Correlating stroke risk with non-invasive cerebrovascular perfusion dynamics using a portable speckle contrast optical spectroscopy laser device

Research Article

Vol. 15, No. 10 /1 Oct 2024 /

Biomedical Optics Express

6083

Correlating stroke risk with non-invasive

cerebrovascular perfusion dynamics using a

portable speckle contrast optical spectroscopy

laser device

UANG

†

IMON

AHLER

1,8,9,

†

IDIN

BEDI

2,3,4

ULIAN

ICHAEL

YSZKA

UNG

2,6

ATRICK

D. L

YDEN

ONATHAN

USSIN

2,3

HARLES

2,3,10,

‡

AND

HANGHUEI

ANG

‡

Department of Electrical Engineering, California Institute of Technology, Pasadena, CA 91125, USA

USC Neurorestoration Center, Department of Neurological Surgery, Keck School of Medicine, University

of Southern California, Los Angeles, CA 90033, USA

Rancho Research Institute, Rancho Los Amigos National Rehabilitation Center, Downey, CA 90242, USA

Department of Urology, University of Toledo College of Medicine and Life Sciences, Toledo, OH 43614,

USA

Division of Humanities and Social Sciences, California Institute of Technology, Pasadena, CA 91125, USA

Department of Neurosurgery, National Neuroscience Institute, Singapore 308433, Singapore

Department of Physiology and Neuroscience, Zilkha Neurogenetic Institute, and Department of Neurology,

Keck School of Medicine, University of Southern California, Los Angeles, CA 90033, USA

mahler@caltech.edu

sim.mahler@gmail.com

cliu@usc.edu

†

These authors contributed equally to this work.

‡

These authors jointly supervised this work.

Abstract:

Stroke poses a significant global health threat, with millions affected annually, leading

to substantial morbidity and mortality. Current stroke risk assessment for the general population

relies on markers such as demographics, blood tests, and comorbidities. A minimally invasive,

clinically scalable, and cost-effective way to directly measure cerebral blood flow presents

an opportunity. This opportunity has the potential to positively impact effective stroke risk

assessment prevention and intervention. Physiological changes in the cerebrovascular system,

particularly in response to hypercapnia and hypoxia during voluntary breath-holding can offer

insights into stroke risk assessment. However, existing methods for measuring cerebral perfusion

reserves, such as blood flow and blood volume changes, are limited by either invasiveness or

impracticality. Herein we propose a non-invasive transcranial approach using speckle contrast

optical spectroscopy (SCOS) to non-invasively monitor regional changes in brain blood flow and

volume during breath-holding. Our study, conducted on 50 individuals classified into two groups

(low-risk and higher-risk for stroke), shows significant differences in blood dynamic changes

during breath-holding between the two groups, providing physiological insights for stroke risk

assessment using a non-invasive quantification paradigm. Given its cost-effectiveness, scalability,

portability, and simplicity, this laser-centric tool has significant potential for early diagnosis and

treatment of stroke in the general population.

1. Introduction

Stroke is a global health concern, with a distressingly high incidence and substantial morbidity

and mortality [1]. Annually, more than ten million people worldwide are impacted by strokes,

#534796

https://doi.org/10.1364/BOE.534796

Received 8 Jul 2024; revised 26 Aug 2024; accepted 7 Sep 2024; published 30 Sep 2024

Research Article

Vol. 15, No. 10 /1 Oct 2024 /

Biomedical Optics Express

6084

imposing a heavy toll on affected individuals and their families, with significant health, financial,

and quality of life burdens [2]. From the age of 25, the lifetime risk of stroke is around 25%

[3,4]. In the US, stroke affects nearly 800,000 individuals each year, contributing to significant

healthcare challenges [5]. Early detection of stroke risk is pivotal in reducing its incidence and

severity and improving outcomes. A variety of interventions, such as lifestyle modifications

(including diet, exercise, and smoking cessation) and medical management strategies (such as

blood pressure control), exist to mitigate stroke risk [5–8].

Traditionally, stroke risk estimation relies on indirect markers such as demographics and

comorbidities. As such, the most widely used questionnaires for stroke risk estimation capture

factors such as diabetes and hypertension, along with lifestyle-related factors such as smoking

and low physical activity [2,9–14]. These risk profiles are based on population data and are

useful for general assessments. However, they are not definitive for determining the need for

invasive or non-invasive evaluations, nor for guiding surgical or pharmacological interventions in

individual patients to prevent future strokes. The community has pushed forward exploration

in using non-invasive models to detect and characterize after the onset of stroke, but not for

the quantization of stroke risk [15,16]. This gap highlights the pressing need for objective and

cost-effective measures of cerebral perfusion reserve. This need is particularly evident when

compared to the management of myocardial ischemia through stress testing. Stress tests serve

as a pivotal, objective method to evaluate cardiac health, functional capacity, and myocardial

ischemia risk. An equivalent direct and objective measure for assessing neurovascular health that

is low-cost enough to deploy broadly in the community—akin to but potentially more available

than cardiac stress testing—is conspicuously absent. We believe that this study introduces a

simple, cost-effective solution that does not require ionizing radiation or radioactive isotopes that

has the potential to lead to better development of personalized strategies for reducing stroke risk.

