Abstract
In this paper, a measurement system of swallowing motion is developed. A piezoelectric film is used as sensitive material of the sensor to measure the body surface deformation due to swallowing. We focused on laryngeal prominence and digastric muscle which are active during swallowing and developed sensors to attach to skin above laryngeal prominence and digastric muscle. As a result of experiments, it was confirmed that output fluctuation occurred at the swallowing. Next, by comparing the frequency spectral density distribution of sensor outputs during swallowing and resting, it was found that the output waveform during swallowing is mainly composed of larger frequency components below 20 [Hz] than that during resting. From the results, it was suggested that the developed measurement system is available to detect the swallowing from skin deformation without obstructing the motion.
Introduction
Swallowing is a motion to send food and water taken into the oral cavity to the stomach. Since the function related to swallowing is weakened due to aging or illness, aspiration, which food and water enter the trachea, is likely to occur. Aspiration may cause pneumonia. In Japan, pneumonia is the third most frequent cause of death in elderly people aged 65 years and older [1], so aspiration is a serious problem. As a method of examining the swallowing function, palpation and auscultation at the time of swallowing are performed. However, palpation and auscultation are subjective and require skill [2]. In addition, although radiographic examination and endoscopic examination are also performed, there are problems such as obstructing swallowing movement by the measurement system [3]. Therefore, development of a simpler and more objective swallowing function evaluation method is required.
There are some studies to evaluate swallowing by measuring electromyograms of the digastric muscle [4] and swallowing sound [5]. However, in order to detect swallowing accurately, discrimination between swallowing and movements related to the larynx is not sufficiently performed.
In this paper, as a basic research to develop a sensor that evaluates swallowing function, we developed a piezoelectric sensor and investigated the sensor output characteristics during swallowing. We focused on the laryngeal prominence and digastric muscles which are active during swallowing. The laryngeal prominence can be observed from the outside during swallowing. And the digastric muscle was a measurement place where you would not feel discomfort during swallowing. We constructed a system that measures skin deformation during swallowing with a piezoelectric film sensor and verified the effectiveness of swallowing movement measurement through experiments.
Measurement system for the Laryngeal prominence movement
Experimental setups
Figure 1 shows the measurement system used in this experiment. In the experiment, a piezoelectric film sensor using a polyvinylidene fluoride (PVDF) film (Tokyo Sensor, DT1-028K) attaches to the laryngeal prominence. Figure 2 shows the developed PVDF sensor. The PVDF film is protected with polyurethane tape. The sensor is 45 mm long, 18 mm wide and 0.50 mm thick. For the comparison data, the laryngeal movement was measured using a 3D motion analysis system (Inter Reha, MX-T160 (VICON)). The 3D motion analysis system measures the position information of the infrared reflective marker using four infrared cameras. Figure 3 shows the position of the markers. The markers for the motion analysis system were attached to the mylohyoid muscle (top) and laryngeal prominence (middle) involved in laryngeal elevation during swallowing. In addition, markers were attached to the chin (chin) and the left and right sides of the throat (left, right) to compensate the influence of body movement.
Measurement system.
PVDF sensor.
Sensor and marker positions.
As the coordinates, the head side surface direction was set as the X axis, the head front direction as the Y axis, and the vertical direction as the Z axis. In this measurement system, the displacement in the X-axis direction is ignored, and the amount of movement of the marker in the YZ plane is analyzed. The distances between the top marker and the three markers for body movement compensation were calculated sequentially. And the change in each distance from the start of measurement was calculated. Then, the total of the three distances was derived as a parameter for the movement amount of the top marker. Similarly, the movement amount of the middle marker was derived. The sensor outputs were recorded on the PC through a data logger (Made by National Instruments, USB-6210) with the motion analysis system. Sampling rate is 1000 [Hz] for PVDF sensor and 100 [Hz] for VICON.
Experimental subjects were four healthy male who had no problem with swallowing function. The measurement time was set to 10 seconds. After attaching sensors and markers to the subjects, they were seated in a chair and raised their chin in order to make it easier to record each marker on the submandibular and throat with VICON. Then, the subject swallowed water or saliva once at any timing in the measurement time. For comparison, we measured at rest. The amount of water when swallowing water was 3 conditions of 10 ml, 20 ml and 30 ml. The number of trials for each condition was 5 times. Measurement at rest was performed once. Therefore, total of 21 trials were performed.
Results
Figures 4 and 5 show sensor outputs and marker movements during swallowing saliva and at rest of subject A, respectively. In these figures, (a) is the waveform of the PVDF sensor, (b) is marker “top” movement, and (C) is marker “middle” movement. From the results, it was confirmed that the movement of the marker “top” and “middle” become larger during swallowing. As for PVDF sensor output, sensor output fluctuation was observed at the timing when the marker movement became large. From the results, it was confirmed that the PVDF sensor can obtain output fluctuations during swallowing.
Sensor waveform and marker movement (Swallowing saliva).
Sensor waveform and marker movement (Rest).
Extracted sensor waveform.
In order to analyze the characteristics of the sensor outputs in more detail, the waveform during swallowing were cut out. The amount of movement of the top marker is used as a reference for cutting out the waveforms, because a large peak appears during swallowing. The time interval to cut out was set to 1.5 seconds, because it is confirmed that time length of muscle activity per one swallowing was within about 1.5 seconds [6]. Specifically, a total of 1.5 seconds was cut out in 0.75 seconds before and after the time when the peak appears in the movement amount of the top marker. Figure 6 shows the extracted PVDF sensor outputs from the outputs in Figs 4 and 5.
