JP7184731B2 - Biological signal processing device, program and method for detecting bias in bilaterally acquired biological signals - Google Patents

Description

TECHNICAL FIELD The present invention relates to a technique for detecting body habits that are biased to the left or right of a human body.

2. Description of the Related Art In recent years, techniques have been developed for detecting various biological signals caused by various activities of humans and animals using sensors and using the biological data obtained by signal processing in various situations. Examples of sensors include wristwatch-type pulse sensors, earphone-type pulse sensors, headband-type electroencephalogram sensors, and the like. Also, biosignals detected by such sensors can be processed by, for example, a smartphone carried by the user and used in various applications.

2. Description of the Related Art Conventionally, there is a technique for stably detecting biological signals using a glasses-type device (see, for example, Patent Document 1). According to this technology, a myoelectric sensor that detects a myoelectric signal as a biological signal is provided, and facial expressions such as a smile or a clenched bite can be identified. Specifically, for myoelectric signals, the power value in the frequency band related to the artifact (noise signal other than the target signal) and the power value in the frequency band related to the signal at the time of a specific facial expression are calculated. Determine whether it is facial expression time. As a myoelectric sensor, the reference electrode and the detection electrode are arranged so as to contact the skin surface at some point near the cheek from the left and right (or right and left) auricles.

There is also a technique of detecting myoelectric signals by acceleration (second-order difference) in order to detect myoelectric signals from noise-like signals with many artifacts (see, for example, Patent Document 2).
Furthermore, there is also a technique for detecting an electromyographic signal of mastication as a periodic signal from noise-like signals (see Patent Document 3, for example).
Furthermore, there is also a technique in which a headband-type device is used to bring a sensor into contact with the temporalis muscle, and the change in contact pressure caused by the swelling of the muscle due to mastication is captured by the change in capacitance (see, for example, Patent Document 4). According to this technology, it is possible to obtain a detection signal related to the movement of the left and right temporal muscles of the head with respect to uneven chewing. can also be analyzed.

Furthermore, there is also a technique of detecting changes in the movement of the mandible due to mastication with a distance measuring sensor in a one-ear device (see Patent Document 5, for example). According to this technique, for uneven mastication, the waveform of the output voltage when the biological body-worn measuring device is worn on the left ear is compared with the waveform of the output voltage when the device is worn on the right ear. This makes it possible to determine the presence or absence of left and right uneven chewing and the degree of uneven chewing. For example, when the peaks (amplitudes) of the output voltage are significantly different between the left and right sides, it is determined that there is uneven mastication. Further, when the time intervals between adjacent peaks of the output voltage are significantly different between the left and right, it is determined that there is uneven mastication.
The term “mastication” refers to biting off food with the front teeth to reduce the size of the food to a size that can be easily swallowed with the back teeth so as to grind the food more finely (see, for example, Non-Patent Document 1).

JP 2019-017945 A JP 2019-107067 A JP 2019-115410 A JP-A-2018-33568 JP 2016-140478 A

Japanese Prosthodontic Society, Completion of "Clinical Guidelines for 3 Diseases in Dental Medicine", [online], [October 26, 2019], Internet <URL: http://www.hotetsu.com/s/ doc/Guidelines.pdf＞

"Unbalanced chewing" based on human chewing behavior is known as a cause of poor physical condition. This causes unilateral tooth decay, periodontal disease, temporomandibular disorders, facial distortion, and the like.
On the other hand, the inventor of the present application considered that it is important to detect a body habit that causes a bias in a biosignal obtained symmetrically from a human body.

According to the techniques described in Patent Literatures 1 to 3, the electrodes of the myoelectric sensor are arranged in the spectacle-type device, but body habits resulting in left-right symmetry cannot be detected.
In addition, according to the technology described in Patent Document 4, it is possible to analyze mastication information and mastication-related information related to uneven mastication of a diner (left-right balance of mastication), but a specific analysis method is disclosed. It has not been.
Furthermore, according to the technique described in Patent Document 5, only when the magnitude of the amplitude of the left and right output signals or the time difference between the peaks is large is regarded as eccentricity, and a temporal analysis method is not described.
According to the description of Non-Patent Document 1, there is no description regarding the evaluation of uneven chewing.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a biological signal processing apparatus, a program, and a method for detecting body habits that cause bias in biological signals obtained symmetrically using a plurality of myoelectric sensors.

According to the present invention, there is provided a biosignal processing device for determining left and right eccentric mastication using biosignals acquired from left and right electrode groups arranged symmetrically in a living body,
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
left and right body habit determination means for determining left and right uneven mastication according to left and right bias of the number of times of mastication based on the number of times of mastication.

According to another embodiment of the biological signal processing device of the present invention,
Each of the left and right electrode groups is composed of a positive electrode (detection electrode), a negative electrode (reference electrode) and a ground electrode (ground electrode), and is preferably in contact with the skin of the living body.

According to another embodiment of the biological signal processing device of the present invention,
The electrode group is arranged bilaterally symmetrically on the head of the living body,
The left biosignal analysis means and the right biosignal analysis means calculate the left biosignal and the right biosignal based on the myoelectric signal related to mastication.
is also preferred.

According to another embodiment of the biological signal processing device of the present invention,
The positive electrode is in contact with the skin position near the cheek from the left and right auricles,
The negative electrode is in contact with the skin around the nose,
It is also preferable that the ground electrode is arranged so as to be in contact with the skin position on the back of the head from around the root of the back of the ear.

According to another embodiment of the biological signal processing device of the present invention,
The electrode group is fixed to the glasses type device,
The positive electrode is arranged to contact the skin of the living body from the temple part,
The negative electrode is arranged so as to be in contact with the skin of the living body from the clings part,
It is also preferred that the ground electrode is positioned so as to contact the skin of the living body from the modern portion.

According to another embodiment of the biological signal processing device of the present invention,
The right and left strength determination means is
If the mean square value of the left side is greater than the mean square value of the right side by more than a predetermined value, it is determined as the left side,
If the mean square value of the left side is smaller than the mean square value of the right side by more than a predetermined value, it is determined as the right side,
Otherwise , it is also preferable to determine as both sides.

According to another embodiment of the biological signal processing device of the present invention,
The left/right transition recording means records the left/right transition using a tree whose node is the transition to either the left, right, or both sides, and records the transition to either the left, right, or both sides for each mouth each time mastication occurs. Traversing the nodes and counting the number of times each node is traversed,
It is also preferable that the left/right body habit determining means determines the left/right bias of the tree.

According to another embodiment of the biological signal processing device of the present invention,
The left-right transition recording means adds and records the number of occurrences and/or the dwell time to each node of the tree,
It is also preferable that the left/right body habit determining means determines the bias of the number of occurrences and/or dwell time of each node of the tree.

According to another embodiment of the biological signal processing device of the present invention,
It is preferable to further include signal counting means for adding the left representative value output from the left biosignal analysis means and the right representative value output from the right biosignal analysis means, and counting the number of occurrences of myoelectricity.

According to another embodiment of the biological signal processing device of the present invention,
The signal counting means is
A data value related to an input signal that may include a periodic biological signal is sequentially acquired, and based on the acquired data value, a lower reference value corresponding to the minimum value of the data value and an upper reference value corresponding to the maximum value of the data value are obtained. (a) the lower reference value and the next upper reference value are determined or updated, and a data value smaller than the upper reference value satisfying a predetermined condition is taken in; or (b) when the upper reference value and the subsequent lower reference value are determined or updated, and a data value larger than the lower reference value satisfying a predetermined condition is acquired. It is also preferable to count the wave number of the biological signal.

According to another embodiment of the biological signal processing device of the present invention,
The left biosignal analysis means and the right biosignal analysis means are
acceleration component generation means for generating acceleration component data of the input biological signal;
It is also preferable to include representative value calculation means for calculating a representative value relating to the state of occurrence of the biosignal in the predetermined time interval in the acceleration component data.