Efforts to characterize cerebral perfusion capacity have been underway for the past four decades.

Neuroimaging modalities such as PET, SPECT, xenon-enhanced CT, and perfusion CT can

be coupled with acetazolamide administration to augment brain perfusion and reveal areas of

diminished reserve [17–19]. However, all existing imaging modalities share several common

limitations: they cannot be performed under exercise stimuli such as a treadmill stress test; they

all require expensive equipment; and their lack of scalability and logistic limitations do not

permit implementation at bedside, in the clinic, or widespread use for screening in the community

settings. Each method also has its specific limitations. For instance, the acetazolamide challenge

has various side effects and contraindications in the presence of certain comorbidities, while the

reference measurements used in CT perfusion assessment are derived from a single unaffected

major cerebral vessel in each individual, which in many cases may lead to the underestimation

of cerebral blood flow (CBF) in affected regions [18]. The Framingham Stroke Risk Profile

has recently been augmented with multimodal magnetic resonance imaging (MRI) to quantify

the risk of acute stroke in patients with symptomatic ischemia [20,21]. This approach requires

access to contrast-enhanced MRI, rendering it impractical for routine stroke risk screening.

Further, with the advent of ischemic symptoms, primary prevention likely becomes moot, as

the presence of symptoms suggests the progression of vascular disease to a clinically significant

stage. There have been efforts using transcranial Doppler (TCD) ultrasonography to evaluate

brain blood flow [22–24] with technician-dependent results. In addition, TCD only targets larger

arteries and can only be used on limited areas of the skull, most commonly the temporal acoustic

window [25]. A cost-effective and scalable stroke risk assessment system does not exist currently.

This complicates the effective long-term prevention of stroke. Having established through a

questionnaire that a patient is at high risk for stroke, a physician has no good way to know whether

a patient’s risk is increasing or stable, and whether or when a patient needs intervention.

Physiologically, a stroke occurs when the brain’s vascular system is compromised, either

through the blocking of blood floow (ischemic stroke) or a rupture of a blood vessel (hemorrhagic

Research Article

Vol. 15, No. 10 /1 Oct 2024 /

Biomedical Optics Express

6085

stroke). Both events lead to reduced blood flow and consequently oxygen supply to brain tissues.

Risk factors of stroke include age, cardiovascular diseases, high blood pressure, high blood

sugar, smoking, and elevated cholesterol [10]. These same factors also impact how the brain

reacts to oxygen deprivation or carbon dioxide buildup during events like breath-holding. During

breath-holding, both cerebral blood volume and cerebral blood flow increase in response to

higher carbon dioxide and lower oxygen levels. To mitigate the increased shear pressure of the

blood caused by the rising blood flow, the cerebral autoregulation system prompts blood vessel

dilation. Such cerebral autoregulation is impaired in diseased atherosclerotic vessels, including

in response to other hemodynamic perturbations such as hypotension and upstream stenosis (e.g.,

carotid stenosis).

The skull blocks most means for direct, non-invasive measurements of the brain’s blood

dynamics. Fortunately, infrared light can transmit relatively well through the skull and brain

[26–26]. By transmitting infrared light through one location on the skull and collecting its

transmission from another location, it is possible to track the relative blood volume by measuring

the light attenuation rate [25,29,30]. If the light used is coherent, such as a laser, mutual

interference of light following different trajectories causes speckle patterns, which depend on

the coherence of the laser source [31–34]. By observing how fast the transmitted laser speckle

field fluctuates, it is possible to also determine relative blood flow by using a camera [25,35–37].

This technique is commonly called speckle contrast optical spectroscopy (SCOS) [25,35,31,38]

or speckle visibility spectroscopy (SVS) [28,35,36,39]. The movement of blood within the

brain causes this speckle field to fluctuate – the faster the blood flow, the faster the fluctuation.

Compared to diffuse wave spectroscopy, also known as diffuse correlation spectroscopy, which

utilizes a photodetector and requires a sampling frequency in the MHz range, SCOS, which utilizes

a CMOS sensor is not only more cost-effective to implement but also enhances signal sensitivity

by leveraging the large pixel count of the camera [36,39,40]. We previously demonstrated the

use of a compact SCOS device on the frontal area of the head to non-invasively measure brain

blood flow, by using a large source-to-detector (S-D) distance, greater than 3 cm [36].