And the power spectral density (PSD) distribution was calculated from the PVDF sensor output as shown in Fig. 6. Figure 7 shows the calculated PSD distribution. The frequency on the horizontal axis of the graph is a logarithmic plot. From this figure, it was confirmed that the frequency component below 20 [Hz] was larger during swallowing than that at rest.
PSD.
Figure 8 shows summation of PSD value from 0 to 20 [Hz] for each measurement. The values in the figure are average values, and error bars are standard deviations. From the results, the value during swallowing is clearly larger than that at rest. However, the relationship between the amount of water in swallowing and the summation of PSD value from 0 to 20 [Hz] was different for each subject.
Summation of PSD value from 0 to 20 [Hz].
Subsequently, since an increase in frequency components below 20 [Hz] was confirmed as a feature of the output by swallowing, a low-pass filter of 20 [Hz] was applied to the PVDF sensor output. Figure 9 shows the PVDF sensor output with the low-pass filter. The sensor output fluctuation during swallowing was confirmed in all subjects. Figure 10 shows the maximum amplitude of the waveform after the low-pass filter. The values in the figure are average values for 5 trials for each condition, and error bars are standard deviations. From these results, the sensor output fluctuation during swallowing showed a larger value than that at rest, and the difference could be made clearer. However, there was no common trend among subjects in relation to the amount of water and the maximum amplitude.
PVDF output after low-pass filter.
Comparison of maximum amplitude in filtered PVDF output.
From the above results, it was confirmed that the sensor output fluctuation during swallowing was confirmed in the piezoelectric sensor output. In particular, the output fluctuation due to swallowing became clearer in the PVDF sensor output to which a 20 [Hz] low-pass filter was applied. And it was found that the amplitude of the sensor output is available as an index for swallowing. However, when the PVDF sensor is attached to the laryngeal prominence, it causes discomfort during swallowing. In the next chapter, we conducted a swallowing measurement experiment focusing on the digastric muscles.
When measuring swallowing in daily life, it is important not to make the subject feel uncomfortable. It is also necessary to take into account the ease of attaching. We focused on the digastric muscles, and measured skin deformation due to the activity of the digastric muscle by using the developed PVDF sensor.
Experimental setup
First, we examined whether swallowing could be detected from skin deformation caused by the activity of the digastric muscle. As shown in Fig. 11, the PVDF sensor was attached to the skin above the digastric muscle. The PVDF sensor was attached to the laryngeal prominence too. The PVDF sensor outputs were recorded on the PC through a data logger. The sampling rate was set to 1000 [Hz]. Experimental subjects were one healthy male (subject E) who had no problem with swallowing function. The measurement time was set to 5 seconds. The subject performed the experiment in the sitting position. The subjects swallowed saliva once at any timing of the measurement time while facing forward, and the sensor output at that time was measured. The number of trials was 3 trials.
Sensor position (Digastric muscle).
Sensor waveform.
Moreover, in order to investigate the characteristics of sensor output. we measured the sensor output during swallowing and at rest and analyze the characteristics of sensor output during swallowing. Experimental subjects were four healthy male (subject E, F, G, H) who had no problem with swallowing function. The experimental system and conditions were the same as in the above experiment. However, the PVDF sensor was attached only to the digastric muscle and the number of trials is 5 trials for both swallowing and resting.
Figure 12 shows the PVDF sensor output waveform during swallowing. In the figure, (a) and (b) are the output of the PVDF sensor attached to the laryngeal prominence and that of the digastric muscle, respectively. It was confirmed that sensor output fluctuations in both figures were observed at almost the same timing. As a result, it was found that the skin deformation during swallowing can be measured by the PVDF sensor attached to the skin above the digastric muscle.
Figure 13 shows the sensor output waveform of subject E. Figure 14 shows the sensor output waveform of subject H. In these figures, (a) are the output waveforms during swallowing and (b) are the output waveforms at rest.
Sensor waveform (Subject E).
Sensor waveform (Subject H).
Subsequently, PSD distribution was calculated for these sensor output waveforms. Figure 15 shows the PSD distribution of subject E. Figure 16 shows the PSD distribution of subject H. In these figures, (a) are the PSD distributions during swallowing and (b) are the PSD distributions at rest. It was confirmed that the frequency components below 20 [Hz] are larger at swallowing than at rest.
PSD (Subject E).
Figure 17 shows summation of PSD value from 0 to 20 [Hz] for each subject. The values in the figure are average values for each condition, and error bars are standard deviations. From this result, it was confirmed that the value of swallowing was larger than the value at rest. Subsequently, we performed t tests between during swallowing and at rest. Subject H showed no significant difference, but other subjects showed significant difference. It is considered that the shape of the face and the amount of fat affect the magnitude of the sensor output fluctuation. In the future, it will be necessary to increase the number of subjects and investigate how the sensor output during swallowing will be affected by subject characteristics.
PSD (Subject H).
Summation of PSD value from 0 to 20 [Hz].
In this study, as a basic research for developing a sensor for evaluating swallowing function, we fabricated a piezoelectric sensor using PVDF film and developed a measurement system for swallowing movement. In the experiment, the PVDF sensors were attached to the laryngeal prominence and the digastric muscle, and the sensor output during swallowing was measured. The sensor output fluctuation was confirmed during swallowing, and it was mainly composed of frequency components below 20 [Hz]. From the results, it was suggested that the developed sensor system is available to measure the swallowing without obstructing the motion.