Retinal neurovascular signatures in response to 6-month yoga exercise in patients with Parkinson’s Disease

Abstract

Purpose:

To characterize retinal neurovascular signatures in response to a 24-week yoga intervention in patients with Parkinson’s disease (PD).

Methods:

Seventeen patients with PD (mean age: 71.4 ± 6.7 years; 8 females) were enrolled and participated in three 1-hour yoga sessions per week for 24 consecutive weeks. Retinal neurovascular parameters were assessed at baseline and at 6-month follow-up. Retinal blood flow (RBF) was measured using a Retinal Function Imager. Retinal vessel density (RVD), retinal vessel length density (RVLD), and retinal tissue volume (RTV) were assessed using optical coherence tomography angiography (OCTA). Retinal capillary function (RCF) and retinal tissue perfusion (RTP) were calculated based on these variables.

Results:

RBF significantly increased following 6 months of yoga training (P = 0.049). Changes in RVD were significantly correlated with disease duration (r = −0.67; P = 0.007) and Hoehn and Yahr (H&Y) stage (r = −0.77; P < 0.001). Patients with disease duration < 5 years demonstrated greater improvements in RVD compared with those whose duration was ≥ 5 years (P = 0.004). Similarly, patients at H&Y stage 1 exhibited greater RVD improvements compared to those at H&Y stage ≥ 2 (P = 0.002).

Conclusion:

This study is the first to investigate retinal neurovascular responses to yoga training in patients with PD. While RBF marginally increased, RVD exhibited disease-severity and duration-dependent variations, suggesting RBF and RVD are potential biomarkers for monitoring disease progression and the efficacy of aerobic interventions such as yoga in PD.

Introduction

Parkinson's disease (PD) is a progressive neurodegenerative disorder primarily characterized by the degeneration of dopaminergic neurons in the substantia nigra pars compacta.[1] Diagnosis of PD is primarily clinical, and by the time patients are diagnosed, a substantial loss of dopaminergic neurons has already occurred.[1] No widely recognized biomarkers currently exist for the early diagnosis of PD. Current therapeutic strategies focus on symptomatic relief, primarily through dopaminergic medications such as levodopa and dopamine agonists.[1]

Exercise has a substantial positive impact on PD, improving mobility, balance, strength, and endurance while reducing bradykinesia and fall risk.[2] Additionally, exercise positively influences cognitive function, mood, and sleep quality.[2,3] Mechanistically, physical activity promotes neuroprotection, reduces inflammation, and increases cerebral blood flow, which may contribute to slowing disease progression.[3] A structured regimen that includes aerobic, strength, flexibility, and balance training is recommended to maximize benefits.[2,3] Yoga is a form of exercise that integrates physical postures, controlled breathing, and focused attention.[4] Regular practice can improve muscular strength, balance, mobility, and lower-body flexibility, making yoga a meaningful form of physical training.[5]

The eye serves as a valuable portal for studying PD. Multimodal imaging modalities have also been used to characterize retinal neurovascular signatures in PD, including retinal blood flow (i.e., microcirculatory hemodynamics), vessel density (i.e., microvascular structure), and tissue volume (i.e., microstructure) in patients with PD.[6] A prior study demonstrated that these elements in the neurovascular system may independently change relative to one another during pathological conditions.[6] The present study continued to characterize retinal neurovascular signatures in response to a 6-month yoga intervention in patients with PD.

Methods

Subjects

The present study was approved by the institutional review board for human research at the University of Miami. Participants were fully informed of their potential risks and benefits. Afterwards, informed consent was signed by each subject prior to their participation in the study. All subjects were treated in accordance with the Tenets of the Declaration of Helsinki.

Twenty patients with clinically diagnosed Parkinson’s disease (8 female; mean age 72.1 ± 7.1 years; disease duration 6.6 ± 4.8 years) were recruited, of whom 17 underwent successful retinal imaging. Inclusion criteria required age ≥55, MoCA >18, Hoehn and Yahr stage 1–3, and the ability to ambulate ≥50 feet and rise from the floor with minimal assistance. Exclusion criteria included significant refractive error (>±6.0 diopters), known ocular disease, stroke, coagulopathy, hypertension or diabetes.

