# Short-Time Impedance Spectroscopy Using a Mode-Switching Nonsinusoidal Oscillator: Applicability to Biological Tissues and Continuous Measurement

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

- (1)
- Design a nonsinusoidal oscillation circuit capable of periodically switching oscillation frequency;
- (2)
- Show that the proposed method can estimate circuit parameters for circuit models close to actual measured objects;
- (3)
- Apply the proposed method to the impedance measurement of biological tissues;
- (4)
- Show that the continuous measurement of impedance is possible.

## 2. Capacitive Coupling Impedance Spectroscopy

#### 2.1. Mode-Switching Nonsinusoidal Oscillator

#### 2.2. Experimental Method

#### 2.2.1. CIS of Parallel RC Circuits and Bioimpedance Models

#### 2.2.2. Application of CIS to Biological Tissues

#### 2.2.3. Continuous Impedance Measurement Using CIS

#### 2.2.4. Data Acquisition and DFT Analysis

## 3. Results and Discussion

#### 3.1. CIS of Parallel RC Circuits and Bioimpedance Models

#### 3.2. Application of CIS to Biological Tissues

#### 3.3. Continuous Impedance Measurement Using CIS

## 4. Conclusions

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## References

- Grossi, M.; Riccò, B. Electrical impedance spectroscopy (EIS) for biological analysis and food characterization: A review. J. Sens. Sens. Syst.
**2017**, 6, 303–325. [Google Scholar] [CrossRef] [Green Version] - Szuster, B.; Roj, Z.S.D.; Kowalski, P.; Sobotnicki, A.; Woloszyn, J. Idea and measurement methods used in bioimpedance spectroscopy. Adv. Intell. Syst. Comput.
**2017**, 623, 70–78. [Google Scholar] - Davies, S.J.; Davenport, A. The role of bioimpedance and biomarkers in helping to aid clinical decision-making of volume assessments in dialysis patients. Kidney Int.
**2014**, 86, 489–496. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Piccoli, A. Bioelectric impedance measurement for fluid status assessment. Contrib. Nephrol.
**2010**, 164, 143–152. [Google Scholar] [PubMed] [Green Version] - Jaffrin, M.Y.; Morel, H. Body fluid volumes measurements by impedance: A review of bioimpedance spectroscopy (BIS) and bioimpedance analysis (BIA) methods. Med. Eng. Phys.
**2008**, 30, 1257–1269. [Google Scholar] [CrossRef] [PubMed] - Kyle, U.G.; Bosaeus, I.; De Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gómez, J.M.; Heitmann, B.L.; Kent-Smith, L.; Melchior, J.-C.; Pirlich, M.; et al. Bioelectrical impedance analysis—Part I: Review of principles and methods. Clin. Nutr.
**2004**, 23, 1226–1243. [Google Scholar] [CrossRef] - Kyle, U.G.; Bosaeus, I.; De Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gómez, J.M.; Heitmann, B.L.; Kent-Smith, L.; Melchior, J.-C.; Pirlich, M.; et al. Bioelectrical impedance analysis—Part II: Utilization in clinical practice. Clin. Nutr.
**2004**, 23, 1430–1453. [Google Scholar] [CrossRef] - Demura, S.; Sato, S.; Kitabayashi, T. Percentage of total body fat as estimated by three automatic bioelectrical impedance analyzers. J. Physiol. Anthropol. Appl. Hum. Sci.
**2004**, 23, 93–99. [Google Scholar] [CrossRef] [Green Version] - Cha, K.; Chertow, G.M.; Gonzalez, J.; Lazarus, J.M.; Wilmore, D.W. Multifrequency bioelectrical impedance estimates the distribution of body water. J. Appl. Physiol.
**1995**, 79, 1316–1319. [Google Scholar] [CrossRef] - Groenendaal, W.; Lee, S.; van Hoof, C. Wearable bioimpedance monitoring: Viewpoint for application in chronic conditions. JMIR Biomed. Eng.
**2021**, 6, e22911. [Google Scholar] [CrossRef] - Sel, K.; Osman, D.; Jafari, R. Non-invasive cardiac and respiratory activity assessment from various human body locations using bioimpedance. IEEE Open J. Eng. Med. Biol.
**2021**, 2, 210–217. [Google Scholar] [CrossRef] [PubMed] - Zink, M.D.; König, F.; Weyer, S.; Willmes, K.; Leonhardt, S.; Marx, N.; Napp, A. Segmental bioelectrical impedance spectroscopy to monitor fluid status in heart failure. Sci. Rep.
**2020**, 10, 3577. [Google Scholar] [CrossRef] [PubMed] - Collins, C.T.; Reid, J.; Makrides, M.; Lingwood, B.E.; McPhee, A.J.; Morris, S.A.; Gibson, R.A.; Ward, L.C. Prediction of body water compartments in preterm infants by bioelectrical impedance spectroscopy. Eur. J. Clin. Nutr.
**2013**, 67, S47–S53. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Piccoli, A.; Rossi, B.; Pillon, L.; Bucciante, G. A new method for monitoring body fluid variation by bioimpedance analysis: The RXc graph. Kidney Int.
