Dynamic Corneal Response for Different Air-Puff Stimuli in Ex Vivo Animal Model Measured with SS-OCT System with Two Sample Arms

Featured Application

Measurement of air-puff induced dynamic corneal response (DCR) has several possible applications in ophthalmology and vision science. DCR provides information on corneal biomechanics, thus allowing for more precise intraocular pressure measurement, which is important in glaucoma diagnosis and management. Moreover, DCR helps identify the eyes at risk of ectatic diseases such as keratoconus.

Abstract

This study demonstrates the use of a dual-sample arm swept-source optical coherence tomography (SS-OCT) instrument coupled with air-puff stimulation to assess corneal displacement in an ex vivo porcine eye model. The air-puff SS-OCT system enables correction of corneal deformation for eye globe retraction, providing a comprehensive quantitative analysis of corneal apex deformation dynamics under varying intraocular pressure (IOP) levels and air-puff stimulus strengths. Spatio-temporal characterization of those stimuli was performed. The results showed that the cornea stiffened with increased IOP, and reducing the stimulus amplitude decreased the correlation between parameters describing corneal dynamics and IOP. However, maximum displacement and corneal response time exhibited very strong correlations regardless of the strength of the applied air-puff. These findings suggest that softening air-puff stimulation may impact the accuracy of non-contact tonometers in measuring IOP and corneal biomechanical properties.

1. Introduction

The research on air-induced dynamic deformations of the corneal surface has so far relied exclusively on limited air pulses strengths proposed in commercial tonometers [1,2,3,4,5]. Commercial air-puff tonometers measure corneal deformation indirectly, based on the amount of light reflected from the corneal surface [6]. Over a decade ago, improvements enabling direct measurements of these deformations were proposed. Notably, Alonso-Caneiro et al. combined an air-puff chamber with optical coherence tomography (OCT) [7]. This approach was explored in both ex vivo and in vivo measurements over the years [8,9,10,11,12]. Shortly after the first publication on air-puff OCT, Oculus GmbH introduced the commercial Corvis ST device, which uses a Scheimpflug camera [13,14]. Similar to the OCT-based system, the Corvis ST prompted numerous studies focusing on clinical applications, data analysis and interpretation [15,16,17,18], as well as computer simulations (modeling) of the biomechanical properties of corneal and scleral structures [19,20,21].

To date, the application of custom air pulses with different spatial and temporal properties, and their impact on corneal-surface deformations, has not been explored. Such customization could be motivated by the need to optimize the eye and air-puff stimuli distance, which is particularly important when imaging beyond the central cornea (i.e., the corneal apex) [22]. However, achieving pulse customization requires specialized tools to characterize its spatial and temporal properties.

Because corneal response to air-pulse stimulation also depends on intraocular pressure (IOP), air-puff OCT studies are typically performed at various IOP levels. These levels may be reached through its natural fluctuations (in vivo measurements) [23] or by using well-controlled IOP adjustment systems (ex vivo measurements) [24].

To focus solely on the corneal surface, one must exclude (or subtract) the dynamic behavior of the rest of the eye. For in vivo measurements, methods such as peripheral corneal displacement and ultra-long-range OCT have been proposed to correct for eye retraction [11,25]. However, for ex vivo porcine eye studies, postmortem changes (e.g., the dramatic reduction in crystalline-lens transparency) necessitate alternative approaches.

In this paper, we evaluate how the spatial and temporal characteristics of custom air pulses influence the dynamic displacement of the ex vivo porcine corneal surface under various IOP levels. To isolate corneal dynamics from overall eyeball movements, we developed a dedicated dual-arm air-puff OCT system. Finally, we investigate how the customization of these air pulses affects the parameters extracted from the resulting temporal displacement profiles.

2. Materials and Methods

The instrumentation used in this study consists of the modules for air-puff stimulation, imaging induced eye deformation, and IOP control.

2.1. Air-Puff Stimulation

The air pulses were generated by the rapid movement of a piston, driven by a solenoid motor (Figure 1A). The activation of these pulses was synchronized with OCT acquisition. By adjusting the capacitor’s discharging voltage (C in Figure 1A)), we customized the spatio-temporal characteristics of the air pulses.

