Functional OCT reveals anisotropic changes of retinal flicker-evoked vasodilation

Abstract

The purpose of this study is to verify the effect of anisotropic property of retinal biomechanics on vasodilation measurement. A custom-built optical coherence tomography (OCT) was used for time-lapse imaging of flicker stimulation-evoked vessel lumen changes in mouse retinas. A comparative analysis revealed significantly larger (18.21%) lumen dilation in the axial direction compared to the lateral (10.77%) direction. The axial lumen dilation predominantly resulted from the top vessel wall movement toward the vitreous direction, whereas the bottom vessel wall remained stable. This observation indicates that the traditional vasodilation measurement in the lateral direction may result in an underestimated value.

The retina, a pivotal component of the central nervous system, is responsible for converting light photons into bioelectrical signals, initiating the process of visual information processing. Neurovascular coupling, the coordinated interaction between neural activity and blood flow regulation, is crucial for maintaining the retina’s intricate functionality [1,2]. Perturbations in this coupling could potentially indicate neural dysfunctions or vascular anomalies, which are closely associated with the pathology of retinal diseases such as diabetic retinopathy [3], glaucoma [4], retinitis pigmentosa [5], hypertensive retinopathy [6], and retinopathy of prematurity [7].

Despite progress in retinal imaging technologies that enable the extraction of quantitative vascular biomarkers, these advancements primarily help in diagnosing and treating retinal diseases at more advanced stages, where morphological abnormalities become evident. Such late detection can result in irreversible vision loss, underscoring the significance of early disease identification. Early detection allows for timely intervention, potentially preventing the progression to vision loss or blindness. Functional impairments in neurovascular coupling might precede these morphological abnormalities, making the assessment of retinal hemodynamics a valuable tool for early disease detection [8]. Recent studies have shown that retinal flicker-evoked vasodilation holds potential as a biomarker for the early detection of eye diseases [9].

Retinal vasodilation, defined as the widening response of the vessel lumen to stimuli, has been explored using imaging modalities such as fundus photography [10], fluorescein angiography [11], and scanning laser ophthalmoscopy (SLO) [12]. These techniques, while invaluable, typically measure blood vessel diameter changes in a single dimension and do not account for the anisotropic biomechanical properties of the retina. The retinal structure is a layered construct, bound on the outer side to the retinal pigment epithelium (RPE) and on the inner side to the vitreous fluid, presenting a unique challenge due to the anisotropic elasticity that can lead to multidirectional changes in vasodilation.

Optical coherence tomography (OCT) provides three-dimensional (3D) structural information, allowing for the assessment of blood vessel diameters beyond the conventional single dimension, a significant advantage over other retinal imaging modalities [13,14]. Although OCT and its functional extension, OCT angiography (OCTA), enhance the visualization of retinal vessels, they are not without limitations, such as the difficulty in the precise separation of vessels from surrounding tissues and the low temporal resolution of OCTA images [15]. Additionally, adaptive optics (AO) modalities, while offering detailed en-face imaging, are limited to measuring lateral diameter changes.

This study seeks to overcome these limitations by employing Doppler OCT to measure flicker-evoked retinal vasodilation, factoring in the anisotropic biomechanical properties of the retinal tissue. Doppler OCT not only distinguishes active blood flow within vessel lumens from the surrounding static tissue but also allows for the observation of vasodilatory changes in full 360-degree detail within the cross-sectional lumen profile. This modality is advantageous for its high temporal resolution, enabling the monitoring of dynamic vasodilation changes over time [16,17].

A custom-built functional OCT was used for this study (Fig. 1). The imaging system was augmented with a green light-emitting diode (LED) to facilitate flicker light stimulation. The pulse train parameters, carefully chosen to induce vasodilation considering literature findings [11], included a duration of 100 ms, a duty ratio of 50%, a frequency of 10 Hz, and an overall duration of 5 s. The light pulse was synchronized with the OCT imaging system. The power and spectrum of the light-emitting diode were measured using a light meter to calibrate the light stimulus. The measured spectrum showed a center wavelength of 504 nm with a bandwidth of 27 nm and about 460 lx.