**1994**, 46, 534–539. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Nwosu, A.C.; Mayland, C.R.; Mason, S.; Cox, T.F.; Varro, A.; Ellershaw, J. The association of hydration status with physical signs, symptoms and survival in advanced cancer—The use of bioelectrical impedance vector analysis (BIVA) technology to evaluate fluid volume in palliative care: An observational study. PLoS ONE
**2016**, 11, e0163114. [Google Scholar] [CrossRef] [Green Version] - Toso, S.; Piccoli, A.; Gusella, M.; Menon, D.; Crepaldi, G.; Bononi, A.; Ferrazzi, E. Bioimpedance vector pattern in cancer patients without disease versus locally advanced or disseminated disease. Nutrition
**2003**, 19, 510–514. [Google Scholar] [CrossRef] - Nescolarde, L.; Piccoli, A.; Román, A.; Núñez, A.; Morales, R.; Tamayo, J.; Doñate, T.; Rosell, J. Bioelectrical impedance vector analysis in haemodialysis patients: Relation between oedema and mortality. Physiol. Meas.
**2004**, 25, 1271–1280. [Google Scholar] [CrossRef] - Piccoli, A.; Italian CAPD-BIA Study Group. Bioelectric impedance vector distribution in peritoneal dialysis patients with different hydration status. Kidney Int.
**2004**, 65, 1050–1063. [Google Scholar] [CrossRef] [Green Version] - Bozzetto, S.; Piccoli, A.; Montini, G. Bioelectrical impedance vector analysis to evaluate relative hydration status. Pediatr. Nephrol.
**2010**, 25, 329–334. [Google Scholar] [CrossRef] - Da Fonseca, R.D.; Santos, P.R.; Monteiro, M.S.; Fernandes, L.A.; Campos, A.H.; Borges, D.L.; Rosa, S.S.R.F. Parametric evaluation of impedance curve in radiofrequency ablation: A quantitative description of the asymmetry and dynamic variation of impedance in bovine ex vivo model. PLoS ONE
**2021**, 16, e0245145. [Google Scholar] [CrossRef] - Barkagan, M.; Rottmann, M.; Leshem, E.; Shen, C.; Buxton, A.E.; Anter, E. Effect of baseline impedance on ablation lesion dimensions: A multimodality concept validation from physics to clinical experience. Circ. Arrhythm. Electrophysiol.
**2018**, 11, e006690. [Google Scholar] [CrossRef] - Sulkin, M.S.; Laughner, J.I.; Hilbert, S.; Kapa, S.; Kosiuk, J.; Younan, P.; Romero, I.; Shuros, A.; Hamann, J.J.; Hindricks, G.; et al. Novel measure of local impedance predicts catheter—Tissue contact and lesion formation. Circ. Arrhythm. Electrophysiol.
**2018**, 11, e005831. [Google Scholar] [CrossRef] - Chinitz, J.S.; Michaud, G.F.; Stephenson, K. Impedance-guided radiofrequency ablation: Using impedance to improve ablation outcomes. Innov. Card. Rhythm. Manag.
**2017**, 8, 2868–2873. [Google Scholar] [CrossRef] [Green Version] - Yamaguchi, T.; Ueno, A. Capacitive-coupling impedance spectroscopy using a non-sinusoidal oscillator and discrete-time Fourier transform: An introductory study. Sensors
**2020**, 20, 6392. [Google Scholar] [CrossRef] - Takano, M.; Ueno, A. Noncontact in-bed measurements of physiological and behavioral signals using an integrated fabric-sheet sensing scheme. IEEE J. Biomed. Health Inform.
**2019**, 23, 618–630. [Google Scholar] [CrossRef] - Elghajiji, A.; Wang, X.; Weston, S.D.; Zeck, G.; Hengerer, B.; Tosh, D.; Rocha, P.R.F. Electrochemical impedance spectroscopy as a tool for monitoring cell differentiation from floor plate progenitors to midbrain neurons in real time. Adv. Biol.
**2021**, 5, 2100330. [Google Scholar] [CrossRef] [PubMed] - Sato, H.; Nakamura, T.; Kusuhara, T.; Kenichi, K.; Kuniyasu, K.; Kawashima, T.; Hanayama, K. Effectiveness of impedance parameters for muscle quality evaluation in healthy men. J. Physiol. Sci.
**2020**, 70, 53. [Google Scholar] [CrossRef] [PubMed] - Lu, F.; Wang, C.; Zhao, R.; Du, L.; Fang, Z.; Guo, X.; Zhao, Z. Review of stratum corneum impedance measurement in non-invasive penetration application. Biosensors
**2018**, 8, 31. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Hernández-Balaguera, E.; López-Dolado, E.; Polo, J.L. Obtaining electrical equivalent circuits of biological tissues using the current interruption method, circuit theory and fractional calculus. RSC Adv.
**2016**, 6, 22312. [Google Scholar] [CrossRef] - Münkler, P.; Gunawardene, M.A.; Jungen, C.; Klatt, N.; Schwarzl, J.M.; Akbulak, R.Ö.; Dinshaw, L.; Hartmann, J.; Jularic, M.; Kahle, A.K.; et al. Local impedance guides catheter ablation in patients with ventricular tachycardia. J. Cardiovasc. Electrophysiol.
**2020**, 31, 61–69. [Google Scholar] [CrossRef]