All participants underwent a comprehensive ophthalmologic exam, including visual acuity, intraocular pressure, anterior segment, and posterior segment examination, to confirm eligibility. One eye per subject was imaged (OD preferred; OS if ineligible). Imaging was performed while subjects remained on stable dopaminergic therapy, with follow-up imaging conducted 2–7 days after the final yoga session. All subjects were receiving only Carbidopa/Levodopa treatment for an average of 67 ± 56 months (mean ± SD).

RVD, RVLD, and RTV Measurement

Retinal capillary vessel density (RVD), retinal capillary vessel length density (RVLD), and retinal tissue volume (RTV) were assessed using optical coherence tomography angiography (OCTA, AngioVue, Optovue Inc., Fremont, CA). Use of the device has been previously described, operating at a speed of 70,000 A-scans per second with an axial resolution of 5 μm.[6] Scans were centered on the fovea and only scans reaching Optovue’s internal quality metric of ≥7/10 were used for analysis.

The calculation of VD and VLD has been described in detail previously.[6] OCTA images were processed with illumination correction, noise reduction, and non-vessel removal, then binarized. Vessels ≥25 μm were excluded. An annulus 0.6–2.5 mm diameter centered on the fovea, was analyzed, removing vessels inside and outside this region. ImageJ (National Institutes of Health, Bethesda, MD, USA; version 1.54) was then utilized to calculate VD as the percentage of vessels covering the total area of the annulus. VLD was calculated after the original binary image was skeletonized and defined as the total vessel length in millimeters per square millimeter of the annulus (mm⁻¹). Retinal tissue volume was assessed within a 2.5 mm diameter circle centered on the fovea and spanned the retinal nerve fiber layer (RNFL), to the outer plexiform layer (OPL).

RBF Measurement

Retinal blood flow (RBF) was assessed using the Retinal Function Imager (RFI, Optical Imaging Ltd., Rehovot, Israel), and its use is well-described.[7] Briefly, it employs a fundus camera with stroboscopic illumination and high-speed imaging to track red blood cells and quantify blood flow velocity noninvasively.[7] Image acquisition is synchronized with the cardiac cycle, enabling reproducible measurements.[7] Combining vessel diameter and velocity allows calculation of macular blood flow.[7]

Participants rested for 15 minutes in a semi-dark room, with dilation using 1% tropicamide. Imaging used a 35° field of view (7.3×7.3 mm²), capturing ≥4 images per session across 3–5 sessions. Blood flow velocities were calculated within a 2.5 mm foveal-centered region using proprietary (Browse, ver. 2.2.0.236) and Matlab-based software, with total macular flow defined as the average of summed arteriolar and venular flow rates, which were approximately equal.[7]

Yoga Intervention

The two tested yoga styles employed predominantly standing postures found in Hatha yoga, such as Warrior poses 1 and 2 (Virabhadrasana 1 and 2), mountain (Tadasana), triangle pose (trikonasana), tree pose (Vrikshasana), and the airplane pose (Dehasana).[8] Sessions were preceded and succeeded by passive movements and postures of rest.[8] The YogaCue program developed by two of the authors (JFS and KJM) incorporated two modifications: shortened recoveries between poses and the use of visual and auditory cues to indicate pose sequences. Both programs included three 1-hour sessions per week. These poses were selected to highlight lower-extremity strength, flexibility, postural control, and balance during exercise periods. All subjects in the current study were able to perform these poses and no adverse events occurred during training.