In this paper, we combine a compact and comprehensive brain blood flow and blood volume

SCOS system with a breath-holding exercise to provide a direct physiology-based measurement of

the brain blood vessel condition. Such a system consists solely of a laser diode and a CMOS-based

board camera that can be placed on the head with no external optical elements, providing a

lightweight, portable, and budget-friendly design [36]. Monitoring of local cerebral blood flow

and blood volume rate changes during breath-holding was performed on a cohort of 50 subjects,

divided into two groups of 25 each. Subjects were evaluated using the Cleveland Clinic Stroke

Risk Calculator where the lower the score, the lower the risk [9]. One group was at low risk for

stroke (low-risk group, score 1), and the other was at higher risk for stroke (higher-risk group,

score

≥

4). By comparing the ratio of blood flow changes over blood volume changes—a proxy

to blood pressure changes—we aim to observe a highly significant difference between the groups

[25,41]. These findings would align with our expectations: during breath-holding among those at

elevated stroke risk, less flexible blood vessels impede dilation (lack of compliance), prompting

accelerated blood flow in response to the brain’s heightened demand for oxygen caused by

carbon dioxide buildup (hypercapnia). We further stratified the higher-risk cohort into smaller

sub-groups according to the Cleveland Clinic Stroke Risk Calculator. Our analysis revealed

significant correlations between stroke risk scores and flow to volume ratio changes during

breath-holding. This study offers preliminary physiological insights into stroke risk assessment

by establishing correlations between these two methods.

Research Article

Vol. 15, No. 10 /1 Oct 2024 /

Biomedical Optics Express

6086

2. Methods

2.1. Participants

Participants for this study were recruited from the Caltech, Pasadena, and greater Los Angeles local

communities, selected among adult humans aged from 18 to 65 years. Subjects uncomfortable

with holding their breath or with major respiratory diseases were excluded from this study. Before

the experiments, each participant completed a health questionnaire (showed in Supplement 1),

and their blood pressure was recorded. Informed consent was obtained from the partricipants

beforehand. The human research protocol for this study received approval from the Caltech

Committee for the Protection of Human Subjects and the Institutional Review Board (IRB). To

simplify the experiment and implementation of the device, measurements were conducted on

hairless areas, such as the forehead. The total illumination power was within the American

National Standards Institute (ANSI) laser safety standards for maximum skin exposure of a 785

nm laser beam.

A total of fifty-three subjects were enrolled. Three subjects were excluded from the analysis

following quality control of raw data (noise and artifact levels) or significant movements during

the recording. The final dataset encompasses 160 breath-holding entries from 50 subjects.

According to the prior questionnaire, the mean(std) ages of the low and higher risk groups were

31(5) and 60(4) years old respectively, 33 subjects were Female and 17 were Male. Traditional

stroke risk assessment was performed using the Cleveland Stroke Risk Calculator, which provides

a probability stroke risk score ranging from 1 to 10, with lower scores indicating lower risk.

Among the 50 subjects, 25 scored 1, 6 scored 4, 10 scored 5, and 9 scored between 6 and 7 in the

Cleveland Stroke Risk Calculator.

2.2. Compact SCOS device

The experimental configuration of the wearable SCOS device is shown in Fig. 1(A). It includes a

laser source for illumination and a board camera for detection. See Ref. [36] for a more detailed

presentation of the apparatus. For this study, we used a single-mode continuous wave

785

nm laser diode [Thorlabs L785H1] which can deliver up to 200 mW. To ensure control over the

illumination spot size and prevent undesirable laser light reflections, we housed the laser diode

within a 3D-printed mount [36]. The laser source and detecting units were positioned on one side

of the participants’ foreheads (see Fig. 1). The forehead was chosen as the device location, as

SCOS devices works better on hairless areas. The laser diode was set 5 mm away from the skin of

participants such that the illumination spot diameter was 5 mm. The total illumination power was

limited to 45 mW to ensure that the laser light intensity level of the area of illumination is well

within the American National Standards Institute (ANSI) laser safety standards for maximum

permissible exposure (2.95 mW/mm

) for skin exposure to a laser beam at 785 nm [42]. We used

a USB-board camera [Basler daA1920-160um (sensor Sony IMX392)]. For optimal performance

and stability, we typically operated the camera at a framerate of 60 frame-per-second (fps) and

with a global shutter setting recording 12-bit images. The camera featured a pixel pitch of 3.4

which offers a balance between the average intensity per pixel and the number of speckles per

pixel [43].

The source injects coherent laser light into the head of the participant. At a S-D distance from

the illumination spot on the scalp, a 60 frames per second camera captured the light that traveled

through the brain. The S-D distance was set at 3.2 cm for 43 subjects and 4.0 cm for 7 subjects,

with larger distances used when possible to optimize brain sensitivity. The S-D distance was

chosen based on a benchmark measurement before the breath-holding experiment. Cerebral

blood flow index (CBFI) and cerebral blood volume index (CBVI) were recorded at 3.2 cm and

4.0 cm over 20 seconds, and analysis of these time traces identified the optimal S-D distance by

checking for a clear heart rate peak in the Fourier transform of the CBFI and CBVI traces and