According to the present invention, a spectacles-type device that determines left and right mastication using biological signals acquired from left and right electrode groups arranged symmetrically on the head of a living body,
The left and right electrode groups are
A positive electrode (detection electrode) in the temple part and in contact with the skin position near the cheek from the left and right auricles,
Negative electrodes (reference electrodes) in the clings area and in contact with skin locations around the nose on the left and right sides;
It consists of a ground electrode (ground electrode) in the modern part and in contact with the skin position behind the head from around the left and right ear roots,
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
left and right body habit determination means for determining left and right uneven mastication according to left and right bias of the number of times of mastication based on the number of times of mastication.

According to the present invention, there is provided a program for causing a computer installed in a device for determining left and right mastication to function using biosignals acquired from left and right electrode groups arranged symmetrically in a living body, comprising:
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
The computer is caused to function as left/right body habit determining means for determining left/ right eccentric mastication according to left/right deviation of the number of times of mastication based on the number of times of mastication.

According to the present invention, there is provided a left and right mastication determination method for a device for determining left and right uneven mastication using biosignals acquired from left and right electrode groups arranged symmetrically in a living body, comprising:
The device
A left side biomedical signal is inputted from a group of electrodes arranged on the left side of the living body, and a left representative value relating to the state of generation of the left side biomedical signal is calculated at predetermined time intervals,
a first step of inputting a right biomedical signal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the generation of the right biomedical signal at predetermined time intervals;
A second step of calculating the mean square value for a predetermined unit time for each of the left representative value and the right representative value, and determining the left side or the right side having the larger mean square value;
a third step of counting the number of times each of the left and right sides each time a mouthful action from ingestion to swallowing occurs;
and a fourth step of judging uneven mastication to the left or right according to the bias of the number of times of mastication to the left or right based on the number of times of mastication.

According to the biological signal processing apparatus, program, and method of the present invention, a plurality of myoelectric sensors can be used to detect a body habit that causes a bias in bilaterally acquired biological signals.

1 is an external view of a wearable device according to the present invention; FIG. 1 is a functional configuration diagram of a biological signal processing device according to the present invention; FIG. 1 is a system configuration diagram for transferring a biomedical signal from a biomedical signal processing device to a mobile terminal; FIG. FIG. 2 is a functional configuration diagram in which a biological signal processing function is implemented in a mobile terminal; 4 is a functional configuration diagram of a biosignal analysis unit; FIG. FIG. 4 is an explanatory diagram showing processing of a left/right intensity determination unit and a left/right transition recording unit; FIG. 10 is an explanatory diagram for judging a left-right bias based on a numerical value; 4 is a flow chart showing processing of a signal counting unit in the present invention; 4 is a flow chart showing processing in states 1 to 3; 10 is a flowchart showing processing in state 4;

EMBODIMENT OF THE INVENTION Below, embodiment of this invention is described in detail using drawing.

FIG. 1 is an external view of a wearable device according to the present invention.

FIG. 1(a) shows an external view of a spectacles-type device according to the present invention.
The spectacles-type device 1 is a device that is attached as spectacles to the head of a living body (for example, a user who is a human being) and acquires a biomedical signal. This biomedical signal is a "myoelectric signal" as an electrical signal generated due to the movement of the intra-face part. In addition, the biological signal may be mixed with a "(noise) signal caused by electrode misalignment" generated by movement.

Here, for example, mastication is an example of the motion of the intra-face part to be detected. The spectacles-type device 1 of the present invention can detect myoelectric signals associated with "chewing", which are "periodic biological signals" that have temporal periodicity due to repetitive motions. Muscle activity related to "chewing" occurs simultaneously on both the left and right sides, and muscle activity on the left or right side may be stronger than muscle activity on the other side. For this purpose, a set of electrodes is arranged separately on the left side and the right side, the myoelectric signals generated by the left and right muscles are captured respectively, and the myoelectric signals generated simultaneously on the left and right sides are added together to obtain a stable and large electromyographic signal. get the number.

According to FIG. 1( a ), the spectacles-type device 1 includes a biological signal processing device 10 , a temple section 11 as a spectacles structure, a clings section 12 , and a modern section 13 .
In addition, the electrode groups for acquiring biosignals are arranged symmetrically on the left and right sides of the median line when the glasses-type device is worn on the head of the living body. This makes it possible to capture the activity of the same type of muscle that exists on each side.
(1) The positive electrode (detection electrode) is arranged from the temple portion 11 so as to be in contact with the skin position from the vicinity of the auricle to the vicinity of the cheek via the elastic support portion.
(2) A negative electrode (reference electrode) is arranged from the clings part 12 so as to be in contact with the skin around the nose.
(3) The ground electrode (ground electrode) is arranged from the modern part 13 so as to come into contact with the skin position behind the head from the vicinity of the root of the back of the ear.
A biological signal is detected as a potential difference between the positive electrode and the negative electrode.

The ground electrode is for noise cancellation and may be a DRL (Driven Right Leg) electrode that reduces common mode noise caused by a commercial power supply or the like.
Also, the ground electrodes may be short-circuited without being divided into left and right sides. Furthermore, the ground electrode may be placed anywhere, and may be placed in common with the negative electrode (near the nose), although the detection of myoelectric signals will be weaker.

The spectacles-type device 1 is stably supported by the skin of the head at the positions where the electrodes are arranged. Human cheekbones protrude sideways when viewed from the front of the face, but if the positive electrode, for example, is brought into contact with the skin slightly above the widest part of the cheekbone, the distance between the left and right "electrode parts" can be reduced. It is narrower than the maximum width of the cheekbone and gets caught in the wider part of the upper cheekbone.

Each electrode can maintain stable contact by contacting the skin of the living body via the elastic supporting portion, even if the head moves greatly, for example.
The positive electrode is placed on the skin below the temple and slightly above the widest point of the cheekbones when the face is viewed from the front.
In addition, the position adjustment unit adjusts the position of the elastic support part of the temple part so that the positive electrode can contact the skin at any position within the range from the upper cheek to the base of the ear via the temple. can do.

FIG. 1(b) shows an external view of a hair band type device according to the present invention.

The hairband type device 1 is a device that is wrapped around the head of a living body to acquire biological signals. In FIG. 1(b), similarly to FIG. 1(a), electrode groups for acquiring biosignals are arranged symmetrically on the left and right sides of the midline.
(1) The positive electrode is arranged so as to come into contact with the skin from the vicinity of the auricle to the vicinity of the cheek.
(2) The negative electrode is placed in contact with the skin around the upper end of the nose (between the eyebrows).
(3) The ground electrode is arranged so as to be in contact with the skin position behind the head from the vicinity of the root of the back of the ear.
The biological signal acquisition device 10 is also mounted on a hair band, and transmits myoelectric signals received from each electrode to the mobile terminal 2 .

The biomedical signal processing device 10 transmits myoelectric signals to the mobile terminal 2 wirelessly or by wire. The mobile terminal 2 may be another information processing device such as a smart phone, a mobile phone, a PDA (Personal Digital Assistant), a tablet computer, or a personal computer.
When wireless communication is performed between the biological signal processing device 10 and the mobile terminal 2, a wireless LAN such as Bluetooth (registered trademark) or Wi-Fi (registered trademark) may be used. . In the case of wired communication, for example, it may be connected to an analog audio input/output terminal (jack) for a headphone/microphone of the mobile terminal 2, or may be connected via a USB (Universal Serial Bus). may

FIG. 2 is a functional configuration diagram of the biological signal processing device according to the present invention.

The biological signal processing apparatus 10 receives biological signals from each electrode via conductive paths along the frame of the glasses-type device 1 . According to FIG. 2, the biomedical signal processing device 10 is arranged outside the left temple, but it can be incorporated in the temple so that the entire spectacles-type device 1 can be designed so as not to differ greatly in appearance from ordinary spectacles. is also preferred. The biological signal processing apparatus 10 incorporates a driving battery, drives a computer with electric power supplied from the battery, and executes processing of the biological signal.
Of course, the biological signal processing device 10 may be divided into left and right, and arranged with approximately (almost) the same weight. The right and left balance of the weight of the spectacles type device 1 can be balanced, and a good wearing feeling without bias can be realized.