**Figure 1.**Mode-switching nonsinusoidal oscillator. The oscillation frequency ${f}_{0}$ and the oscillation period ${T}_{0}$ of the oscillator are switched between four different frequencies by 4-bit binary counters and bilateral analog switches. CLK of 4-bit binary counters denotes a clock signal, and RST represents a reset signal. ${Q}_{0}$, ${Q}_{1}$, ${Q}_{2,}$ and ${Q}_{3}$ of 4-bit binary counters denote zero bits, and the first, second, and third bits of the output signals, respectively. ${C}_{\mathrm{O}}$ represents a bypass capacitor to remove the switching noise in the oscillator, and its value is 0.47 nF. CTR1 and CTR2 are binary counters to control the repeat count and the mode number of the mode switching, respectively.

**Figure 2.**Impedance measurement of porcine myocardium: (

**a**) photograph of the porcine left ventricle coagulated using electrocautery; (

**b**) measurement electrode fixture.

**Figure 3.**Constant current circuit for the LED current control of the photocoupler. The CdS photocell was used as (

**a**) a single resistor, (

**b**) a resistor ${R}_{\mathrm{X}}$ of the bioimpedance model, or (

**c**) ${R}_{\mathrm{PX}}$ of its model.

**Figure 4.**Four-mode oscillation waveforms of ${v}_{1}\left(t\right)$ and ${v}_{12}\left(t\right)$ obtained through the mode-switching nonsinusoidal oscillator. ${R}_{\mathrm{X}}$, ${R}_{\mathrm{PX}}$, ${C}_{\mathrm{PX}}$, and ${C}_{\mathrm{SX}}$ were set to 0.10 k, 0 Ω, 1.016 nF, and 9.76 nF, respectively. The oscillation frequency was switched every two cycles of each ${R}_{\mathrm{F}}$. The DFT analysis was performed on the oscillation waveforms of the second half of two cycles.