Three distinct settings for high-, medium-, and low-amplitude stimuli were selected (70 V, 55 V, and 45 V, respectively) and are referred to throughout the article as Type 1, Type 2, and Type 3 air-puffs, respectively. The generated air pulses were characterized using a pressure sensor (Ps1; SenSym, Milpitas, CA, USA) incorporated into the air-puff chamber (Figure 1A). Ten readouts from the Ps1 sensor were used to extract both the peak pressure and the duration of the air pulse temporal profiles (Figure 1B). This duration was defined as the time span during which the pressure magnitude exceeded 10% of the peak value. The mean and standard deviations of the amplitude and duration of the air puff were calculated. The sensor Ps1 was also calibrated by an external force sensor, which allowed for expressing the pressure signal from Ps1 in force units. The procedure of calibration was described in an earlier study [11].

We used an MPX5050D pressure sensor (NXP Freescale Semiconductor Inc., Austin, TX, USA) to characterize the spatial profile of each tested air-puff configuration (Figure 1A). The sensor was moved along the Z axis to determine axial air-puff pressure amplitude profiles. The translation range was 18 mm, and the measurements were taken in 1 mm steps. Then, the sensor was placed at the distance of 10 mm in front of the air-puff nozzle, which corresponded to the position of the corneal apex in ex vivo measurements. To investigate the lateral (X and Y axis) distribution of the air pulse, the sensor was moved along the X and Y axes in 0.5 mm increments, covering a range of ±2 mm distance around the center. Five measurements were completed in each sensor position.

2.2. Dual-Arm Air-Puff SS-OCT Setup

We examined the corneal response to three types of air-puff stimuli under systematically controlled IOP. The setup (Figure 2A,B) comprised of: (i) a customized dual-sample-arm swept-source OCT (SS-OCT) system, (ii) an air-puff chamber, and (iii) an IOP-control system, consisting of a saline column connected to the eye via plastic tubing and monitored by a differential pressure sensor.

The core component of the SS-OCT system was a short external cavity swept light source operating at a central wavelength of 1310 nm, with a bandwidth of 100 nm and a sweep rate of 50 kHz (Axsun Technologies Inc., Billerica, MA, USA). To enable simultaneous imaging of the anterior and posterior segments of the eye, a 50/50 fiber coupler was integrated into the sample arm, creating two sample arms with distinct optical path lengths. These sample beams illuminated the cornea and sclera, respectively (Figure 2A,B). The collected light from the sample interfered with the reference arm signal. The path lengths of the sample arms were adjusted to ensure that the corresponding OCT images did not overlap (Figure 2C). This approach utilized the depth multiplexing method, commonly employed in polarization-sensitive OCT and space-division multiplexed OCT systems [26,27,28]. The interference signal was detected using a balanced photodetector (PDB110C-AC, Thorlabs Inc., Newton, NJ, USA) and digitized with an acquisition board (Gage Compuscope 14200, Gage Applied Inc., Lockport, IL, USA; 200 MS/s, 14-bit resolution). The system’s sensitivity was measured to be 99 dB for both sample arms. The measured depth range was 9.33 mm, and the axial resolution was measured to be 9.3 μm.

To induce dynamic responses in the cornea, the corneal sample arm was coupled to an air-puff chamber dismounted from a commercial non-contact tonometer (XPert NCT; Reichert Inc., Buffalo, NY, USA) [7]. The beam illuminating the cornea was aligned collinearly with the air-puff stimulus direction. The pressure generated within the air-puff chamber during activation was measured using a pressure sensor (Ps1 in Figure 2A; SenSym, Milpitas, CA, USA) and recorded alongside the OCT signal. The optics of the sample arms and the air-puff chamber were mechanically separated to minimize interferometric signal instabilities caused by vibrations from the air-puff piston movement.