Calculation of RCF and RTP

Retinal capillary function (RCF) and retinal tissue perfusion (RTP) are ratios which serve as representations of the interactions between RBF, RVLD, and RTV. RCF was calculated by dividing RBF by VLD and represented the amount of blood flowing through a single millimeter of tissue. RTP was calculated by dividing RBF by RTV and represented the amount of blood in a given cubic millimeter of tissue. The purpose of each calculation was to normalize changes in RBF by another retinal neurovascular variable.[6]

Statistical Analysis

Descriptive statistics and data analysis were conducted using the Statistical Package for the Social Sciences (SPSS for Windows 27.0; SPSS Inc., Chicago, Illinois, USA). A Student's paired t-test assessed changes to retinovascular variables between baseline and follow-up. Pearson correlation assessed for linear relationships between changes in the retinovascular variables following yoga exercise and PD duration, Hoehn and Hahr stage, and Unified Parkinson’s Disease Scale part III – Mobility (UPDRS III) scores. Patients were stratified into two subgroups based on disease duration and Hoehn and Yahr stage. A Shapiro-Wilk test assessed the normality of data in each subgroup, and unequal variance was assumed among all cohorts due to the small and unequal sample sizes of the subgroups. Based on these results, between-group comparisons were either made with a Welch’s t-test or with a Mann-Whitney U test. The significance level was set at 0.05.

Results

Retinal Neurovascular Signatures Among the Whole Cohort

RBF increased by 0.36 ± 0.70 nL/s (mean ± SD; P = 0.049) among the entire cohort over the six-month course of the present study. RVD increased by 0.23 ± 1.66% (P = 0.67) and RVLD increased by 0.54 ± 1.32 mm⁻¹ (P= 0.13). RTV decreased by 0.02 mm³ (P = 0.06). RCF increased by 0.02 ± 0.04 nL/s/mm (P= 0.14) and RTP increased by 0.32 ± 0.61 nL/s/mm³ (P= 0.06). PD motor symptom severity, assessed by the UPDRS part III, increased among the overall subject pool from 27 ± 12 to 34 ± 15 (P = 0.003) over the six months of the study. While examining perparticipant changes to retinal neurovascular signatures, it was found that changes in RVD over the study period (ΔRVD) were linearly correlated to PD symptom severity and disease duration (Figure 1). ΔRVD negatively correlated with PD duration (r = −0.67; P = 0.007). ΔRVD additionally correlated with Hoehn and Yahr (H&Y) stage (r = −0.77; P < 0.001). UPDRS III, was related to ΔRVD at baseline (r = −0.56; P = 0.03) as well as with the relative changes in patients’ UPDRS III (ΔUPDRS III) from baseline to follow-up (r = −0.55; P = 0.04). ΔRBF, ΔVLD, ΔRTV, ΔRCF, and ΔRTP were not related to PD duration or severity (P > 0.05).

Figure 1:

The relationships between changes in retinal vessel density (ΔRVD) and PD duration, Hoehn and Yahr (H&Y) Scale measurements, the Unified Parkinson’s Disease Score part III – Mobility (UPDRS III) at baseline, and changes to the UPDRS III following yoga exercise (ΔUPDRS). (A) ΔRVD was correlated with PD duration (r = −0.67; P = 0.007). (B) ΔRVD was correlated with H&Y stage at baseline (P < 0.001). (C) ΔRVD was related to UPDRS III at baseline (P = 0.03). (D) ΔRVD correlated with ΔUPDRS III (P = 0.04).

Subgroup Analysis

Patients were stratified into subgroups based on disease duration and baseline H&Y stage (Table 1). Among patients with a PD duration of < 5 years, RVD increased 1.7 ± 0.9% (mean ± SD; P = 0.02), RVLD increased by 1.3 ± 0.6 mm⁻¹ (P = 0.01), and RTV decreased by 0.04 ± 0.03 mm³ (P = 0.04). In contrast, patients with PD for ≥ 5 years did not exhibit changes in RVD, RVLD, or RTV; however; these patients experienced significant within-group improvements to RBF and RTP which were not found in patients with a disease duration less than 5 years. RBF in patients with PD for ≥ 5 years increased 0.2 ± 0.3 nL/s (P = 0.01) and RTP increased by 0.2 ± 0.2 nL/s/mm³ (P = 0.04).