**Figure 5.**Frequency spectra of the (

**a**) real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ and (

**b**) imaginary part $\mathrm{Im}\left({\dot{Z}}_{\mathrm{X}}\right)$ of the impedance ${\dot{Z}}_{\mathrm{X}}$ comprising the resistance ${R}_{\mathrm{X}}$ of 0.10 kΩ, resistance ${R}_{\mathrm{PX}}$ of 0 Ω, capacitance ${C}_{\mathrm{PX}}$ of 1.016 nF (parallel RC circuit), and capacitance ${C}_{\mathrm{PX}}$ of 9.76 nF. The dashed lines represent the theoretical curves calculated from Equations (9) and (10).

**Figure 6.**Frequency spectra of the (

**a**) real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ and (

**b**) imaginary part $\mathrm{Im}\left({\dot{Z}}_{\mathrm{X}}\right)$ of the impedance ${\dot{Z}}_{\mathrm{X}}$ comprising the resistance ${R}_{\mathrm{X}}$ of 0.91 kΩ, resistance ${R}_{\mathrm{PX}}$ of 0 Ω, capacitance ${C}_{\mathrm{PX}}$ of 9.85 nF (parallel RC circuit), and capacitance ${C}_{\mathrm{SX}}$ of 9.76 nF. The dashed lines represent the theoretical curves calculated from Equations (9) and (10).

**Figure 7.**Absolute relative estimation errors of the circuit parameters of parallel RC circuits: (

**a**) ${R}_{\mathrm{X}}$ = 0.10 kΩ, ${C}_{\mathrm{PX}}$ = 1.016 nF, $\mathrm{and}{C}_{\mathrm{SX}}$ = 9.76 nF; (

**b**) ${R}_{\mathrm{X}}$ = 0.91 kΩ, ${C}_{\mathrm{PX}}$ = 9.85 nF, and ${C}_{\mathrm{SX}}$ = 9.76 nF.

**Figure 8.**Frequency spectra of the (

**a**) real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ and (

**b**) imaginary part $\mathrm{Im}\left({\dot{Z}}_{\mathrm{X}}\right)$ of the impedance ${\dot{Z}}_{\mathrm{X}}$ comprising the parallel RC circuits for ${C}_{\mathrm{PX}}=$ 1.016 nF and coupling capacitance ${C}_{\mathrm{SX}}=$ 9.76 nF. The DFT data were fitted to Equations (9) and (10) to calculate ${R}_{\mathrm{X}}$, ${C}_{\mathrm{PX}}$, and ${C}_{\mathrm{SX}}$ (dashed lines).

**Figure 9.**${R}_{\mathrm{X}}$, ${C}_{\mathrm{PX}},$ and ${C}_{\mathrm{SX}}$ of the parallel RC circuits estimated using oscillation waveforms and DFT: (

**a**) the estimated ${R}_{\mathrm{X}}$, (

**b**) relative error of the estimated ${C}_{\mathrm{PX}}$, and (

**c**) relative error of the estimated ${C}_{\mathrm{SX}}$. The insets are the circuit model used for fitting, and the estimated parameters are highlighted in pink.

**Figure 10.**Cole–Cole plot of ${\dot{Z}}_{\mathrm{X}1}$ for the parallel RC circuits with ${C}_{\mathrm{PX}}=$1.016 nF and ${C}_{\mathrm{SX}}=$9.76 nF. The dashed lines represent the fitting curves obtained from Equations (9) and (10).