Fig. 1.

Schematic diagram of the functional OCT. CL, collimation lens; DM, dichroic mirror; L1, L2, L3, lens; PC, polarization controller; RS, retinal stimulator; SLD, super luminescent diode; SM, scanning mirror. Retinal stimulator consists of a narrowband LED light (λ = 505 nm).

C57BL/6J strain mice were used for this study. Prior to imaging, mice were anesthetized intraperitoneally with a ketamine and xylazine mixture. Phenylephrine hydrochloride 2.5% and tropicamide 1% were administered for full pupil dilation, and a cover glass with GenTeal eye gel was applied to the cornea to prevent drying and reduce spherical aberrations. Animal care and examination protocols were approved by the University of Illinois Chicago’s Animal Care Committee, adhering to the Association for Research in Vision and Ophthalmology statement for animal use in ophthalmic and vision research.

The experiments were conducted in a dark room with no ambient light. The mice were dark-adapted for 1–2 h prior to the experiment. After 10–15 min for full pupil dilation, the mice were moved to the animal holder. Circular scanned OCT images were acquired around the optic nerve head area for Doppler OCT imaging. A total of 1050 frames were captured at a recording speed of 35 FPS during each imaging session. The session was organized into three distinct phases: an initial 3 s pre-stimulation phase designed to establish a reliable baseline, comprising 105 images; a second 5 s light stimulation phase, resulting in 175 images for in-depth analysis during stimulation; a third 22 s post-stimulation phase, generating 770 images to capture both immediate and gradual hemodynamic responses. Each mouse in the study underwent a single imaging session. The OCT imaging was performed continuously during the entire recording session. All data was saved to a computer hard drive for post-processing. The custom-designed OCT system was controlled using lab-built software written in LabVIEW. Structural OCT and Doppler OCT images were reconstructed. Approximately 8 to 12 blood vessels per Doppler image in the mouse retina were manually selected for quantitative analysis, prioritizing those near the optic nerve head for their visibility, ease of differentiation, and higher density. This deliberate choice enhances the statistical robustness of our findings, supported by circular scanning around the optic nerve head for a comprehensive capture of the vessel network, reinforcing the overall robustness of our analysis. Both axial and lateral profiles of blood vessels were measured from OCT and Doppler OCT images recorded at different time points. A total of 80 vessels, comprising 41 arteries and 39 veins, from seven mice were analyzed in this study. All statistical analyses were conducted using Origin Lab and a custom-developed Matlab algorithm. To compare changes in lateral and axial diameters, a two-sample t-test was performed, assuming unequal variance, with significance defined as p < 0.001.

The en-face OCT image reveals the presence of major blood vessels surrounding the optic disk (Fig. 2(A)). OCT B-scan, on the other hand, provides detailed information about the different layers in the retina, allowing for the visualization of individual blood vessels (Fig. 2(B)). In Fig. 2(C), a magnified OCT B-scan from the white box in Fig. 2(B) shows a clear cross-sectional view of a blood vessel. Following stimulation, the vessel wall expansion is evident in the axial profile of the OCT B-scan (Fig. 2(D1)). Specifically, the top vessel wall and signal intensity peak move toward the vitreous, while the bottom vessel wall remains unchanged. The lateral profile shows a relatively smaller change compared to the axial profile (Fig. 2(D2)).

Fig. 2.

(A) Representative en-face OCT image. (B) Circular OCT B-scan image corresponding to the orange circle in (A). (C) Magnified OCT B-scan image corresponding to the white box in (B). (D) Flicker-evoked vessel profile changes measured from time-lapse OCT B-scan images of the (D1) axial and (D2) lateral direction in (C).