As another embodiment, when the positive electrode is in contact with the skin at a position within the range from the upper cheek to the base of the ear via the temple, the biosignal that can be obtained is not limited to myoelectric signals. .
For example, in addition to biosignals based on biopotentials such as electroencephalograms and electrooculography signals that can be detected from a position near the ear, signals related to body temperature and perspiration (although sensor devices other than biopotential sensors are required), A pulse wave or the like can also be detected.

The biomedical signal processing device 10 uses biomedical signals acquired from the left and right electrode groups arranged symmetrically on the living body to determine the body habit of left and right bias.
According to FIG. 2, the biosignal processing device 10 includes a left-side signal conversion unit 100a and a right-side signal conversion unit 100b, a left-side biosignal analysis unit 101a and a right-side biosignal analysis unit 101b, a left-right intensity determination unit 102, and a left-right transition. It has a recording unit 103 , a left/right body habit determination unit 104 , and a signal counting unit 105 . These functional components are implemented by executing a program that causes a computer installed in the device to function. In addition, the flow of processing by these functional components can also be understood as a biological signal processing method.

FIG. 3 is a system configuration diagram for transferring a biomedical signal from a biomedical signal processing device to a mobile terminal.
According to FIG. 3, the biosignal processing device 10 generates biosignals (potential difference over time) from the positive electrode, the negative electrode and the ground electrode on the left side, and the biosignals from the positive electrode, the negative electrode and the ground electrode on the left side. are converted into digital signals by the left signal conversion section 100a and the right signal conversion section 100b and transferred to the portable terminal 2. FIG. This digital signal is transmitted to the mobile terminal 2 by radio such as Bluetooth (registered trademark). Actual biological signal processing is performed by the mobile terminal 2 .

[Left signal converter 100a, right signal converter 100b]
The signal converter 100 amplifies the AC component of the potential difference between the positive electrode and the negative electrode by differentially amplifying the potential of the ground electrode, and generates one digital signal from this analog biological signal at a constant sampling frequency. This differential amplification may use a DRL circuit for reducing common mode noise caused by a commercial power supply or the like.

This enables skin potential detection in the range of plus or minus 0.1 to several hundred microvolts, for example. As a condition for this digitization, it is also preferable to perform analog/digital (A/D) conversion with a sampling frequency of 500 Hz or higher and a quantization of 10 bits or higher. Such a circuit configuration can be realized using, for example, TGAM1 manufactured by Neurosky.

FIG. 4 is a functional configuration diagram in which a biological signal processing function is implemented in a mobile terminal.
Each functional component of the portable terminal 2 in FIG. 4 is the same as the biological signal processing device in FIG. The left-right body habit information determined by the mobile terminal 2 can also be transmitted to the analysis server 3 .

<Left biosignal analysis unit 101a, right biosignal analysis unit 101b>
The left biosignal analysis unit 101a receives the left biosignal from the electrode group arranged on the left side of the living body, and calculates a left representative value relating to how the left biosignal is generated at predetermined time intervals.
Similarly, the right biomedical signal analysis unit 101b also receives a right biomedical signal from the electrode group arranged on the right side of the living body, and calculates a right representative value relating to the generation of the right biomedical signal at predetermined time intervals.
Both the left biosignal analysis unit 101a and the right biosignal analysis unit 101b perform the same processing (hereinafter commonly referred to as the "biological signal analysis unit 101").

FIG. 5 is a functional configuration diagram of the biosignal analysis unit.

According to FIG. 5 , the biological signal analysis unit 101 has a prefiltering unit 1011 , an acceleration component generation unit 1012 and a representative value calculation unit 1013 .

[Pre-filtering unit 1011]
The pre-filtering unit 1011 performs band-elimination filtering on the input signal before the acceleration component data is generated to reduce periodic noise associated with the commercial power source (which is often mixed). For example, TGAM1 manufactured by Neurosky has a notch filter for reducing noise derived from a commercial power supply.

[Acceleration component generator 1012]
The acceleration component generator 1012 generates acceleration component data of the input biological signal.
Specifically, the acceleration component generation unit 1012 has a second-order difference filter, and performs difference filter processing on the input signal twice. A difference equation showing the principle of the difference filter used here is as follows.
y[n] = x[n] - x[n-1]
n: sample position (sample index)
x[n]: input signal at sample position n
y[n]: output signal at sample position n

In general, when using a digital filter like the acceleration component generator 1012, the more advanced the digital filter, the greater the amount of calculation. This increased computational complexity results in reduced battery life in mobile devices such as wearable devices. On the other hand, the acceleration component generator 1012 does not use a filter including a trigonometric function, for example, and uses a filter with a small order to process the biosignal, so it is possible to suppress an increase in the amount of calculation.

[Representative value calculation unit 1013]
The representative value calculation unit 1013 calculates a representative value relating to the state of occurrence of the biological signal in a predetermined time interval (window analysis interval) in the acceleration component data.

In general, time-series data output from a biosensor can be fed back to the user in real time via a user interface by sequentially analyzing it in real time. At this time, by dividing the data into predetermined window analysis intervals and sequentially performing analysis while shifting the analysis intervals, substantially real-time analysis processing becomes possible.

For example, when the sampling frequency for digitization in the signal conversion unit 1011 is 512 Hz, the representative value calculation unit 1013 performs window analysis on the most recently input 128 samples every time 64 samples of the time-series data of the acceleration component are input. Calculate the standard deviation SD as an interval.
Further, according to another embodiment, the window analysis interval is similarly set to 128 samples, and the time-series data of the acceleration component is divided every 0.25 seconds (every 128 samples). A standard deviation SD may be calculated.

Note that the value calculated here is not limited to the standard deviation SD, and various values can be adopted as long as they are values related to the degree of deviation of the acceleration component data in the window analysis interval.

In addition, the representative value calculation unit 1013 calculates a weight W that is a monotonically decreasing function for the length of the time interval (sample number length len_th) in which the acceleration component in each window analysis interval remains continuously within a predetermined range. .
The inventors of the present application have found that this time interval is significantly longer when no myoelectric signal is present compared to when it is present.
Therefore, a "weight W" is determined that rapidly decreases as this time interval lengthens (or at least is a monotonically decreasing function for this time interval). By including such a characteristic in the representative value SD _W , it is possible to more reliably avoid erroneous determination of myoelectric signal generation when there is no myoelectric signal or when there is only noise. In other words, noise such as electrode misalignment is removed before SDw is calculated.

Here, first, in the acceleration component data generated by the acceleration component generation unit 1012, the amplitude of the acceleration component is scanned from the beginning of the window analysis interval for weight calculation, and the amplitude less than the preset threshold th continues. Determine the number of samples length len_th. Also, the number of observed samples obs that defines the noise interval is set in advance. For example, we can set th=10 and obs=15.

Incidentally, when there are a plurality of continuous amplitude intervals below the threshold th within the window analysis interval, the sample number length len_th may be the total number of samples in those intervals. Alternatively, the number of samples in the amplitude continuous section with the longest time section can be set to the sample number length len_th.

The exponential weight W as a function of the sample number length len_th is represented by the following formula.
W = exp(1-len_th/obs)
For example, when obs=15 and Len_th=15, W=1.0.
Furthermore, as len_th increases, the exponential weight W decreases sharply and approaches zero asymptotically. In fact, when len_th corresponds to the window analysis interval length (128 samples), the exponential weight W becomes approximately zero.

On the other hand, similarly for obs=15, the inverse proportional weight W as a function of the sample number length len_th is represented by the following formula.
W = 1 / ((len_th - obs) / a + 1)
Here, a preferably takes a value exceeding obs (a>obs) to make the denominator a positive value.
The inverse weight W is for a=obs*obs (=225), and if Len_th=15, W=1.0. Also, as len_th increases, the inverse weight W decreases and approaches zero.