**Figure 11.**Frequency spectra of the (

**a**) real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ and (

**b**) imaginary part $\mathrm{Im}\left({\dot{Z}}_{\mathrm{X}}\right)$ of the impedance ${\dot{Z}}_{\mathrm{X}}$ comprising the bioimpedance models for ${R}_{\mathrm{X}}=$0.30 kΩ, ${C}_{\mathrm{PX}}=$3.22 nF, and coupling capacitance ${C}_{\mathrm{SX}}=$9.76 nF. DFT data were fitted to Equations (9) and (10) to determine ${R}_{\mathrm{X}}$, ${R}_{\mathrm{PX}}$, ${C}_{\mathrm{PX}},$ and ${C}_{\mathrm{SX}}$ (dashed lines).

**Figure 12.**${R}_{\mathrm{X}}$, ${C}_{\mathrm{PX}}$, and ${C}_{\mathrm{SX}}$ of the bioimpedance models estimated using oscillation waveforms and DFT: (

**a**) the estimated ${R}_{\mathrm{X}}$, (

**b**) the estimated ${R}_{\mathrm{PX}}$, (

**c**) relative error of the estimated ${C}_{\mathrm{PX}}$, and (

**d**) relative error of the estimated ${C}_{\mathrm{SX}}$. The insets are the circuit models used for fitting, and the estimated parameters are highlighted in pink.

**Figure 13.**Cole–Cole plot of ${\dot{Z}}_{\mathrm{X}1}$ for the bioimpedance models with ${C}_{\mathrm{PX}}=$3.22 nF and ${C}_{\mathrm{SX}}=$9.76 nF. The dashed lines represent the fitting curves obtained from Equations (9) and (10).

**Figure 14.**Frequency spectrum of the real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ of the impedance ${\dot{Z}}_{\mathrm{X}}$ of the porcine myocardium and coupling capacitance. DFT data were fitted to Equation (9) to determine ${R}_{\mathrm{X}}$, ${R}_{\mathrm{PX}}$, and ${C}_{\mathrm{PX}}$ (dashed lines).

**Figure 15.**Frequency spectra of the impedance ${\dot{Z}}_{\mathrm{X}1}$ of the porcine myocardium: (

**a**) the Cole–Cole plot, (

**b**) the real part $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}1}\right)$, and (

**c**) the imaginary part $\mathrm{Im}\left({\dot{Z}}_{\mathrm{X}1}\right)$. ${\dot{Z}}_{\mathrm{X}1}$ represents the impedance obtained by subtracting the impedance of the estimated ${C}_{\mathrm{SX}}$ from ${\dot{Z}}_{\mathrm{X}}$. $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}1}\right)$ equals $\mathrm{Re}\left({\dot{Z}}_{\mathrm{X}}\right)$ in Figure 14.

**Figure 16.**Comparison of equivalent circuit parameters between the untreated and ablated left ventricles of porcine hearts: (

**a**) ${R}_{\mathrm{X}}$, (

**b**) ${R}_{\mathrm{PX}}$, and (

**c**) ${C}_{\mathrm{PX}}$. The sample size is 12.

**Figure 17.**Single-mode oscillation waveforms of ${v}_{12}\left(t\right)$ and ${v}_{1}\left(t\right)$ for the nonsinusoidal oscillator incorporating the CdS photocell (Figure 3a): (

**a**) full waveform and (

**b**) enlarged waveform. The photocurrent ${I}_{\mathrm{Ph}}$ of the photocoupler is changed at 10 and 23 ms. The high and low levels of ${I}_{\mathrm{Ph}}$ are 2.0 and 0.6 mA, respectively. Circuit parameters are estimated for each cycle (blue dashed line). Representative points of time in each cycle are the ends of the cycles such as ${t}_{1}$ and ${t}_{2}$.

**Figure 18.**Resistance ${R}_{\mathrm{X}}$ of the photocell estimated using oscillation waveforms in Figure 17a and DFT: (

**a**) time–resistance curves and (

**b**) saturation values at a high ${I}_{\mathrm{Ph}}$ period. Time of the time–resistance curves corresponds to the representative points of time in Figure 17b.