Table 1:

Subgroup statistics

Parameter	Unit	Subgroup 1	Subgroup 1	Subgroup 1 P-value	Subgroup 2	Subgroup 2	Subgroup 2 P-value	Between Group P- Value for Δ
Disease Duration		< 5 years	< 5 years	< 5 years	≥ 5 Years	≥ 5 Years	≥ 5 Years	< 5 vs ≥ 5 Years
RBF	nL/s	2.4 ± 0.7	2.9 ± 1.6	0.27	2.5 ± 0.5	2.7 ± 0.6	0.01	0.81
RVD	%	31.6 ± 1.7	33.3 ± 1.15	0.02	32.2 ± 1.1	31.7 ± 1.3	0.31	0.004
RVLD	mm−1	19.1 ± 1.1	20.41 ± 1.4	0.01	19.61 ± 1.0	19.8 ± 1.3	0.73	0.05
RTV	mm3	1.24 ± 0.05	1.20 ± 0.05	0.04	1.22 ± 0.09	1.22 ± 0.09	0.77	0.03
RTP	nL/s/mm3	1.7 ± 0.4	2.3 ± 1.4	0.26	2.0 ± 0.4	2.2 ± 0.5	0.04	0.64
RCF	nL/s/mm	0.11 ± 0.03	0.14 ± 0.09	0.42	0.13 ± 0.02	0.14 ± 0.03	0.11	0.44
H&Y Scale		H&Y 1	H&Y 1	H&Y 1	H&Y ≥ 2	H&Y ≥ 2	H&Y ≥ 2	1 vs ≥ 2
RBF	nL/s	2.2 ± 0.5	2.9 ± 1.6	0.22	2.6 ± 0.5	2.8 ± 0.6	0.02	0.53
RVD	%	31.3 ± 1.0	33.1 ± 1.0	0.005	32.5 ± 1.3	31.7 ± 1.4	0.06	0.002
RVLD	mm−1	19.0 ± 0.9	20.1 ± 1.2	0.04	19.8 ± 1.1	19.9 ± 1.4	0.74	0.15
RTV	mm3	1.22 ± 0.05	1.19 ± 0.05	0.13	1.23 ± 0.09	1.22 ± 0.10	0.42	0.22
RTP	nL/s/mm3	1.8 ± 0.4	2.4 ± 1.3	0.19	2.0 ± 0.4	2.2 ± 0.5	0.08	0.86
RCF	nL/s/mm	0.12 ± 0.03	0.14 ± 0.08	0.31	0.12 ± 0.03	0.13 ± 0.03	0.21	0.33

Δ: Change from baseline to follow-up; RBF: retinal blood flow; RVD: retinal vessel density; RVLD: retinal vessel length density; RTV: retinal tissue volume. H&Y: Hoehn and Yahr scale. RCF: retinal capillary function; RTP: retinal tissue perfusion. Values are represented as mean ± standard deviation.

The between-group difference in ΔRVD among patients with PD for < 5 years vs. ≥ 5 years was significant (Table 1; P = 0.004). RVLD additionally significantly increased by 1.3 ± 0.6 mm⁻¹ after 24 weeks of yoga exercise (P = 0.01); however, patients with PD for 5 years or more experienced no significant within-group changes in RVLD (P = 0.73). The between-group changes in RTV were significant (P = 0.03).

Patients with a baseline H&Y stage of 1 demonstrated a significant within-group increase in RVD and RVLD (Table 1). RVD increased by 1.8 ± 0.9% (mean ± SD; P = 0.005) after yoga exercise and RVLD increased by 1.8 ± 0.9 mm⁻¹ (P = 0.04). Conversely, patients with a baseline H&Y stage of 2 or higher exhibited a decline in RVD which trended toward, but did not reach, significance (P = 0.06). Patients with a baseline H&Y stage of two or three also experienced a within-group increase in RBF of 0.2 ± 0.3 nL/s (P = 0.02). The between-subgroup difference in ΔRVD between patients with a baseline H&Y of 1 vs ≥ 2 was significant (P = 0.002).