Of course, the weight W is not limited to the one explained above. A monotonically decreasing function of len_th can be used as the weight W. Preferably, a function that approaches zero as len_th increases, more preferably a function that approaches zero asymptotically, can be used as the weight W. be able to. Here, for example, the weight W may be a linear function of len_th having a negative slope, but the exponential weight W contributes to more reliable biological signal generation determination.

Finally, the representative value calculation unit 1013 calculates a representative value SD _W by weighting the calculated standard deviation SD with the similarly calculated weight W as shown in the following equation.
SD _W [k] = W [k] * SD [k]
k: window position (window index)

FIG. 6 is an explanatory diagram showing the processing of the left/right intensity determination section and the left/right transition recording section.

[Right/left strength determination unit 102]
The left/right intensity determination unit 102 determines one side (left side, right side, or both sides) where a strong biosignal is generated according to the difference between the left representative value and the right representative value.

Specifically, the left/right strength determination unit 102 calculates as follows.
(S1) The mean square value of the predetermined unit time Tlr is calculated from the left representative value, and the mean square value of the predetermined unit time Tlr is calculated from the right representative value. The predetermined unit time Tlr is, for example, approximately 1 to 2 seconds.
Root mean value = 1/(Tlr*Ssdw) Σ(SDw[idx] ² )
SDw: typical value
Ssdw[sample/sec]: Number of samples per second of SDw
idx: Sample position of SDw (up to past Tlr x Ssdw, with the most recent as 0).
Number of SDw input in Tlr time (number of samples) = Tlr x Ssdw

(S2) If the mean square value of the left side is greater than the mean square value of the right side by exceeding the determination threshold value Zlr, the left side is determined. If the mean square value of the left side is smaller than the mean square value of the right side by exceeding the determination threshold value Zlr, the right side is determined.
Dlr = Left Root Mean Square - Right Root Mean Square

In some cases, the muscles of the left side, the right side, or both of the muscles are used for mastication, and a certain determination threshold value Zlr is set in advance in order to prevent left-right determination from fluttering left and right.
Dlr>Zlr: Judged as chewing on the left side
Dlr<-Zlr: Judged as chewing on the right side
Zlr ≧ Dlr ≧ -Zlr : Judged as chewing on both sides

According to FIG. 6, the left-right strength determination unit 102 counts a predetermined action (one chewing) notified from the signal counting unit 105 during a predetermined action (one bite from eating to swallowing). Determine left, right, and both sides. The determination result is output to the left/right transition recording unit 103 . Since the period of one bite from ingestion to swallowing is unknown in this system, the periodical signal generation period of the signal counting unit 105 is set as a counting period.

[Left-right transition recording unit 103]
For example, one mouthful from ingestion to swallowing is defined as one predetermined action. In addition, one mouthful of predetermined action includes mastication as a predetermined action of one or more times.
At this time, the left-right transition recording unit 103 records the transition of either the left side, the right side, or the both sides for each predetermined action (chewing) each time a predetermined action (one bite from eating to swallowing) occurs. .
Since body habits that are biased to the left or right cannot be determined not only by one prescribed action but also by one prescribed action, it is necessary to record the prescribed action repeatedly a plurality of times.

According to FIG. 6, two recording methods for the left-right transition recording unit 103 will be described below.
<Recording method 1>
The left-right transition recording unit 103 records the left-right transition for each predetermined action (chewing) using a tree whose nodes are transitions to either the left side, the right side, or both sides.
The left-right transition recording unit 103 traces the node toward the lower stage, triggered by the transition of each predetermined action to either the left side, right side, or both sides, each time a predetermined action of the living body occurs. Then, each time each node of the tree is traced, 1 may be added and recorded.
In addition, each node in the tree is not limited to the number of times that each node was traced, but the number of times that a predetermined action occurred (the number of times of mastication) was added, and the number of times that the dwell time continued on either the left side, the right side, or both sides was added and recorded. The sum of the chewing strengths (root mean square of SDw, left, right) at that node may also be recorded. Of course, they may be recorded in combination.

According to recording method 1 in FIG. 6, eight mouths are repeated from ingestion to swallowing. That is, each tree is merged and the number of times each node is traced is counted.
Specifically, immediately after eating, there were 3 mouths where mastication was started on the left side, and there were 5 mouths where mastication was started on the right side. From this, it can be understood that chewing often begins on the right side immediately after eating.
In addition, it can be seen that of the 5 mouths that were masticated on the right side immediately after eating, 4 mouths were masticated on the left side, and 1 mouth was masticated on both sides. From this, it can be understood that, immediately after eating, the right side starts chewing and then the left side often chews.
By recording in the tree in this way, it is possible to know the frequent pattern by counting the pattern traced for each mastication from a plurality of mouth counts.

As in recording method 1, if a tree is constructed by transitioning nodes that determine left and right for each mastication, the number of mastications for a single mouthful is several tens of times, so the depth of the tree is the maximum number of mastications. It becomes uselessly deep. As a result, there arises a problem that it is difficult to see the overall left/right bias and it is difficult to judge uneven mastication.

For this reason, the left-right transition recording unit 103 records the left-right transition only when the left-right transition occurs a predetermined number of times in order to transition the node of the tree to either the left-hand side, the right-hand side, or both sides. can be anything. That is, only when chewing occurs 3 or more times on the left (or right), the nodes of the tree are traced to the left (or right) and recorded, and when chewing 3 or more times does not occur on the left (or right) may record the nodes of the tree as both sides. Consecutive sides are aggregated as one node.

<Recording Method 2>
According to the recording method 2, the left-right transition recording unit 103 has a table structure of three vertical columns, each column has the meaning of left side, right side, and both sides, and one mastication is defined in the row direction. For each predetermined action (one mouthful from ingestion to swallowing), 1 is added each time it is determined whether the chewing is performed on the left side, the right side, or both sides.
Also, the row may be set to a predetermined unit time Tlr, instead of being set to each mastication, and may be added for each left/right intensity determination result for each Tlr. Compared with recording method 1, by comparing the total value of each column of the table, it is possible to determine the uneven chewing. You can visualize the body habits of

[Left and right body habit determination unit 104]
The left/right body habit determination unit 104 determines left/right biased body habits according to transition records for either the left side, the right side, or both sides based on a plurality of predetermined actions (one bite from eating to swallowing). The determination result is output to an application that executes various processes based on body habits.

FIG. 7 is an explanatory diagram for judging left-right bias based on numerical values.

For example, the left-right bias can be determined by the following numerical values.
(1) Add up the numerical values of all the nodes in the tree, left (eg 1.0), right (eg -1.0), and both sides (eg 0.0), and calculate the average centroid. The closer the center of gravity is to 1.0, the more it is determined to be leftward, and the closer it is to -1.0, the more it is determined to be rightward.
(2) Calculate the center of gravity, which is the average value, by totaling the numerical values of the nodes for each stage of the tree, with the left side (eg 1.0), the right side (eg -1.0), and both sides (eg 0.0). The closer the center of gravity is to 1.0, the more it is determined to be leftward, and the closer it is to -1.0, the more it is determined to be rightward. Arrange the calculated center-of-gravity values for each stage from the top.
(3) The route (most frequent route) that has been traced the most in the node transitions of the tree is set as the body habit pattern.
(4) Representative values obtained from the tree structure, such as the shortest route length with the fewest number of steps traced, the longest route length with the largest number of steps traced, and the average root length of the number of traced steps in the node transition of the tree. Make it a habit.

According to the recording method 1 of FIG. 7, the left-right body habit determination unit 104 may determine the bias of the number of times each node of the tree has been passed.
Σ _{Each node} ^{All nodes} (left side 1.0 or right side −1.0×number of passages)/total number of passages at all nodes Here, the numerical value is 0.071, and it is determined that the chewing is biased to the left.