Alignment of the imaging beam with the corneal apex was achieved via lateral scanning in preview mode. The apex was identified by observing the central reflex in both horizontal and vertical sectional images. The scleral sample arm was mounted on a positioning system (translational stages) to focus the optical beam on the posterior pole of the eye. After aligning the eye globe with respect to the illumination beams, a series of 4000 A-scans at both poles of the eye was acquired (M-scan) (Figure 2C).

2.3. IOP Control System

The scheme of the IOP-control system is presented in Figure 2A. A saline column was connected to the eye globe using plastic tubing and an 18G needle. The IOP generated in the eye was measured using a differential pressure sensor (Ps2 in Figure 2A; Sen-02, Experimetria, Budapest, Hungary) integrated with the needle and an oscilloscope (DSO5202B, Qingdao Hantek Electronic Co., Ltd., Qingdao, China). The height of the saline column was manually adjusted using a syringe mounted on a dedicated holder, allowing precise regulation of the saline volume within the tubing.

2.4. Ex Vivo Eye Preparation

Freshly enucleated, virgin porcine eyes (n = 9) were obtained from a certified local abattoir and used in accordance with ethical, safety, and sanitary guidelines for animal tissue research. After enucleation, the eyes were stored in saline solution at refrigeration temperatures and utilized within 8 h postmortem. Selection criteria included intact corneas and the absence of edema.

Each eye was placed in a custom mount for measurements. To prevent corneal dehydration, the mount was positioned inside a wet chamber lined with moistened cotton at the bottom. Humidity and temperature inside the chamber were continuously monitored throughout the sessions. No artificial eye drops were applied to maintain corneal hydration. Openings in the chamber and mount were strategically positioned to allow the delivery of illuminating beams and connection to the IOP control system (Figure 2B).

2.5. Pilot Testing for Protocol Refinement

Before the main study, we performed optimization tests to refine the research protocol. Because corneal properties can be influenced by tissue hydration (ongoing dehydration or edema), stable and controlled environmental conditions are essential for ex vivo eye examinations. In this phase, we also evaluated the repeatability of the dynamic air-puff–induced corneal response.

To investigate how ambient conditions (i.e., humidity and temperature) and repeated mechanical stimulation affect the corneal tissue, we designed the following experiment. The humidity and temperature were maintained at (65 ± 5)% and (21 ± 1) °C, respectively, mirroring the conditions used in the main study. Three porcine ex vivo eyes were tested at a constant intraocular pressure (15 mmHg) using a prototype air-puff SS-OCT system. Each cornea was stimulated 40 times with Type 1 air puffs at a maximum force of 128 mN, applied at approximately one-minute intervals. The corneal and scleral responses were measured after each stimulus, and the entire procedure took about 50 min.

2.6. Final Study Protocol

The impact of air pulse stimulation characteristics on dynamic corneal displacement was investigated using a multi-stage protocol (

Table 1. Protocol for the main study.

	Air-Puff Type	IOP [mmHg]	Number of OCT Scans
Pre-conditioning	Type 1	15	12
Stage 1	Type 1	10–30 (step 5)	5 (1 per each IOP level)
Stage 2	Type 2	10–30 (step 5)	5 (1 per each IOP level)
Stage 3	Type 3	10–30 (step 5)	5 (1 per each IOP level)

). Initially, the tested eye underwent pre-conditioning with a series of 12 high-amplitude air-puffs (Type 1 air-puff) while maintaining a constant IOP level of 15 mmHg. Following this pre-conditioning phase, separate measurements were performed for each air pulse stimulus type across a range of IOP levels.

In the first stage (Stage 1 in Table 1), Type 1 air-puff stimuli were applied, and OCT measurements were conducted for IOP levels ranging from 10 to 30 mmHg, incremented in 5 mmHg steps. Subsequently, the same measurement protocol was repeated using Type 2 and Type 3 air-puff stimuli (Stages 2 and 3 in Table 1), respectively. This approach allowed for a systematic evaluation of how different air pulse characteristics influence corneal displacement dynamics under varying IOP conditions.