**Figure 19.**Two-mode oscillation waveforms of ${v}_{12}\left(t\right)$ and ${v}_{1}\left(t\right)$ for the nonsinusoidal oscillator incorporating the CdS photocell and circuit elements of the bioimpedance model (Figure 3b): (

**a**) full waveform and (

**b**) enlarged waveform. The photocurrent ${I}_{\mathrm{Ph}}$ of the photocoupler is changed at 10 and 23 ms. The high and low levels of ${I}_{\mathrm{Ph}}$ are 8.0 and 4.0 mA, respectively. Circuit parameters are estimated for each cycle (blue dashed line). Representative points of time in each cycle are the ends of the cycles such as ${t}_{1}$ and ${t}_{2}$. The frequency f is the oscillation frequency obtained from the second half cycle of the oscillation waveform at ${R}_{\mathrm{F}}$ = 3.3 kΩ.

**Figure 20.**Time–resistance, time–capacitance, and time–frequency curves of the resistance of the photocell and circuit elements of the bioimpedance model estimated using oscillation waveforms and DFT. The photocell was used as (

**a**) ${R}_{\mathrm{X}}$ or (

**b**) ${R}_{\mathrm{PX}}$. The other resistance is 0.20 kΩ; the capacitance ${C}_{\mathrm{PX}}$ is 3.22 nF; and the capacitance ${C}_{\mathrm{SX}}$ is 9.76 nF. The times of these curves correspond to the representative points of time in Figure 19b. The insets are photocell-incorporated circuits.

Mode | ${\mathit{Q}}_{1}$ | ${\mathit{Q}}_{0}$ | SW1 | SW2 | ${\mathit{R}}_{\mathit{F}}$ |
---|---|---|---|---|---|

1 | 0 | 0 | OFF | OFF | ${R}_{\mathrm{F}1}+{R}_{\mathrm{F}2}+{R}_{\mathrm{F}3}$ |

2 | 0 | 1 | OFF | ON | ${R}_{\mathrm{F}1}+{R}_{\mathrm{F}3}$ |

3 | 1 | 0 | ON | OFF | ${R}_{\mathrm{F}1}+{R}_{\mathrm{F}2}$ |

4 | 1 | 1 | ON | ON | ${R}_{\mathrm{F}1}$ |

**Table 2.**Two sets of the oscillation frequencies obtained by combining ${R}_{\mathrm{F}1}$, ${R}_{\mathrm{F}2}$, and ${R}_{\mathrm{F}3}$.

Set | ${\mathit{R}}_{\mathit{F}1}$ (kΩ) | ${\mathit{R}}_{\mathit{F}2}$ (kΩ) | ${\mathit{R}}_{\mathit{F}3}$ (kΩ) | 4 Modes of
${\mathit{R}}_{\mathit{F}}$ (kΩ) | Estimated Oscillation Frequencies (kHz) |
---|---|---|---|---|---|

1 | 3.3 | 1.8 | 3 | 8.1, 6.3, 5.1, 3.3 | 16.9, 22.3, 28.4, 47.9 |

2 | 5.1 | 1 | 2 | 8.1, 7.1, 6.1, 5.1 | 16.9, 19.5, 23.1, 28.4 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Yamaguchi, T.; Ogawa, E.; Ueno, A.
Short-Time Impedance Spectroscopy Using a Mode-Switching Nonsinusoidal Oscillator: Applicability to Biological Tissues and Continuous Measurement. *Sensors* **2021**, *21*, 6951.
https://doi.org/10.3390/s21216951

**AMA Style**

Yamaguchi T, Ogawa E, Ueno A.
Short-Time Impedance Spectroscopy Using a Mode-Switching Nonsinusoidal Oscillator: Applicability to Biological Tissues and Continuous Measurement. *Sensors*. 2021; 21(21):6951.
https://doi.org/10.3390/s21216951

**Chicago/Turabian Style**

Yamaguchi, Tomiharu, Emiyu Ogawa, and Akinori Ueno.
2021. "Short-Time Impedance Spectroscopy Using a Mode-Switching Nonsinusoidal Oscillator: Applicability to Biological Tissues and Continuous Measurement" *Sensors* 21, no. 21: 6951.
https://doi.org/10.3390/s21216951