According to recording method 2 in FIG. 7, numerical values are calculated for all predetermined actions (chewing).
Σ(left side 1.0 or right side −1.0×predetermined number of motions)/predetermined number of motions (chewing) Here, the numerical value is 0.224=(29−18)/(29+18+2), which is determined as uneven chewing to the left. .

[Signal counting unit 105]
The signal counting unit 105 can count periodic biosignals without resorting to frequency analysis.
The signal counting unit 105 adds the left representative value output from the left biosignal analysis means and the right representative value output from the right biosignal analysis means, and counts the number of occurrences of myoelectricity. By adding, the power saving signals on both sides generated at the same time on both sides can be largely captured.

Specifically, the signal counting unit 105 performs processing as follows.
(A) Sequentially capture data values related to an input signal that may include a periodic biological signal, and based on the captured data values, correspond to a “lower reference value” corresponding to a minimum value of data values and a maximum value of data values. Sequentially determine or update the "upper reference value" to
(B) The “lower reference value” and the next “upper reference value” are determined or updated, and a data value smaller than the “upper reference value” satisfying a predetermined condition (for example, exceeding the set threshold value) is captured. when
(C) Count the number of waves of the biological signal.

By the way, instead of (B) above,
(B') The "upper reference value" and the next "lower reference value" are determined or updated, and a data value larger than the "lower reference value" satisfying a predetermined condition (for example, exceeding the set threshold value) is found. when taken in,
may be adopted. In any case, by sequentially determining or updating the “lower reference value” and the “upper reference value” as described above and using them appropriately, the periodic biological signal can be counted without relying on frequency analysis. can do. That is, counting can be performed with less processing load than frequency analysis processing that requires a large amount of calculation.

In addition, the strength (amplitude) of biological signals generally varies greatly depending on the individual differences of the signal source, such as a person or animal, and the environment in which they are located. there is On the other hand, the "lower reference value" and the "upper reference value" described above are dynamically determined and updated based on the data values that are sequentially acquired. Counting processing is performed using a reference value that dynamically changes according to the As a result, it becomes possible to carry out counting processing of periodic biological signals, which conventionally had a large error, with higher accuracy.

Incidentally, the counting process by the signal counting unit 105 described above can be applied to various signals as long as they have temporal periodicity. However, it is a particularly suitable processing method for biosignals that have the characteristics described above (which are usually unstable).

Also, as the “data value” that the signal counting unit 105 takes in, various values can be adopted as long as they relate to an input signal that can include a periodic biological signal. In this respect, the inventors of the present application have proposed that when dealing with myoelectric signals related to mastication as periodic biosignals, the "representative We have found that it is highly preferable to employ "value" for "data value". Here, for example, the standard deviation SD may be used as this “representative value”, but as will be described later, it is more preferable to adopt a value SD _W weighted with respect to the standard deviation SD.

In fact, biological signals such as myoelectric signals are generated due to, for example, the activities of many cells, unlike vibrations caused by artificial machines, so they are originally AC signals having a wide frequency range. . By adopting the above-mentioned "representative value" in performing such counting processing of biological signals, it is possible to avoid frequency analysis, which requires a large amount of calculation, and to be robust against the generation of noise due to, for example, electrode misalignment. processing can be realized.

Furthermore, the inventors of the present application have reliably identified large-amplitude artifacts that are mixed in by "chewing," which is of particular interest, through the biosignal counting process described above, and have found robustness against such artifacts. have also found that it is possible to improve

By reliably determining whether or not "chewing" has occurred, it is possible to robustly handle large-amplitude artifacts that may be mixed.

The signal counting unit 105 has <rising phase detection state>, <upper threshold determination state>, <falling phase detection state>, and <lower threshold determination state>, and sequentially shifts these four detection/determination states. At the same time, the representative value _SDW data that is sequentially taken in is processed, and the wave number of the biosignal is counted. In this embodiment, the representative value SD _W data is not subjected to resonator filtering in order to minimize the occurrence of erroneous counting due to artifacts.

Furthermore, according to the present embodiment, the periodic biosignals to be counted are obtained from the glasses-type device 1 worn on the head, so that they are determined to be myoelectric signals caused by "chewing". be done. That is, the signal counting unit 105 counts the number of occurrences of myoelectric signals related to mastication (the number of times of mastication) in this embodiment.

Next, an outline of the processing contents in the above four detection/determination states will be described. Here, it is assumed that the data value (representative value SD _W in this embodiment) is taken into the signal counting unit 105 moment by moment at predetermined time intervals.

<rising phase detection state>
(a) If the captured representative value SD _W (data value) is less than or equal to the value at the previous point in time, determine this representative value SD _W as the “temporary lower reference value”,
(b) If the representative value SD _W then taken in is greater than the "temporary lower reference value", the "temporary lower reference value" is determined as the "lower reference value", and the next upper threshold judgment state is entered. It is in the detection state to transition.

<Upper threshold judgment state>
(C) Determine whether the captured representative value SD _W exceeds the predetermined "upper threshold value" when viewed from the "lower reference value", and determine true, that is, When it is determined that it exceeds the value, it is in a determination state in which the transition to the next falling phase detection state is made.

<Descent Phase Detection State> is
(d) If the captured representative value SD _W is equal to or greater than the value at the previous time point, determine this representative value SD _W as the “temporary upper reference value”,
(e) If the representative value SD _W then taken in is smaller than the "temporary upper reference value", the "temporary upper reference value" is determined as the "upper reference value", and the next lower threshold judgment state It is in the detection state that shifts to

<Lower threshold judgment state>
(f) Determine whether the captured representative value SD _W is a value below the predetermined "lower threshold value" when viewed from the "upper reference value", and determine true, that is, When it is determined that the value has fallen below, it is in a determination state of shifting to the next rising phase detection state.

Signal counting unit 105, each time one set of processing consisting of <rising phase detection state>, <upper threshold determination state>, <falling phase detection state>, and <lower threshold determination state> described above is completed Then, the number of waves of the biological signal is counted. for example,
(a) A series of processes starting with <rising phase detection state> and ending with <lower threshold determination state>,
(b) A series of processes starting with <upper threshold determination state> and ending with <rising phase detection state>,
(c) A series of processes starting with <falling phase detection state> and ending with <upper threshold determination state>, or (d) a series of processing starting with <lower side threshold determination state> and ending with <falling phase detection state> Each time it is completed, the wavenumber of the biosignal can be incremented by one.

The “upper threshold value” in (c) above is determined based on the standard deviation of the representative value SD _W for a predetermined period, and the “lower threshold value” in (f) above is determined from the representative value SD _W It is also preferable that it is determined based on "reference value".

Here, especially when determining the "upper threshold", if the input signal has a bias component, the average value of the bias component is subtracted from the input signal, and the representative value SD _W (input signal data value) fluctuates around the zero line. It is also preferable to perform monitoring for a predetermined period of time after adjusting so as to obtain the "upper threshold value" by calculating the standard deviation.

FIG. 8 is a flow chart showing the processing of the signal counting section in the present invention.

Here, the periodic biosignal is counted based on the time-series data of the representative value SD _W calculated in the above steps and the biosignal generation time interval. Incidentally, hereinafter, the <rising phase detection state>, the <upper threshold determination state>, the <falling phase detection state>, and the <lower threshold determination state> are referred to as states 1, 2, 3, and 4, respectively.

(S301, S302) The representative value SD _W data is sequentially acquired, and when the acquired representative value SD _W is a value related to the biological signal generation time interval, the state to which the representative value SD _W corresponds (out of states 1 to 4) one of the above). The details of the processing in each state will be described in detail later with reference to FIGS. 9 and 11. FIG.
(S321) On the other hand, in step S301, if the captured representative value SD _W is not a value related to the biological signal generation time interval, it is determined that the biological signal is not generated at the present time, and the captured representative value SD _W is Alternatively, the biological signal processing is terminated for a predetermined period thereafter.