The entire protocol for each eye included the acquisition of 87 datasets in total. Although a single measurement took ca. 80 ms, the duration of the experiment was ca. 90 min due to the time required for data saving, data transfer, and the changes in IOP and air-puff parameters.

2.7. Data Analysis

Custom design of the instrument facilitated recording the dynamic response of the cornea and sclera before, throughout, and after the air-puff stimulus is applied to the cornea apex. To provide more advanced analysis of the cornea displacement, corresponding pressure values (generated within the chamber) were collected and, in further steps, recalculated into force values.

After segmentation of all interfaces of interest (temporal displacement of both cornea surfaces and the back of the sclera surface; cf. Figure 2C), optical distances measured by OCT were mapped to geometrical ones by correcting for the refractive index of the cornea n = 1.389 [7]. Temporal evolution of the central corneal thickness (CCT) was determined as a difference between anterior and posterior cornea surface profiles. We further analyzed CCT before air-puff activation (CCT_bef) and during maximum displacement amplitude (CCT_max). To isolate the dynamic response of the cornea (corneal displacement in Figure 3A) from the eye retraction, we subtracted scleral displacement profiles from the cornea profiles.

The maximum apex displacement (MAD) and maximum eye retraction (MER) were extracted accordingly from segmented temporal profiles of the cornea and the sclera, respectively (Figure 3A). We identified two distinct phases in the displacement profiles: loading phase (L in Figure 3A), representing time span from the beginning of the air-puff stimulus to the maximum displacement, and recovery phase (R in Figure 3A), corresponding to the time span from maximum displacement to the instant when the cornea apex reaches its initial axial position. Later, we determined the velocity of the cornea surface by calculating time derivative of the cornea displacement profile. From this signal, the maximum velocity was extracted for the loading phase (CV_L) and the recovery phase (CV_R). The time interval between those velocity peaks was also measured (Δt_v), as presented in Figure 3B.

The time required for the cornea to respond to the air-puff stimulus was evaluated. Corneal response time (CRT in Figure 3C) was defined as the interval between the moment when the air-puff force reached 10% of its maximum amplitude and the moment when the corneal displacement reached 10% of MAD.

The corneal response to the air-puff stimulus differs between the loading and recovery phases. To characterize this response, the air-puff-induced force F(t) was plotted against corneal surface displacement (hysteresis curve), as shown in Figure 3D. The relation between the signal from the pressure sensor and force was used to calculate F(t) [11]. The difference in corneal movement between the loading and recovery phases was quantified using the hysteresis area (HA). To analyze the shape of the force-displacement hysteresis, the slope S of the loading curve is defined as the slope of the line connecting points corresponding to 10% and 90% of the MAD on the loading curve (Figure 3D).

2.8. Statistical Analysis

Statistical analysis was performed by using PS IMAGO PRO 8.0 and Microsoft Excel 2013 software. The Shapiro–Wilk test assessed the normality of the parameter distributions. To detect differences between groups, we used the Wilcoxson signed-rank sum, with significance levels adjusted for multiple comparisons using the formula ${\alpha }_A={\displaystyle \frac{\alpha }{N}}$, where α = 0.05 and N stands for the number of comparisons. The relationships between selected parameters and IOP were evaluated using Spearman correlation coefficients r_s. The significance of correlation coefficients was tested at α = 0.05. The results are presented as box plots showing median values, interquartile ranges, and outliers.

3. Results

We performed characterization of all three types of air-puff stimuli. We observed that Type 2 and Type 3 air pulses reached 63% and 37.9% of the peak pressure of the Type 1 pulse, respectively. The reduction in peak pressure in our custom air-puff system was associated with an elongation of pulse duration: from 17.31 ms for Type 1 to 19.57 ms for Type 2, and 21.99 ms for Type 3 (

Table 2. The maximum force and time duration of the selected air puff types.

Air-Puff Type	Capacitor Charging Voltage (V)	Peak Force (mN)	Time Duration (ms)
Type 1	70	127.7 ± 0.3	17.31 ± 0.02
Type 2	55	79.3 ± 0.1	19.57 ± 0.02
Type 3	45	47.3 ± 0.1	21.99 ± 0.03

).