Discussion

This is the first study to explore the changes of the retinal neurovascular signatures in response to yoga exercise and to elucidate the relationship of these signatures with clinical PD measurements. It builds on prior work establishing a baseline retinal neurovascular profile in PD.[6] The key findings that RVD changes following yoga exercise correlated with disease duration, baseline H&Y stage, and UPDRS III score signify that alterations to the retinal neurovascular signature may serve as noninvasive biomarkers of disease progression. The association between ΔRVD and ΔUPDRS III further indicates potential prognostic value for short-term motor outcomes and supports ΔRVD as a marker of exercise treatment efficacy. Notably, greater improvements in patients with shorter disease duration (<5 years) or lower H&Y stage highlight the importance of early intervention.

Prior studies demonstrated reduced retinal blood flow (RBF) and nonsignificant reductions in RVD and RVLD in PD.[6] The present study sought to understand if yoga exercise would alter this signature. Exercise has been shown to improve retinal vessel density, reflecting enhanced vascular health and neurovascular integrity.[9,10] OCTA-based studies indicate that physical activity increases capillary and vessel density, improving perfusion and oxygen delivery.[9,10] The mechanisms underlying these effects may involve enhanced endothelial function, increased expression of vascular growth factors such as vascular endothelial growth factor (VEGF) from improved cardiovascular fitness.[9]

RVD and RVLD among the whole sample did not align with the aforementioned trend of increased vessel density. This likely reflects heterogeneity in response by disease duration and motor symptom severity. In contrast, patients with shorter disease duration and milder motor impairment demonstrated the expected increase in RVD. Considered alongside the significant association between ΔRVD and PD duration, these findings support earlier initiation of exercise therapy in PD.

Emerging evidence suggests that metabolically demanding exercise promotes neuroprotection and delays PD progression by slowing motor decline in early-stage PD.[11] Mechanistically, such exercise enhances neuroplasticity and resilience through adaptive stress responses.[12] The observed relationship between ΔRVD and ΔUPDRS III supports these mechanisms and underscores ΔRVD as a potential biomarker of microvascular response to exercise.

Improved RBF across the cohort is meaningful, as increased perfusion enhances oxygen and nutrient delivery which may counteract neurodegeneration.[13] RTV did not significantly change overall, though greater reductions were seen in early-stage patients. This may reflect a “floor effect,” where advanced PD patients with preexisting retinal thinning exhibit slower degeneration.[14]

The current study had a few limitations. First, the study duration was relatively short. Although the intervention lasted 6-month, this duration may not be sufficient to observe all of the retinal neurovascular changes that take place in response to yoga. Several metrics, such as the difference in RVLD among patients with a greater or lesser PD duration, trended towards significance. More extended follow-up periods help clarify the potential long-term benefits of yoga on retinal health. Second, there was no control group. The absence of a control group limits the ability to draw definitive conclusions about the effectiveness of yoga as an intervention. Comparing results with a control group would strengthen the evidence for causality.

Figure 2:

Representative yoga poses included in the intervention. The yoga sequences were as follows: (A) Warrior (Virabhadrasana) I, (B) Warrior (Virabhadrasana) II, (C) Mountain (Tadasana), (D) Triangle (trikonasana), (E) Tree (Vrikshasana), and (F) Airplane (Dehasana) poses. The yoga protocol was created to emphasize lower-extremity strength, flexibility, postural control, and balance.

Highlights.

• Yoga increased retinal blood flow in patients with Parkinson’s disease.

• Retinal vessel density changes correlated with PD duration and severity.

• Early-stage PD patients showed greater microvascular improvements after yoga.

• RVD changes may serve as a signature of PD progression and exercise response.

Data Availability

Data may be made available upon request.