The Doppler OCT image provides clear differentiation of the vessel lumen from the surrounding tissue, utilizing distinct colors to distinguish between arteries and veins (Fig. 3(A)). In Fig. 3(B), the Doppler OCT images represent the blood flow within the vessel lumen, revealing a noticeable increase in blood flow after the stimulation. Furthermore, the images show uneven blood flow distribution within the vessel lumen. The Doppler OCT images also confirm vasodilation as a response to the stimulus by the flow area expansion which reflects the vessel lumen. Notably, this vasodilation is observed to be non-uniform across the vessel’s circumference. The expansion of the blood flow area is found to be more significant in the axial direction than in the lateral direction. Specifically, the axial profile displays a substantial increase in the distance between the top and bottom boundaries of the vessel lumen, indicating a significant expansion of the blood vessel (6.84 μm) after the stimulation. In contrast, the lateral profile shows a relatively smaller increase (3.16 μm) in vessel diameter.

Fig. 3.

(A) Doppler OCT corresponding to Fig. 2(B). (B) Time-lapse Doppler OCT of the vessel lumen shown in Fig. 2(C), representing changes in response to flicker stimulation. (C) Time-lapse flow profile changes in both the (C1) axial and (C2) lateral directions of the vessel lumen in (B).

For a better understanding of the anisotropic property of the retinal flicker-evoked vasodilation, we quantitatively evaluated the vasodilation magnitudes in all directions, including the relocation of the vascular center (Fig. 4). To analyze variations of vasodilation in response based on the type of blood vessel, arteries and veins are assessed separately. In the case of arteries, as shown in Fig. 4(A1), all vessels are observed to vasodilate after light stimulation, although there are differences in the degree of change. The vasodilation is more pronounced in the axial direction than that in the lateral direction. Also, in the axial direction, the top vessel wall expands more than the bottom vessel wall. The vasodilation of veins shows a similar trend to that of arteries (Fig. 4(B1)). However, when comparing the degree of dilation, it is found that the degree of venous dilation is not as great as that in arteries. This directionality of vasodilation is further confirmed by observing changes in the center of the vessel lumen. As shown in Fig. 4(A2), the center of the vessel lumen is moving in the axial direction, especially toward the vitreous. The veins show a similar trend but to a lesser extent than the arteries (Fig. 4(B2)).

Fig. 4.

Flicker-evoked vasodilation in the (A1) artery and (B1) vein, along with alterations in their (A2) and (B2) central positions, computed from time-lapse Doppler OCT of the vessel lumen in Fig. 3(A).

Figure 5 illustrates the quantitative directionality of vasodilation, delineating changes in axial and lateral diameters over time alongside the total area and dynamic volumetric blood flow changes. Each graph is represented with the mean and standard deviation (σ). Lateral changes align with the traditional light-induced vasodilation. Vasodilation onset occurs when responses exceed 3σ above the baseline for over 1 s, with peak times identified at the graph’s first plateau lasting similarly. Both arteries and veins show latency-based diameter increases post-stimulation, with a pronounced rise in axial over lateral diameters. Arterial axial diameters surge post-stimulation 1.26 s, peaking at 6.72% by 2.51 s, and eventually reaching 18.21% (Figs. 5(A1) and 5(B1)). Veins begin their increase at 1.4 s, peaking at 4.41% by 3.03 s, and climb to 14.25% (Figs. 5(A1) and 5(B1)). Lateral diameters for arteries rise sharply post-stimulation 1.26 s, peaking at 3.98% by 2.45 s, and ascend to 10.77% (Figs. 5(A2) and 5(B2)). For veins, the increase starts at 1.46 s, peaking at 2.49% by 3.06 s, and culminates at 8.87% (Figs. 5(A2) and 5(B2)). Total area increases mimic axial and lateral summations, with arterial areas spiking at 13.05% by 2.57 s and reaching 35.79%, and venous areas peaking at 8.6% by 3.06 s, eventually hitting 27.6% (Figs. 5(A3) and 5(B3)). Both vessel types exhibit similar onset times across measurements, with arteries peaking sooner. Volumetric blood flow in arteries begins to rise post-stimulation 1.26 s, peaking at 26.81% after a 9.57% increase by 2.54 s, while veins show a delayed start at 1.46 s, peaking at 4.98% by 3 s and reaching 17.6% (Figs. 5(A4) and 5(B4)).