We investigated spatial distributions of three types of air puffs. The results, shown in Figure 4A, indicate that peak pressure remained constant (variations < 1.3%) within the distance of 10 mm from the nozzle but declined more significantly beyond this point, dropping to 30% of the initial value at 18 mm. It is therefore crucial to ensure the measurements are taken at a distance not exceeding 10 mm from the nozzle end. Moreover, the horizontal (X) and vertical (Y) pressure profiles taken in a plane 10 mm away from the nozzle end revealed differences in peak pressure but generally followed a Gaussian-like distribution (Figure 4B,C).

Figure 2C shows the cornea and sclera before, during, and after mechanical stimulation. Because dehydration reduces corneal thickness, we monitored CCT prior to each stimulation to identify any changes over time. Finally, we used these data to assess the repeatability of the air-puff–induced displacement amplitude at the corneal apex and in the sclera.

Analysis of the maximum amplitude of corneal displacement (MAD) indicated that the corneal response stabilized after about 12 air-puff stimulations (Figure 5A). Additionally, eye retraction during the experiment suggests that air-puff stimulation induces slight ocular movements (MER in Figure 5B). These findings underscore the importance of monitoring scleral behavior in the posterior eye pole to account for whole-eye movement when measuring corneal displacements.

Figure 5C shows a slight decrease in CCT during repeated mechanical stimulation. For all three tested eyes, CCT remained relatively stable for the first 10–20 min (around measurements 9–17), exhibiting a maximum reduction of about 2%. After approximately 40 min (around measurement 30), CCT decreased by around 5%, which is consistent with the maximum duration of examinations in the main study. The differences in baseline CCT values observed for all three air-puff types are not related to air-puff type, but to the population CCT variation in porcine eyes.

Following the final study protocol, we measured nine porcine eyes ex vivo and analyzed each measurement to quantify the displacement. Figure 6A–E presents the results for key parameters (MAD, MER, CRT, HA, and S) as boxplots, grouped by different IOP levels and air-puff types (colors). MAD, MER, and HA showed negative correlations with IOP. The correlation between MAD and IOP was particularly strong and statistically significant (Figure 6A and

Table 3. Spearman’s correlation coefficients (r_s) for selected parameters versus IOP.

Parameter	Air−Puff Type
	Type 1		Type 2		Type 3
	rs	p	rs	p	rs	p
MAD	−0.972	<0.001	−0.876	<0.001	−0.854	<0.001
MER	−0.455	0.002	−0.392	0.009	−0.105	0.498
CCTbef	−0.208	0.170	−0.171	0.266	−0.227	0.139
CCTmax	−0.150	0.330	−0.052	0.738	−0.177	0.250
CRT	0.963	<0.001	0.880	<0.001	0.748	<0.001
CVL	−0.980	<0.001	−	−	−	−
CVR	−0.906	<0.001	−	−	−	−
Δtv	−0.962	<0.001	−	−	−	−
HA	−0.792	<0.001	−0.670	<0.001	−0.740	<0.001
S	0.967	<0.001	0.842	<0.001	0.855	<0.001

), being highest for Type 1 air-puff (r_s = −0.972) and slightly lower for Type 2 and Type 3 (r_s = 0.876 and r_s = −0.850, respectively).

The eye retraction (MER) was significantly lower (Figure 6B) than corneal displacement (MAD), and MER correlated weakly with IOP. Furthermore, reducing the puff stimulus caused a more pronounced decrease in MER–IOP correlation than in the MAD–IOP correlation (Table 3). CCT measured before or during deformation did not show any statistically significant relation to IOP (Table 3).

A high correlation emerged between CRT and IOP (Figure 6C and Table 3), with a slight reduction in correlation for Type 2 and Type 3 air-puffs (r_s = −0.880 and r_s = 0.748, respectively) compared to Type 1 (r_s = 0.963). Notably, CRT depended on IOP in a nonlinear manner, while MAD, MER, HA, and S varied linearly with IOP (Figure 6A–E). We found a strong negative correlation between HA and IOP and a strong positive correlation between S and IOP for all stimulus strengths (Figure 6D,E and Table 3). For Type 1 air-puff, HA reached zero at the highest IOP, indicating a purely elastic response (green curve in Figure 6F).

