Hysteresis-free and high sensitivity strain sensing of ionically conductive hydrogels

Abstract

Hydrogels are promising materials for soft and implantable strain sensors owing to their large compliance (E<100 kPa) and significant extensibility (ε_max >500%) compared to other polymer networks. Further, hydrogels can be functionalized to seamlessly integrate with many types of tissues. However, most current methods attempt to imbue additional electronic functionality to structural hydrogel materials by incorporating fillers with orthogonal properties such as electronic or mixed ionic conduction. Although composite strategies may improve performance or facilitate heterogeneous integration with downstream hardware, composites complicate the path for regulatory approval and may compromise the otherwise compelling properties of the underlying structural material. Here we report hydrogel strain sensors composed of genipin-crosslinked gelatin and dopamine-functionalized poly(ethylene glycol) for in vivo monitoring of cardiac function. By measuring their impedance only in their resistive regime (>10 kHz), hysteresis is reduced and the resulting gauge factor is increased by ~50x to 1.02±0.05 and 1.46±0.05 from approximately 0.03–0.05 for PEG-Dopa and genipin-crosslinked gelatin respectively. Adhesion and in vivo biocompatibility are studied to support implementation of strain sensors for monitoring cardiac output in porcine models. Impedance-based strain sensing in the kilohertz regime simplifies the piezoresistive behavior of these materials and expands the range of hydrogel-based strain sensors.

Graphical Abstract

The choice of polymer used in hydrogels for strain sensing applications has been limited by the high resistivity and capacitance of most hydrogels, which causes hysteresis, low sensitivity, and unstable measurements. By measuring their impedance only in their resistive regime (>10 kHz) at a high sampling rate, hysteresis is reduced and gauge factor is improved by a factor of 50. Genipin-crosslinked gelatin and dopamine-functionalized poly(ethylene glycol) are used as model hydrogels for this study. Adhesion and biocompatibility are studied for in vivo implementation. The measurement method and implantable capability are demonstrated in porcine models to monitor relative changes in cardiac output using strain sensing.

1. Introduction

Hydrogels are a class of materials that can seamlessly interact with many types of excitable tissue owing to their tunable low Young’s moduli (<100 kPa) and large equilibrium swelling in water (>95 vol%).^[1,2] In particular, hydrogels have recently been explored as materials for strain sensors.^[3] By imparting conductivity into a hydrogel, flexible strain sensors with storage moduli as low as 100 Pa and maximum strains greater than 500% have been demonstrated.^[4] For example, hydrogel-based strain sensors incorporated with carbon nanotubes,^[5] novel conjugated polymers,^[6] conductive nanoparticles,^[7] and MXenes^[7] have been explored. Although the many combinations of polymers, fillers, and solvents allow for a wide range of possible hydrogels, the ideal hydrogel strain sensor must balance performance metrics such as gauge factor, biocompatibility, and operability including facile handling and tissue integration.^[8] To achieve suitable performance, previously reported hydrogel-based strain sensors use potentially toxic conductive fillers or novel conjugated polymers, which face rigid regulatory barriers.^[9]

Mussel-inspired dopamine-functionalized poly(ethylene glycol) (PEG-Dopa) and genipin-crosslinked gelatin hydrogels are two hydrogel compositions that have seen recent interest in many applications including targeted drug delivery, wound healing, and tissue model applications, among many others.^[10] They both exhibit intrinsic properties that are suitable for use as implantable hydrogel-based sensors. The storage modulus of cross-linked PEG-Dopa+[Fe³⁺] networks can approach G’ = 57 ± 8 Pa and adhesiveness up to 1.5 J/m² to Parylene C substrates.^[11,12] Poly(ethylene glycol) is used in many FDA-approved devices including sutures, drug delivery devices, and tissue engineering scaffolds.^[13] Similarly, genipin-crosslinked gelatin hydrogels are made of naturally derived materials, are easily synthesized, and have seen wide usage in biomedical applications.^[10,14–17] Although each material can conform and adhere to dynamic, hydrated, curvilinear substrates such as organs, they have thus far not been explored as strain sensors primarily due to their low effective conductivities of ~0.1 S/m. The conduction mechanisms in hydrogels presents challenges when implementing these materials as passive elements in piezoresistive sensing. For example, small gauge factors and hysteresis contribute to poor signal-to-noise ratios and degradation in sensor performance over time. Further, the capacitive nature of electrical measurements through water causes instability, leading to imprecise measurements.^[18] A novel approach to measure the strain of these hydrogels would allow for an implantable strain sensor with demonstrated biocompatibility, adhesiveness, low solvent loss, low hysteresis, and high sensitivity – a combination that has yet to be reported.