(S303) Whether or not a periodic biosignal is generated based on whether or not the transition from state 1 to state 4 progresses repeatedly during the execution of processing according to the state corresponding to the captured representative value _SDW . determine whether or not Specifically, thereafter, by performing states 1 to 4 described with reference to FIGS. 9 and 10, a periodic biosignal is generated depending on whether or not counting processing (at least a predetermined amount) is performed within a predetermined period. It can be determined whether or not

(S304) The count number CNT is incremented by 1 each time the count is performed (when it is determined that a periodic biological signal is occurring), and the periodic biological signal is counted (occurrence of a periodic biological phenomenon). number counting). Note that the steps described above are performed each time the representative value SD _W calculated for each window analysis interval is captured, and are repeatedly performed during the counting target period or a predetermined number of times.

(S311) On the other hand, when it is determined in step S303 that the periodic biosignal is not generated even though it is the biosignal generation time interval, the biosignal is counted (the number of occurrences of the biophenomenon is counted). This counting process will also be described later.

FIG. 9 is a flow chart showing the processing of states 1-3.
FIG. 10 is a flow chart showing the processing in state 4. As shown in FIG.

First, using the flowchart of FIG. 9A, the details of the processing performed in state 1 (increasing phase detection state) will be described.
(S401) It is determined whether or not the representative value SD _W fetched this time is greater than the "temporary lower reference value". Here, the “temporary lower reference value” is the representative value SD _W (which is the value determined in step S406 or step S708 in FIG. 9) at the time immediately before (immediately before), or (initially) set in advance can be the default value.

(S402) If a true determination is made in step S401 (determination that SD _W >"temporary lower reference value"), it is determined that the current "temporary lower reference value" is in an upward phase, and the current "temporary lower reference value" is set to " Decide on “lower reference value”.
(S403) Time point information related to the representative value SD _W fetched this time, for example, the sample number, which is the temporal position of the representative value SD _W , is determined as the initial point point position CP of the state cycle.
(S404) It is determined that the representative value SD _W to be taken in next time (at the point after one) is to be shifted to the next "state 2".

(S405) On the other hand, if a false determination is made in step S401 (determination that SD _W ≤ "temporary lower reference value"), it is determined that the situation is in a downward phase, and the representative value SD _W to be taken in next time is also determined to be " Decide to continue in state 1.
(S406) The "temporary lower reference value" to be used in the next "state 1" is updated to the representative value _SDW taken in this time.

Next, the contents of processing performed in state 2 (upper threshold determination state) will be described using the flowchart of FIG. 9B.
(S501) The difference between the sample number of the representative value SD _W taken in this time and the initial time point position CP, that is, the current sample number (elapsed time) from the initial time point position CP is equal to the predetermined P length threshold (the number of sample points). Long threshold) Determine whether the value exceeds ThP.

(S502) If a false determination is made in step S501 (determination that "the number of samples elapsed from CP" ≤ ThP), the representative value SD _W taken in this time is, viewed from the current "lower reference value", It is determined whether or not the value exceeds a predetermined "upper threshold value". That is, the following formula (8) SD _W >"lower reference value" + "upper threshold value"
is established. Here, as the "upper threshold value", for example, a value n times the standard deviation (for example, n=2) in the distribution of the representative value SD _W in the past predetermined period can be adopted.

(S503) If a true determination (determination that the above equation (8) is established) is made in step S502, it is assumed that the "mountain" of the signal is detected, and the representative value SD _W to be taken in next time is determined as follows: It decides to move to state 3”.
(S506) The "temporary upper reference value" to be used in the next "state 3" is determined and updated to the representative value _SDW taken in this time.

(S504) On the other hand, if a false determination is made in step S502 (determination that the above equation (8) does not hold), it is determined that the "mountain" of the signal is not yet detected, and the next representative value SD _W also decides to continue in "state 2".

(S505) Furthermore, if a true determination is made in step S501 (determination that "the number of samples elapsed from CP">ThP), the retention in the state up to the present time is long and the periodic It determines that it is not a signal, resets the signal count number CNT to zero, and decides to return to "state 1" for the representative value SD _W to be taken in next time.

Next, using the flowchart of FIG. 9(C), the details of the processing performed in state 3 (descending phase detection state) will be described.
(S601) The difference between the sample number of the representative value SD _W taken in this time and the initial time point position CP, that is, the number of elapsed samples (elapsed time) at the present time from the initial time point position CP exceeds a predetermined P length threshold ThP. It is determined whether or not the value is

(S602) If a false determination is made in step S601 (determination that "the number of samples elapsed from CP" ≤ ThP), whether or not the representative value SD _W taken in this time is smaller than the "temporary upper reference value" judge. Here, the "temporary upper reference value" can be the representative value SD _W (which is the value determined in step S606 or step S506) at the previous time point (immediate time point).

(S603) If a true determination is made in step S602 (determination that SD _W <"temporary upper reference value"), it is determined that the current "temporary upper reference value" is in a downward phase, and the current "temporary upper reference value" is set to " “Upper reference value”.
(S604) It is determined that the representative value SD _W to be taken in next time is to be shifted to the next "state 4".

(S605) On the other hand, if a false determination is made in step S602 (determination that SD _W ≥ "temporary upper reference value"), it is determined that the phase is in an upward trend, and the representative value SD _W to be taken in next time is also " Decide to continue in state 3.
(S606) The "temporary upper reference value" to be used in the next "state 3" is updated to the representative value _SDW taken in this time.

(S607) Furthermore, when a true determination is made in step S601 (determination that "the number of samples elapsed from CP">ThP), the retention in the state up to the present time is long, and the periodic It determines that it is not a signal, resets the signal count number CNT to zero, and decides to return to "state 1" for the representative value SD _W to be taken in next time.

Finally, the details of the processing performed in state 4 (lower threshold determination state) will be described with reference to the flowchart of FIG. 10 .
(S701) The difference between the sample number of the representative value SD _W taken in this time and the initial time point position CP, that is, the current number of elapsed samples (elapsed time) from the initial time point position CP exceeds a predetermined P length threshold ThP. It is determined whether or not the value is

(S702) If a false determination is made in step S701 (determination that "the number of samples elapsed from CP" ≤ ThP), the representative value SD _W taken in this time is It is determined whether or not the value is lower than a predetermined "lower threshold value". (9) SD _W <"upper reference value" - "lower threshold value"
is established. Here, as the "lower threshold value", for example, a value that is m times the peak value (0<m<1, for example m=0.9) of the current "upper reference value" (peak value) can be adopted.

(S703) If a true determination (determination that the above equation (9) holds true) is made in step S702, the signal is counted as a "mountain" followed by a "trough" of the signal. , Here, it is determined whether or not the current counting process is the first time, that is, whether or not the current signal count number CNT is a positive value (positive integer value).
(S704) If a false determination (CNT=0) is made in step S703, the signal count number CNT is incremented by 1 because this counting process is the first time.

(S705) On the other hand, if a true determination (CNT>0) is made in step S703, the current counting process is the second or subsequent time, so here it is checked again whether or not the signal is a periodic signal to be counted. Confirm. Specifically, the difference between the sample number of the representative value SD _W that was taken in when the previous counting process was performed and the sample number of the representative value SD _W that was taken in this time and this time point position PI, That is, it is determined whether or not the number of elapsed samples (elapsed time) at the present time from the time point PI exceeds a predetermined P length threshold value ThPL.

Here, if a true determination is made in step S705 (determination that "the number of samples elapsed from PL">ThPL), it is determined that the periodic signal to be counted is captured, and the process proceeds to step S704 to perform counting processing. I do. On the other hand, if a false determination is made in step S705 (determination that "the number of elapsed samples from PL" ≤ ThPL), it is assumed that a pulse-shaped signal that is not to be counted is generated at short time intervals, and step S704 is skipped, and the process proceeds to step S706.

(S706) The sample number, which is the temporal position of the representative value SD _W fetched this time, is determined as the time position PI.
(S707) It is determined that the representative value SD _W to be taken in next time (at the point after one) is to be shifted to the next "state 1".
(S708) The "temporary lower reference value" used in the next "state 1" is determined and updated to the representative value _SDW taken in this time.