We also evaluated the corneal apex velocity during mechanical stimulation to broaden our understanding of corneal dynamics for Type 1 air-puff. Figure 7A shows representative corneal velocity profiles from one eye at several IOP levels (10–30 mmHg in 5 mmHg increments). Positive velocity values correspond to the loading phase of corneal deformation, and velocity reaches zero at the point of maximum displacement (MAD). During the recovery phase, velocity values are negative. The sign of velocity value indicates only the direction of the cornea movement. We analyzed the absolute velocity peaks and the time intervals between these peaks (CV_L, CV_R, and Δt_v). Each showed strong negative correlations with IOP (r_s = −0.980, r_s = −0.906, and r_s = −0.962, respectively, Figure 7B,C).

4. Discussion

Ex vivo studies on corneal biomechanical properties have, so far, involved inflation tests with cornea buttons (artificial anterior chambers) or whole eye globes [29,30,31,32]. Different approaches have been used previously to monitor the dynamic response of the cornea [2,8,9,10,29,33]. In this study, ex vivo inflation test using air-puff non-contact indentation was applied to assess the reaction of porcine eye globes to rapid stimulus. The porcine eye model and the IOP control system provided well-defined experimental conditions that are close to in vivo tests but might be extremely hard to achieve in the in vivo experiment.

The measurements of corneal deformation with air-puff SS-OCT offer several advantages over existing technology that combines air puff stimulation and Scheimpflug imaging (Corvis ST). The high acquisition speed of our system is at least one order of magnitude higher than the frame rate of Scheimpflug camera. However, our data acquisition is confined to a single apical location. Moreover, the axial resolution of OCT images is higher than that in Scheimpflug imaging, which is directly associated with the inherent precision of corneal surface segmentation. Despite differences between OCT- and Scheimpflug-based assessment of air-puff induced deformation, both approaches implement methodologies to correct observed effects for whole eye globe retraction.

We completed hardware modifications with respect to other reports on air-puff SS-OCT to enable more precise assessment of corneal reaction to air-puff stimulus. In particular, an additional optical beam, illuminating the back of the eye, was introduced, forming the dual-sample-arm OCT system. Although the presented design cannot be easily implemented in in vivo experiments, such a cost-efficient hardware change enables easy depth multiplexing, and the measurement with this configuration corresponds to operation of two perfectly synchronized OCT set-ups. Therefore, this approach allows for simultaneous monitoring of both the deformation of the corneal apex and backward movement of the sclera in the posterior pole of the eye globe. As a result, we were able to correct the observed effects for eye retraction.

We investigated the behavior of the porcine corneas under different dynamic mechanical stimuli for different IOP levels. This allowed us to determine the relation between the biomarkers extracted from time-dependent corneal deformation and the IOP. The experiments performed in this report were motivated by the quest for a softer and more tolerable mechanical stimulation used in clinical non-contact tonometers. The study comprised three levels (types) of air-puff pressure that were easily set by the puff driver. Therefore, different loading rates can be achieved. However, it must be emphasized here that modifying a single parameter represents a more scientifically sound approach. Thus, the ideal case would include a full control of both amplitude and duration of the air puff to decouple both effects, but it is difficult to obtain experimentally without complex solutions [12].

Corneal deformation is a complex interplay between geometric and biomechanical parameters of the ocular tissues as well as the external factors such as IOP. We do not consider the effects to be associated with the IOP loading rate as the IOP was changed stepwise [33]. The pilot experiments shown in Figure 2 revealed that corneal and scleral displacements are influenced by the number of air-puff tests, and the effects cannot be simply explained by the associated CCT changes. This effect necessitated the implementation of a pre-conditioning stage. Although this is not a clinically acceptable scenario, we wanted to minimize the unambiguity of the measurements.