Transient effects from electrical measurements of hydrogels may potentially be eliminated by measuring only the impedance in the resistive regime. Hydrogels with water as the mobile phase can be modeled using a modified Randles circuit.^[19] The electrode-electrolyte interface is defined by the double layer capacitance (C_DL) that dominates at the low frequencies (~0–10⁴ Hz). The bulk electrolyte contribution is defined by a parallel RC-circuit of electrolyte solution resistance (R_s) and bulk capacitance (C_bulk) that dominates at higher frequencies (>10⁴ Hz). When ionized water (e.g., phosphate buffered saline) is the solvent, the impedance is almost purely resistive at medium frequencies (~10⁴-10⁶ Hz). By measuring only the impedance at specific frequencies (~10⁴-10⁶ Hz) at a high sampling rate, we hypothesize that common biocompatible hydrogels with low ionic conductance (<1 S/m) can be used as strain sensors with low hysteresis (i.e. high stability). As a demonstration of its implantable capabilities, the hydrogel is utilized as a cardiac strain sensor to measure relative changes in cardiac output, where biocompatibility, adhesiveness, and low hysteresis are prerequisites for measurement. Thus, impedance-based strain sensing could enable the usage of biocompatible hydrogels with no conductive fillers as a next-generation implantable electronic material.

2. Results and Discussion

2.1. Preparation of hydrogel coupons and sensors

PEG-Dopa (Figure 1a) and genipin-crosslinked gelatin (Figure 1b) were chosen as implantable strain sensors for several reasons. Catechol-bearing PEG-Dopa exhibits properties that are suitable for use in implantable devices, such as biocompatibility,^[20] low storage modulus (G’ = 10²-10⁴),^[12] adhesiveness to common substrates such as Parylene C,^[11] and demonstrated bioresorpability.^[21] Genipin-crosslinked gelatin exhibits similar properties, but has a higher storage modulus (G’ = 10⁴-10⁵ Pa).^[16,22,23] However, genipin-crosslinked gelatin is simpler and more cost-effective to synthesize, and the storage modulus is within the range of internal organs.

Figure 1.

a) Schematic of the conduction mechanism of a typical hydrogel represented by the proposed equivalent circuit. b,c) Impedance and phase plots of PEG-Dopa at strains ranging from ε = 0–100%. d,e) Impedance and phase plots of genipin-crosslinked gelatin at strains ranging from ε = 10–40%.

Upon synthesis and lyophilization of PEG-Dopa precursors, hydrogel coupons were prepared by mixing the precursor with Fe³⁺ ions in solution and injecting the mixture into polytetrafluoroethylene (PTFE) molds (Figure S1, Supporting Information). For hydrogels used in the implantable device, a custom polyimide flexible PCB was placed between two PTFE molds before injecting the precursor solution for increased robustness. Genipin-crosslinked gelatin hydrogel coupons and hydrogel devices were prepared similarly after mixing gelatin, glycerol, water, and genipin as described in the Experimental Section.

2.2. Electrochemical characterization of hydrogels

Electrochemical impedance spectroscopy (EIS) was performed on hydrogel coupons to calculate relevant the figures-of-merit that are relevant to strain sensing. Resistance measurement of hydrogels using DC voltage can result in electrochemical anomalies due to charge accumulation at the electrode-electrolyte interface and the possibility of charge transfer reactions.^[19] By measuring the impedance of each hydrogel using AC voltage across a spectrum of frequencies, an equivalent circuit model can be determined. The Bode plots of genipin-crosslinked gelatin and PEG-Dopa are shown in Figure 1b-e. The curvatures of each plot follow similar trends and can be modeled using the proposed equivalent circuit (Figure 1a), where R_elec is the resistance of the electrodes and leads, C_DL is the capacitance of the double layer at the interface, R_s is the solution resistance, and C_bulk is the bulk capacitance of the hydrogel. The impedance and phase plots for each hydrogel were recorded for strains between 0–100% (Figure 1b-e), and their impedance and phase at 10 kHz were tabulated in Table S1 (Supporting Information) for PEG-Dopa and Table S2 (Supporting Information) for genipin-crosslinked gelatin. At 10 kHz, the phase was found to be ϕ_PEG-DOPA = (−2.14 ± .05)^o and ϕ_Gelapin = (−13.43 ± .30)^o for PEG-Dopa and genipin-crosslinked gelatin, respectively. These results suggest that the majority contribution from the solution resistance for the impedance, while the impedance magnitude was found to increase with strain.