(S711) On the other hand, if a false determination is made in step S702 (determination that the above equation (9) is not established), it is determined whether the representative value SD _W taken in this time is greater than the "upper reference value". judge. Here, when a true determination (SD _W >“upper reference value”) is made, the “upper reference value” is updated to this representative value SD _W (S712). On the other hand, if the determination is false (SD _W ≦“upper reference value”), step S712 is skipped and the process proceeds to step S713.
(S713) Assuming that the "trough" of the signal has not yet been detected, it is determined to continue "state 4" for the representative value SD _W to be taken in next time.

(S721) Furthermore, when a true determination is made in step S701 (determination that "the number of samples elapsed from CP">ThP), the residence in the state up to the present time is long, and the periodic It determines that it is not a signal, resets the signal count number CNT to zero, and decides to return to "state 1" for the representative value SD _W to be taken in next time.

According to the biological signal processing of the present embodiment, four states are regarded as one cycle, and waveforms (pulses) are sequentially analyzed every time the representative value SD _W is taken in, for example, "chewing", "heartbeat", "pulse ”, “breathing”, “walking”, and other periodic biosignal waves, i.e., “mastication rate”, “heart rate”, “pulse rate”, “breathing rate”, “stepping rate”, etc., are subjected to frequency analysis. It is possible to count more reliably without

Here, especially in this embodiment, the "upper reference value" and the "lower reference value" are not fixed values set in advance, but are dynamically calculated each time by representative values SD _W (input signal data values) that are sequentially taken in. have decided to Furthermore, the "upper threshold value" and the "lower threshold value" are also appropriately dynamically determined based on the representative value SD _W (input signal data value). As a result, even for biological signals that generally have large amplitude fluctuations and tend to be mixed with noise that is not subject to counting, the waveform and wavenumber can be captured more appropriately, and signal counting processing with higher accuracy, that is, with smaller counting errors, can be performed. It can be implemented.

Furthermore, the intensity of a biosignal generally varies greatly between individuals and animals, which are objects to be counted, and usually the parameters in the signal counting process must be manually adjusted for each user and animal to be counted. must. On the other hand, according to the present embodiment, the parameters relating to the accuracy of the counting process such as the "upper reference value", the "lower reference value", the "upper threshold value" and the "lower threshold value" are automatically determined. This eliminates the need for manual adjustment for each object to be counted.

By the way, the above-mentioned initial time point position CP is the "first reference time point" related to the transition from <rising phase detection state (state 1)> to <upper threshold determination state (state 2)>, but here Summarizing the processing related to the initial point-in-time position CP described above, in the end,
(a1) In the <upper threshold determination state (state 2)>, the representative value SD _W reaches the “temporary lower reference value” during the time from the “first reference time (initial time point position CP)” to the elapse of a predetermined time. If it is not determined that the value exceeds
(a2) In the <falling phase detection state (state 3)>, if the "upper reference value" is not determined within a predetermined period of time from the "first reference point (initial point point position CP)", or,
(a3) In the <lower side threshold determination state (state 4)>, the representative value SD _W becomes the “temporary upper reference value” during the time from the “first reference time (initial time point position CP)” to the elapse of a predetermined time. If it is not determined that the value is less than
The number of waves of the biosignal is not counted or reset.

On the other hand, as already mentioned, in another embodiment, one cycle of biomedical signal processing starts from state 3, sequentially goes to states 4, 1 and 2, and ends with signal counting in state 2. It is also possible to perform processing.

In such an embodiment, the above-described initial time point position CP becomes the "second reference time point" related to the transition from <falling phase detection state (state 3)> to <lower threshold determination state (state 4)>, The processing related to this initial time point position CP is, after all,
(b1) In the <lower threshold value determination state (state 4)>, the representative value SD _W is the “temporary upper reference value” during the time from the “second reference time (initial time point position CP)” to the elapse of a predetermined time. If it is not determined that the value is less than
(b2) In <rising phase detection state (state 1)>, if the "lower reference value" is not determined within a predetermined period of time from the "second reference time (initial time point position CP)", or,
(b3) In the <upper threshold determination state (state 2)>, the representative value SD _W reaches the “temporary lower reference value” during the time from the “second reference time (initial time point position CP)” to the elapse of a predetermined time. If it is not determined that the value exceeds
In other words, the number of waves of the biosignal is not counted or reset.

In any case, by using the initial time point position CP, it is determined whether the temporal length of the input signal is within the range expected from the period of the signal to be counted, and the temporal length is appropriate. It becomes possible to subject only the signal to the counting process. As a result, for example, when a weak noise signal not to be counted, such as a "smile" signal, can be mixed with a relatively small amplitude "chewing" signal (although there are individual differences), counting can be performed without lowering the amplitude sensitivity. It is also possible to suppress erroneous counting of noise signals that are not to be counted by using the temporal length (periodicity) constraint expected of the target "chewing" signal.

In addition, the determination processing in step S705 described above can be summarized as follows.
(a) The point in time at which it is determined that the captured representative value SD _W (input signal data value) is a data value smaller than the predetermined condition when viewed from the "upper reference value" is from the most recent point in time when the wave number is counted. If the predetermined time or longer has not yet elapsed, the wave number count is skipped.

On the other hand, as described above, as another embodiment, one cycle of biological signal processing starts from state 3, sequentially transitions to states 4, 1, and 2, and finally, signal counting processing is performed in state 2. is to be performed, as the processing in the step corresponding to step S705 in state 2 (upper threshold determination state),
(b) The point in time at which it is determined that the captured representative value SD _W (input signal data value) is a data value larger than the predetermined condition when viewed from the "lower reference value" is from the point immediately after the wave number was counted. If the predetermined time or longer has not yet elapsed, the process of skipping the counting of wavenumbers is performed.

Further, as still another embodiment, in each of <rising phase detection state>, <upper threshold determination state>, <falling phase detection state>, and <lower threshold determination state>, the time spent in the state exceeds a predetermined time, it is possible not to count the wave number of the biosignal or to reset it. Of course, such processing may be performed together with the above-described processing using the initial time point position CP.

The biomedical signal analysis unit 101 and the signal counting unit 105 described above reduce the amount of calculation for noisy AC signals such as myoelectric signals (including electroencephalograms and the like), and reliably convert them into “periodic biosignals” as AC signals. can be detected. This utilizes the property that the signal detected by a myoelectric sensor or the like using dry electrodes is an alternating current. As a result, an AC signal with a small amplitude is not detected, and noise (artifact) due to displacement of the dry electrodes is not detected as a biological signal.

Finally, the determination result of the left/right body habit determining unit 104 (and the signal counting unit 105) is output to the application.
The application of the mobile terminal 2 can execute command instructions from the user triggered by, for example, a predetermined number of mastication actions by the user, such as camera shutter action, zooming, etc., and even favorite registration of the content being viewed. can.
For example, by increasing the number of times of chewing only on the right side, it is possible to fast-forward the electronic content or transition to the next page. On the other hand, by increasing the number of times of chewing only on the left side, the reverse operation can also be performed. Of course, it can also be applied to zoom-in/out, enlargement/reduction, volume adjustment, and the like.
In addition, it is also possible to transmit, for example, a meal log of a determination result regarding mastication of a meal to the analysis server 3 for each meal or every predetermined unit time (1 hour, 1 day, etc.).

