wave data processing device detecting distinguishing patterns from
individual brain wave data obtained in a single trial, a brain wave data storage
means for storing digital brain wave data, a wavelet conversion means for subjecting
the digital brain wave data read out from the brain wave data storage means
to wavelet conversion to determine a wavelet coefficient, a wavelet coefficient
surface output means for outputting the wavelet coefficient as function values
of a scale parameter and a shift parameter in the wavelet conversion, a wavelet
coefficient window parameter setting means for setting a wavelet coefficient
window, a wavelet coefficient window means for extracting a predetermined area
based on the wavelet coefficient window from a wavelet coefficient surface defined
by the scale parameter, shift parameter, and wavelet coefficient, and a brain
wave data discriminating means for discriminating whether or not the predetermined
area has been extracted from the wavelet coefficient surface by the wavelet
coefficient window means for individual digital brain wave data.
The brain wave data processing device according to the present invention may
further be provided with a brain wave data averaging means for averaging only
the digital brain wave data from which the predetermined area is extracted in
the wavelet coefficient surface and a pattern latency extraction means for determining
a vertex latency of distinguishing patterns included only in the digital brain
wave data from which the predetermined area is extracted in the wavelet coefficient
surface.
In the present invention, the brain wave data in which the distinguishing patterns
have been detected, for example, by inspection are previously prepared and the
corresponding wavelet coefficient surface is determined from these brain wave
data, and the wavelet coefficient window may be set according to the shape and
value of this wavelet coefficient surface. Although various types are considered
as a mother wavelet in the wavelet conversion, Mexican Hat can be exhibited
as a desirable one.
According to the present invention, the wavelet coefficient surface is the result
of subjecting the brain wave data to the wavelet conversion, and by subjecting
this wavelet coefficient surface to the wavelet coefficient window, it is discriminated
whether or not a predetermined area is extracted in the wavelet coefficient
surface, so that all the processing from the measurement of the brain wave data
to the discrimination of whether distinguishing patterns exist in the brain
wave data can be automatically performed.
Furthermore, by providing a brain wave data averaging means, all the processing
from the measurement of the brain wave data to the averaging process can be
automatically executed, and by providing a pattern latency extraction means,
a vertex latency of the extracted pattern can be automatically determined.
Comments