Cyclic voltammetry (CV) was conducted on each hydrogel to determine voltages at which electrochemical reactions may be occurring (Figure S3-S4, Supporting Information). Resistance and the voltage required to measure that resistance are directly related when the current is set following Ohm’s law $\left(V=IR\right)$. Minimum input currents for lab-grade multimeters (e.g. Keithley 2400) are typically on the order of 10 uA. For a hydrogel with a resistance of 100 kΩ, this generates a potential of ~1 V (vs. Ag/AgCl) which may lead to excursions from the water stability window.^[18] Electrochemical reactions occurring in PEG-Dopa (Figure S3, Supporting Information) and genipin-crosslinked gelatin (Figure S4, Supporting Information) were investigated using an Ag/AgCl reference electrode. Minimum voltages of 40–60 mV allowed for the presence of electrochemical reactions to be detected, indicated by peak currents at those voltages for both genipin-crosslinked gelatin and PEG-Dopa. The presence of electrochemical reactions at voltages greater than 40–60 mV causes increases in solution resistance over the duration of the measurement, making long-term DC resistance measurements unsuitable for these hydrogels (Figure S2, Supporting Information).

2.3. Impedance-based piezoresistive strain sensing

Hydrogels, including genipin-crosslinked gelatin and PEG-Dopa, can be modeled using the proposed equivalent circuit (Figure 1a), for which the transfer function is the following:

$$Z=R_{elec}+j\omega C_{DL}+\frac{1}{\frac{1}{R_s}+\frac{1}{j\omega C_{bulk}}}$$ (1)

where Z is the impedance, j is the imaginary unit, and ω is the frequency. At low frequencies (i.e. approaching the DC regime), the capacitive character of the hydrogel dominates, leading to a high time constant and unstable measurements. At high frequencies, the double layer capacitance, which is typically orders of magnitude larger than the bulk capacitance, approaches zero and the impedance is nearly purely resistive (Figure 1c,e). For genipin-crosslinked gelatin, frequencies of f > 2 × 10⁵ Hz allow for contributions from the bulk capacitance, resulting in phases greater than zero. By measuring the impedance at 10⁵ < f < 10⁶ Hz, the impedance can be modeled as purely resistive.

As a direct comparison between DC resistance and high frequency impedance measurement of hydrogels, step input strains of 15 and 30% were performed while measuring the DC resistance with an input current of 10 μA on genipin-crosslinked gelatin and PEG-Dopa hydrogels (Figure 2). The result for the DC measurement of the hydrogel was highly unstable (Figure 2a,b). A base resistance value (R_o) could not be obtained for measuring the change in resistance (ΔR/R_o) due to the continual drift caused by the capacitive nature of the hydrogel at DC. The drift was more significant in PEG-Dopa due to the presence of additional electrochemical reactions caused by oxidative groups (Figure S4, Supporting Information). The experiment was repeated while measuring the impedance at f =10 kHz for genipin-crosslinked gelatin (Figure 2c) and PEG-Dopa (Figure 2d). The base impedance was stable across the full duration of the measurement, and no detectable hysteresis was observed.

Figure 2.

Comparison of electromechanical properties of genipin-crosslinked gelatin (gelapin) and PEG-Dopa using the DC resistance method and the high frequency impedance-based method. a,b) The DC resistance response to step input strains of 15% and 30% for a) gelapin and b) PEG-Dopa. c,d) The 10 kHz impedance response to step input strains of 15% and 30% for c) gelapin and d) PEG-Dopa and its linearity for g) gelapin and h) PEG-Dopa. e,f) Stress-strain cycles at various strains for e) gelapin and f) PEG-Dopa.

To further evaluate the impedance-based method for strain-sensing applications, linearity in the change in impedance at 10 kHz with respect to strain was measured. In Figures 2g-h, the change in impedance at 10 kHz different strains for PEG-Dopa and genipin-crosslinked gelatin respectively are plotted. The result was found to be highly linear (R² = 0.997 and 0.998 respectively). Interestingly, gauge factor was also significantly improved to 1.02±0.05 and 1.46±0.05 respectively, showing a 30–50x improvement over the approximate gauge factor found using DC measurements. Cyclic measurements were more stable for the impedance-based measurement, while the DC resistance measurement demonstrated drift (Figure S9, Supporting Information). Furthermore, the input voltage used to measure impedance at these frequencies was low (20 mV), avoiding potential electrochemical reactions previously seen in DC measurements (Figure S3).