<Analysis Server 3>
The analysis server 3 can analyze the following meal log information from the application of the mobile terminal 2 .
・Frequent pattern of uneven chewing (left-right body habit) ・Total number of chewing times ・Number of chewing times (left, both sides, right)
・Number of changes in chewing position ・Number of chewing times ・Average chewing pace ・Average chewing strength (root mean square of SDw, left, right)
・Meal time ・Meal time ・Meal photo

Moreover, the analysis server 3 may present a screen interface for inputting at least the self-evaluation result of mastication to the portable terminal 2 at the end of one meal of the user. This may be input intuitively by operating a slider such as VAS (Visual Analog Scale) and acquired as a numerical value.
Meal name: [Text input or selection]
Habits on the chewing side: Toward the left <-|-> Toward the right Chewed well: Didn't bite <-|-> Chewed well Scared of pulling with front teeth: Scared <-|-> Not scared Teeth hurts: Painful <-| -> No pain My jaw feels sluggish when I eat hard foods: I feel tired <-|-> I don't feel tired My jaw hurts: Painful <-|-> No pain Mental and physical condition: Absolutely unwell <-|-> Great condition Analysis server 3: User subjective information received from the mobile terminal 2 based on user operations is also recorded as a log.

The analysis server 3 presents the feedback information to the mobile terminal 2 based on the result of left/right body habit determination, the meal log, and the user's subjective information received from the mobile terminal 2 . The feedback information may be, for example, instructional content for the user or evaluation results of medical personnel.

As described in detail above, according to the biological signal processing apparatus, program, and method of the present invention, a plurality of myoelectric sensors are used to detect a body habit that causes a bias in the biological signals acquired bilaterally symmetrically. can be done.
In particular, when detecting uneven mastication, objectively, adults can masticate without difficulty after dental caries (cavities) treatment, dentures or dentures, and infants have no habit of uneven chewing. can be evaluated to

For the various embodiments of the present invention described above, various changes, modifications and omissions within the scope of the technical ideas and aspects of the present invention can be easily made by those skilled in the art. The foregoing description is exemplary only and is not intended to be limiting. The invention is to be limited only as limited by the claims and the equivalents thereof.

1 wearable device, glasses-type device, headband-type device 10 biological signal processing device 100a left signal conversion unit 100b right signal conversion unit 101a left biological signal analysis unit 101b right biological signal analysis unit 1011 pre-filter processing unit 1012 acceleration component generation unit 1013 Representative value calculation unit 102 Left/right strength determination unit 103 Left/right transition recording unit 104 Left/right body habit determination unit 105 Signal counting unit 11 Temple unit 12 Clings unit 13 Modern unit 2 Portable terminal 3 Analysis server

Claims (14)

A biosignal processing device for determining left and right eccentric mastication using biosignals acquired from left and right electrode groups arranged symmetrically in a living body,
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
A biological signal processing apparatus, comprising left-right body habit determination means for determining left- right eccentric mastication according to left-right deviation of the number of times of mastication based on the number of times of mastication . 2. The biological signal according to claim 1, wherein each of the left and right electrode groups comprises a positive electrode (detection electrode), a negative electrode (reference electrode) and a ground electrode (ground electrode), and is in contact with the skin of the living body. processing equipment. The electrode group is arranged bilaterally symmetrically on the head of the living body,
The left biosignal analysis means and the right biosignal analysis means calculate the left biosignal and the right biosignal based on the myoelectric signal related to mastication.
3. The biological signal processing apparatus according to claim 2, characterized in that: The positive electrode is in contact with the skin position near the cheek from the left and right auricles,
The negative electrode is in contact with the skin around the nose,
4. The biosignal processing apparatus according to claim 3, wherein the ground electrode is arranged so as to contact the skin position behind the head from the vicinity of the root of the back of the ear. The electrode group is fixed to the glasses type device,
The positive electrode is arranged to contact the skin of the living body from the temple part,
The negative electrode is arranged so as to be in contact with the skin of the living body from the clings part,
5. The biological signal processing apparatus according to claim 4, wherein the ground electrode is arranged so as to contact the skin of the living body from the modern portion. The right and left strength determination means is
If the mean square value of the left side is greater than the mean square value of the right side by more than a predetermined value, it is determined as the left side,
If the mean square value of the left side is smaller than the mean square value of the right side by more than a predetermined value, it is determined as the right side,
6. The biological signal processing apparatus according to any one of claims 1 to 5, wherein in other cases, it is determined as both sides. The left/right transition recording means records the left/right transition using a tree whose node is the transition to either the left, right, or both sides, and records the transition to either the left, right, or both sides for each mouth each time mastication occurs. Traversing the nodes and counting the number of times each node is traversed,
7. The biological signal processing apparatus according to claim 6 , wherein the left/right body habit determination means determines the left/right bias of the tree. The left-right transition recording means adds and records the number of occurrences and/or the dwell time to each node of the tree,
8. The biological signal processing apparatus according to any one of claims 1 to 7, wherein the left/right body habit determining means determines bias in the number of occurrences and/or dwell time of each node of the tree. A signal counting means for adding the left representative value output from the left biosignal analysis means and the right representative value output from the right biosignal analysis means, and counting the number of occurrences of myoelectricity. The biological signal processing device according to any one of claims 1 to 8. The signal counting means is
A data value related to an input signal that may include a periodic biological signal is sequentially acquired, and based on the acquired data value, a lower reference value corresponding to the minimum value of the data value and an upper reference value corresponding to the maximum value of the data value are obtained. (a) the lower reference value and the next upper reference value are determined or updated, and a data value smaller than the upper reference value satisfying a predetermined condition is taken in; or (b) when the upper reference value and the subsequent lower reference value are determined or updated, and a data value larger than the lower reference value satisfying a predetermined condition is acquired. 10. The biomedical signal processing device according to claim 9, which counts the number of waves of the biomedical signal. The left biosignal analysis means and the right biosignal analysis means are
acceleration component generation means for generating acceleration component data of the input biological signal;
11. The biological signal according to any one of claims 1 to 10, further comprising representative value calculating means for calculating a representative value relating to the state of occurrence of the biological signal in the predetermined time interval in the acceleration component data. processing equipment. A spectacles-type device that determines left and right eccentric mastication using biological signals acquired from left and right electrode groups arranged symmetrically on the head of a living body,
The left and right electrode groups are
A positive electrode (detection electrode) in the temple part and in contact with the skin position near the cheek from the left and right auricles,
Negative electrodes (reference electrodes) in the clings area and in contact with skin locations around the nose on the left and right sides;
It consists of a ground electrode (ground electrode) in the modern part and in contact with the skin position behind the head from around the left and right ear roots,
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
A spectacles-type device, comprising left-right body habit determination means for determining left- right eccentric mastication according to left-right deviation of the number of times of mastication based on the number of times of mastication . A program for causing a computer installed in a device for determining left and right eccentric mastication to function using biosignals acquired from left and right electrode groups arranged symmetrically in a living body,
left side biosignal analysis means for inputting a left side biosignal from an electrode group arranged on the left side of the living body and calculating a left representative value relating to the occurrence of the left side biosignal at predetermined time intervals;
right biosignal analysis means for inputting a right biosignal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the occurrence of the right biosignal at predetermined time intervals;
left and right strength determination means for calculating the mean square value of each of the left representative value and the right representative value for a predetermined unit time and determining the left side or the right side having the larger mean square value;
Left-right transition recording means for counting the number of left and right movements each time one mouthful action occurs from ingestion to swallowing ;
A program for causing a computer to function as left/ right body habit determining means for determining left/right eccentric mastication according to left/right deviation of the number of times of mastication based on the number of times of mastication . A left and right mastication determination method for a device for determining right and left uneven mastication using biosignals acquired from left and right electrode groups arranged symmetrically in a living body, comprising:
The device
A left side biomedical signal is inputted from a group of electrodes arranged on the left side of the living body, and a left representative value relating to the state of generation of the left side biomedical signal is calculated at predetermined time intervals,
a first step of inputting a right biomedical signal from a group of electrodes arranged on the right side of the living body and calculating a right representative value relating to the generation of the right biomedical signal at predetermined time intervals;
A second step of calculating the mean square value for a predetermined unit time for each of the left representative value and the right representative value, and determining the left side or the right side having the larger mean square value;
a third step of counting the number of times each of the left and right sides each time a mouthful action from ingestion to swallowing occurs;
and a fourth step of judging left and right unbalanced mastication according to the left/right deviation of the number of times of mastication based on a plurality of times of mastication.

Images