The results confirmed that elevated IOP increase the stiffness of the cornea, which is harder to deform [33]. Generally, the corneal displacement (MAD) and eye retraction (MER) amplitudes depend also on the strength of the stimulus. However, we observed that the relation between MAD (or MER) and the IOP became weaker (Spearman correlation coefficients are lower) for softer air puffs (Type 2 and Type 3) in comparison with regular air puff (Type 1). This suggests that it might be more challenging to estimate the IOP from air-puff tonometry with softer stimulus. However, recent studies showed that the application of computational models help to maintain the accuracy of IOP estimation by taking into account different air pressure loading for patient-specific parameters of the cornea [22,34].

CRT increased with the IOP, which confirms that the start of deformation of eyes with higher IOP requires higher values of the air-puff pressure. In addition, the relation between CRT and IOP was not linear, and the influence of the air-puff amplitude on observed CRT was less defined.

The results showed no statistically significant impact of the IOP on the CCT measured either before applying the stimulus or during the maximum cornea deformation, which is in line with the observations obtained in the in vivo study [35]. The maximum cornea speeds in the loading (CV_L) and recovery (CV_R) phase, as well as the time interval between them (Δt_v), were highly correlated with the IOP. Since the effective deformation of the cornea becomes lower with increasing IOP, CV_L, CV_R, and Δt_v also decrease with IOP. Similar results were obtained earlier in human corneas in vivo [35]. However, the analysis was performed only for Type 1 stimulation (air puff parameters used in standard devices). The identification of velocity peaks and their times was not reproducible enough to provide reliable results when the air puffs with lower amplitudes were used. The reproducibility of velocity-related parameters could be improved by applying cornea and sclera segmentation technique(s), providing better performance at low signal-to-noise ratio conditions or by implementing signal phase analysis. The latter boosts sensitivity towards very low amplitudes; thus, it would require phase-unwrapping and a higher A-scan rate.

The tissue deformation is composed of loading and recovery phases. Accordingly, the return of the cornea to baseline position enables us to treat the process as a single cycle. Thus, air-puff induced cornea reaction can be plotted against the instant stimulus amplitude. The observed hysteresis is significantly affected by any offset, such as eye retraction. Retraction-corrected displacements allow us to avoid systematic errors in measurements of HA, which corresponds to the visco-elastic properties of corneal tissue. The HA decreased with the IOP and significant correlation coefficients for all types of the air puff have been found. It means that elastic properties dominated the corneal reaction when the IOP increased. However, like in the case of CRT, no particular relation between the air-puff type and HA was found for each IOP level. Finally, the correlations between the slope S and the IOP are similar to those observed for MAD.

The findings from this ex vivo study have important implications for clinical practice, particularly in the assessment of IOP and corneal biomechanics. Softening the air-puff stimulus may be more comfortable for patients; however, the results indicate that some parameters quantifying the dynamic corneal response become more vulnerable to a reduction in correlation with IOP when a gentler air-puff is applied. Consequently, the accuracy of IOP estimation based on corneal deformation data could be compromised, potentially affecting the diagnostic reliability for conditions such as glaucoma. Since the corneal response to air-puff stimulation is also used in the assessment of corneal biomechanics, optimizing stimulation parameters is crucial for the early detection of ectatic diseases such as keratoconus.

5. Conclusions

To summarize, we demonstrated a dual-sample arm SS-OCT instrument coupled with air-puff stimulation for determination of cornea displacement in an ex vivo porcine eye model. The air-puff SS-OCT instrument allowed for correction of measured corneal deformation for eye globe retraction. Comprehensive quantitative analysis of deformation dynamics was performed for different IOP levels and air-puff stimulus strengths. The cornea became stiffer when the IOP increased. Lowering the stimulation amplitude resulted in a decrease in the correlation levels between the parameters describing the induced corneal dynamics and the IOP. However, when MAD or CRT are used as the parameters of interest, they exhibit very strong correlations regardless of the air-puff excitation type studied in this work. The results suggest that softening air-puff stimulation may affect the ability of non-contact tonometers to assess the IOP and biomechanical properties of the cornea [22,34].