To evaluate the hydrogels’ potential as a standalone strain sensor for cardiac in vivo applications, their adhesion to aortic tissue was evaluated (Figure 3). Testing conditions were based on the standard tensile test for tissue adhesives (ASTM F2258). First, ABS holders were 3D printed to a size that is compatible with the machine grips. Fresh aortic tissue was adhered to the 3D printed holders using cyanoacrylate surgical glue. A hydrogel with dimensions 1cm x 1cm x 2 mm was then placed between the holders (Figure 3a). The hydrogel was compressed to a distance of 1 mm and held for 5 seconds before applying a constant strain rate between 10⁻² to 10⁰ mm/s, including the 0.033 mm/s specified in STM F2258 (Figure 3b). It was found that the adhesive work for gelapin varied between .37±.11 to 1.32±.47 J/m² and for PEG-Dopa 0.97±.20 to 4.20±1.86 J/m², increasing per strain rate (Figure 3c). The adhesive work dependance on the strain rate is consistent with that observed for viscoelastic polymers, which have dissipated energy typically proportional to the separated rate with a power between 0.4 and 1.^[24]

Figure 3.

Adhesiveness of gelapin and PEG-Dopa. a) Experimental setup of the adhesion test based on the standard tensile test for tissue adhesives (ASTM F2258). b) Method for determining adhesive work for each sample. c) Adhesive work of gelapin and PEG-Dopa versus porcine aortic tissue.

2.4. In vivo deployment of hydrogel-based strain sensors

Methods of monitoring of cardiac output to evaluate patients after cardiac surgery are currently limited in their non-invasiveness, continuity, and/or portability. The pulmonary artery catheter (PAC), which requires catheterization through the chambers of the heart, is highly invasive and may cause further complications to the heart.^[25] Cardiac echocardiography uses ultrasound to estimate changes in cross-sectional area of the heart (i.e. fractional shortening), thereby calculating cardiac output, but requires the presence of a clinician and is not portable. Relative changes in cardiac output may be continuously monitored using a strain sensor attached to a ventricle. Impedance-based strain sensing using hydrogels may be an ideal approach for this purpose because it allows for the use of biocompatible and clinically approved hydrogels with properties desirable for implantable applications (e.g. adhesiveness, biodegradability) that were previously unexplored due to limitations in sensing ability.

To assess the biocompatibility for each hydrogel for an acute in vivo implantation, the growth of C2C12 mouse myoblasts was monitored using the hydrogels as a substrate (Figure S5). Cell viability determined by scanning laser confocal fluorescence microscopy was calculated to be: 93.7±4.0% (control); 95.1±7.4% (gelapin); and 89.8±4.9% (PEG-Dopa) (n = 4 samples per condition, n = 5 images per sample, Figure S5b), demonstrating that the hydrogels could be used acutely with minimal cytotoxicity.^[20,22,23]

As a demonstration of the in vivo capability of impedance-based hydrogel strain sensing, hydrogel strain sensors were placed on the left ventricle (LV) in two pigs using genipin-crosslinked gelatin. The hydrogel was formed around a custom flexible printed circuit board (PCB) to ensure uniform dimensions and robustness, thereby requiring no additional adhesive between the hydrogel and the electrodes (Figure 4a, Figure S1, Supporting Information). To simulate a low cardiac output state that can be seen after open heart surgery, the pulmonary artery (PA) was compressed for 10 s, thereby limiting blood flow and showing a visible reduction in LV volume (Movie S1). Strain from the hydrogel sensor was measured using single frequency impedance measurements at 10 kHz with a sampling rate of 10 Hz and compared to a PAC, the clinical standard. Mixed venous oxygen saturation (SvO₂) was used for the clinical measurement as opposed to cardiac output due to the insufficient sampling rate of cardiac output measurement from the clinical PAC (Figure S5-S6, Supporting Information). Strain from the hydrogel changed from 9.0±0.4% before the PA compression to 3.1±0.4% during the compression and returned to 7.4±2.9% after the compression for pig 1. SvO₂ changed from 69.6±4.2% to 44.7±4.7% to 68.2±4.4% during the sequence. Similarly, strains for pig 2 changed from 3.8±0.9% to 1.7±0.5% to 3.2±0.8% during the sequence, while the SvO₂ changed from 78.2±1.6% to 61.5±4.9% to 77.4±1.2% before, during, and after the PA compression respectively (Table S3, Supporting Information). Strain and SvO₂ values were found to vary significantly with p < 0.001 when comparing groups before versus during compression and groups during versus after compression using a one-way ANOVA test (Figures 4d). The in vivo demonstration indicated that a hydrogel strain sensor may be a potential method to track relative changes to cardiac state using bioresorbable materials that are compatible with the ionically conductive properties of bodily fluids.