Fig. 5.

(A) Illustration of stimulation-evoked blood vessel changes, encompassing (A1) lateral diameter, (A2) axial diameter, (A3) total vessel area, and (A4) dynamic volumetric blood flow. Green windows indicate retinal stimulation period. (B) Enlarged illustration of the response graphs, corresponding to −1 to 5 s. The red and blue lines represent arteries and veins, respectively. Red and blue arrows denote changes in arteries and veins, respectively, with the first arrow indicating the onset time, while the second arrow signifies the time to peak response.

In summary, functional OCT imaging reveals flicker-evoked anisotropic vasodilation in the mouse retina, shedding light on the mechanisms governing retinal blood flow regulation. The observed increase in blood flow is associated with the release of vasoactive substances, leading to vessel wall expansion through heightened nitric oxide production to meet increased metabolic demand during visual stimulation [18]. The latency before vasodilation onset may be attributed to the time required for neurovascular coupling to sense and signal increased metabolic demand [11], aligning with previous studies [19,20]. Arteries demonstrate a faster onset, quicker peak response, and greater vasodilation compared to veins, as reported in the literature [11]. Distinct arterial and venous characters are likely linked to their anatomical differences and roles in blood flow regulation [21]. Veins, with thinner walls and less active involvement in blood flow regulation, may exhibit less pronounced vasodilation [22].

The observation of notably greater expansion of the vessel lumen in the axial direction, as opposed to the lateral direction, following stimulation can be attributed to several interconnected factors that influence the biomechanics of the retinal vasculature. First, the composition of the surrounding tissue. The vitreous humor, a gel-like substance within the eye, plays a pivotal role in influencing this differential expansion. It exhibits greater flexibility and compliance, allowing it to accommodate changes in the shape and size of the retinal blood vessels [23]. When vessels expand in the axial direction, they encounter minimal resistance from the vitreous, enabling them to dilate more significantly. In contrast, the surrounding tissue in the lateral direction may be composed of structures that are comparatively less elastic and more rigid. This relative rigidity can impede lateral vessel wall expansion, resulting in a relatively smaller change in vessel diameter. Second, mechanical characteristics of the vessel wall. The mechanical properties of the vessel wall are not uniform along all axes [24]. The axial direction may inherently exhibit greater compliance and expandability than the lateral direction. This variation in mechanical characteristics can be attributed to the organization of smooth muscle cells and connective tissues within the vessel wall, which may differ between axes [25].

Despite a consistent trend of larger axial lumen dilation compared to the lateral direction in retinal flicker-evoked vasodilation, certain cases, such as vessel 8 in Fig. 4, deviate, possibly due to inherent differences in vascular architecture, regional variations, and specific physiological attributes. While acknowledging these deviations, the overall pattern provides a robust framework for understanding retinal flicker-evoked vasodilation dynamics, prompting the need for further investigation, including biomechanical modeling or additional experiments, to confirm contributing factors. Additionally, the prolonged duration of vasodilation in both arteries and veins, exceeding typical response times in standard hemodynamic functions, may be influenced by specific stimulus types and intensities, highlighting the importance of exploring factors behind the extended vasodilation and its implications for vascular dynamics [26–28].

Our study provides a foundation for understanding anisotropic vasodilation around the optic nerve head, with potential implications for certain regions of the retina. While extending these findings to capillaries may pose challenges due to their anatomical location, the exploration of Doppler OCT in human trials holds promise, emphasizing the need for imaging stabilization techniques to overcome potential complication of eye movements. Further research and interdisciplinary collaboration will be crucial to fully unravel the perspectives and limitations of our results in the context of broader retinal biomechanics.

In conclusion, functional OCT reveals anisotropic vasodilation within the retinal vessel lumens. Retinal flicker-evoked lumen expansion is significantly larger in the axial direction compared to that in the lateral direction, which can be reasonably explained by the biomechanical anisotropy of the retinal tissue. Traditional vasodilation measurement that considers only the lateral direction may result in an underestimated value.

Data availability.

Data is available upon reasonable request.