Figure 4.

In vivo demonstration of the impedance-based strain sensing method. a) Schematic of the in vivo experiment. b) Image of the hydrogel strain sensor attached to the left ventricle and volume reduction of the heart during constriction of the pulmonary artery. Scale bars: 2 cm. c) Measured LV strain over time before, during, and after PA constriction. d) Statistical comparison of strain and SvO₂ values before, during, and after PA constriction for all trials. *** p <0.001.

3. Conclusion

PEG-Dopa and genipin-crosslinked gelatin are two hydrogel materials that possess many of the desired properties for implantable electronics such as low tunable moduli, adhesiveness, stretchability, and biocompatibility, but have not seen use as a conductor due to their high resistivity and instability. The impedance of such hydrogels is dominated by their double layer capacitance at low frequencies, which contributes to high hysteresis, instability, and comparatively small gauge factor when compared to other piezeoresistive sensors. By measuring the impedance only at a high frequency, hydrogels with no conductive fillers are enabled for electromechanical sensing. PEG-Dopa and genipin-crosslinked gelatin were demonstrated for this purpose and showed little to no hysteresis and 30–50x improved gauge factor. Its capability as an implantable electronic device was demonstrated by measuring relative changes in cardiac state in an in vivo porcine experiment. It is envisioned that the impedance-based strain sensing method presented in this report could be generalized to enable the usage of a wide variety of biocompatible hydrogels for implantable electronics.

4. Experimental Section

Synthesis and Preparation of Genipin-crosslinked Gelatin Hydrogels

Genipin-crosslinked gelatin hydrogels were synthesized using previously published protocol^[16] with modifications. Gelatin (10% w/w, 175 bloom type A, Sigma Aldrich) was heated in 1:1 phosphate-buffered saline (PBS):glycerol at 65 ^oC until fully dissolved. Genipin (10% w/w; Wilshire, Princeton, NJ) was separately dissolved in 3:7 ethanol:PBS at room temperature. 100 μL of genipin solution was pipetted into the gelatin solution and stirred at 65 ^oC for approximately 5 minutes until crosslinking was indicated by a color change, forming a pre-hydrogel. The pre-hydrogel was poured into a custom Teflon mold and gelled at room temperature for 1 h. The samples were then stored overnight in sealed humid containers at 4 ^oC.

Synthesis and Preparation of PEG-Dopa Hydrogels

All chemicals were ordered from Sigma Aldrich unless otherwise specified. PEG-Dopa hydrogels synthesized using previously published protocol.^[11,12] Four-arm poly(ethylene glycol) succinimidyl carboxymethyl ester (M_w ~ 10,000 g mol⁻¹; Jenkem Technology Ltd, Plano, TX) was combined with dopamine hydrochloride (2:3 mol ratio, neutralized with N-methylmorpholine) in anhydrous N,N-dimethylformamide (DMF) for 18 h. The product was dialyzed in regenerated cellulose tubing (Spectra/Por® 3, 9.3 mL/cm; Repligen, Waltham, MA) against H₂O with pH ~ 3.5 (titrated with 1 M HCl) for 24 h and then against doubly-distilled H₂O for 2 h. The PEG-Dopa precursor solution was lyophilized and stored at −15 ^oC until further use. In separate containers, 0.05 M of Tris base in PBS (Tris solution) and 0.15 M of FeCl₃ and 0.05 M of Tris base in PBS (FeCl₃ solution) were prepared. PEG-Dopa hydrogels then were prepared by adding 250 μL of Tris solution and 250 μL of FeCl₃ solution to 0.2 g (0.08 M) of PEG-Dopa, mixing in container, and dispensing into custom Teflon molds using hydrogel dimensions of 25×5×5 mm (Figure S1).

Electrochemical and Mechanical Characterization of Hydrogels

Potentiostatic electrochemical impedance spectroscopy was conducted between f=10⁻² Hz to 10⁶ Hz with an amplitude of 20 mV (Interface 1000e, Gamry Instruments, Warminster, PA) while the hydrogel strain was controlled using a custom machined micromanipulator. Cyclic voltammetry was conducted with the same instrument at −3 to 3 V and −1 to 1 V using an Ag/AgCl reference electrode, a platinum counter electrode, and a gold-coated custom flexible PCB working electrode at scan rates between 40–500 mV/s at 21.4 ^oC. Mechanical properties of hydrogels were measured using uniaxial stress-strain measurements (Instron 5943 with Bluehill Software, Instron, Norwood, MA) at 21.4 ^oC and 43% relative humidity. Electromechanical properties with the hydrogel on the Instron were measured at 10 kHz with an amplitude of 20 mV and a sampling rate of 10 Hz (ADG5940, Analog Devices, Wilmington, MA). DC resistance measurements were conducted using a 10 μA input current (Keithley 2400, Tektronix Inc., Beaverton, OR).

Cell Growth

To prepare substrates for mammalian cell growth, 100 μL of either gelapin or PEG-Dopa was placed on 10 mm glass substrates in a 24-well plate and polymerized at the same conditions as above. Then 1 mL of EtOH was added to each well for 1 h and placed under UV for sterilization. The hydrogels were then washed by in 1x PBS (VWR, catalog no. 02–0119-0500) 3 times for 5 minutes each. To improve cell adhesion to glass and hydrogel substrates, each well was coated with a fibronectin (Advanced BioMatrix, catalog no. 5050–1MG) solution at 50 µg/mL and incubated at room temperature for 1 hour in the cell culture hood. The samples were again washed with 1x PBS 3 times for 5 minutes each and then culture media: 10% fetal bovine serum (Invitrogen, catalog no. 10082147) + 1% penicillin/streptomycin (ThermoFisher, catalog no. 15140122) in Dulbecco’s modified Eagle medium (Corning, catalog no. 15–013-CM). Mouse myoblast cells (C2C12) (ATCC, catalog no. CRL-1772, passage number p3 - p5) were then added to each sample with a cell density of 5,000/cm² counted using a hemocytometer and incubated at 37 ^oC for 24 h with 5% CO₂. For the viability assay, the samples were labeled with a Live/Dead assay kit (ThermoFisher, catalog no. L3224). The live cells were labeled with 2 μM Calcein acetoxymethyl (Calcein); the dead cells were labeled with 4 μM ethidium homodimer-1 (EHD); and the nuclei were labeled with 1 μM Hoechst 33342 (Hoescht, ThermoFisher, catalog no. 62249). After adding the dyes, the samples were incubated for 15 min at 37 °C with 5% CO₂. The samples were washed 3 times with warm carbogenated 1x Hanks’ Balanced Salt Solution (ThermoFisher, catalog no. 14065–056) supplied with 20 mM HEPES solution (Sigma-Aldrich, catalog no. H3537–500ML). Scanning laser confocal fluorescence imaging were performed using an upright confocal microscope and a 20×/0.50 NA water immersion objective (Nikon). Viability quantification was performed for 4 samples per condition and 5 images per dish using ImageJ software. The viability was calculated as (nucleus # - dead #) / nucleus #.

In Vivo Cardiac Strain Measurements

This portion of the experiment was performed using 40–60 kg Yorkshire swine. IACUC approval was approved prior to the experimentation. The pig underwent a deep plane of anesthesia and a full sternotomy was completed. The pericardium was opened and the heart was visualized. A PAC was placed through the superior vena cava and “floated” into the right pulmonary artery. Measurements were recorded that included cardiac output and stroke volume. The left ventricle was then exposed and the genipin-crosslinked gelatin hydrogel sensor was carefully placed on the anterior apical region of the left ventricle after applying small drops of cyanoacrylate (Loctite 4011, Henkel Corps., Stamford, CT) at the ends of the hydrogel. Strain was recorded from the sensor by measuring impedance at 10 kHz with an amplitude of 20 mV and a sampling rate of 10 Hz (ADG5940, Analog Devices, Wilmington, MA). Intermittent manual compression of the main pulmonary artery was done to create an acute low cardiac output state. Release of the compression allowed the heart to recover. Recordings were done on the PAC and the sensor through these changes.

Statistical Analysis

Calculated values are reported as (mean ± SD) unless otherwise specified. MATLAB (v.R2021a) was used to assess the statistical significance of all comparison studies in this Article. Data distribution was assumed to be normal for all parametric tests, but was not formally tested. In the statistical analysis for comparison between multiple samples, a one-way ANOVA test followed by Tukey’s multiple comparison test was conducted.