top of page

Proof-of-Concept Whitepaper, Reports | January 2022 Technology for Real-Time Measurement of Salivary Cortisol

Study Highlights

  •’s breakthrough technology consistently measures salivary cortisol levels

  • provides the ability to measure cortisol levels anywhere, eliminating the need for expensive equipment and delayed results

  • This allows individuals to take actional steps to improve their wellness


In recent years, there has been an increasing need for mass deployable and scalable at-home diagnostics with near-instantaneous results. Currently, several unique challenges hinder the deployment of at-home diagnostics, including the use of high-end laboratory equipment, which means extraordinary costs for the patient. Another challenge includes limitations in current sample collections and analysis using the ordinary hardware that typical households possess. 

Furthermore, a majority of at-home test kits simply involve patients performing sample collection at home, and sending the samples to a laboratory for analysis, which often introduces a weeks-long delay in getting measurement results. 


Breakthrough Technology’s breakthrough technology platform enables full end-to-end at-home diagnostic measurements for key biomarkers.’s breakthrough technology leverages the most advanced computer vision and deep learning technologies (CVDL) to enable diagnostic measurements in conjunction with custom proprietary lateral flow assays (LFA) with only an iPhone and scalable LFA, eliminating the need for large or expensive laboratory equipment and skilled technicians.'s  novel application of this technology will be for the measurement of salivary cortisol, a human hormone known to affect nearly every biological process within the human body. By tracking cortisol on a daily basis, Paradigm’s technology enables home users to take actional steps from their cortisol measurements to improve their wellness. 

Proof-of-Concept Data

After extensive research and development, has demonstrated proof of concept (POC) and commercial viability of its CVDL and LFA technologies for the measurement of salivary cortisol.


In this POC evaluation, pooled and ultra-filtered human saliva were used as the initial inoculation media for the LFAs. The pooled saliva was allotted into 11 different lots and spiked with different concentrations of cortisol from 0-10ng/mL in approximately 1ng/mL steps. From each concentration batch of spiked human saliva, lot concentrations were measured using an FDA-approved salivary cortisol ELISA test in triplicate. Spiked ultrafiltered saliva was used to eliminate the variability in sample collection from actual users for the purposes of research and development (R&D). The mean of each triplicate was used as the concentration level (ground truth) of each spiked saliva sample lot. A manufacturing run of several thousand cortisol LFAs was produced and inoculated with known concentrations of cortisol from the spiked saliva samples previously prepared. In training our CVDL models, half the manufacturing lot was used for training the models in various lighting conditions, while the other half was used for actual cortisol measurements using our iPhone and app. 


For generating the measurement dataset, we chose N=6 for each concentration level and measured the inoculated LFAs with our iPhone and app. Measurement of each LFA was done with the LFA placed on a table with above-average ambient light from an overhead ceiling lamp. No lighting aids or other mechanical devices were between the iPhone camera and the LFA. 


​The results shown below plot the reading from our iPhone and app, demonstrating proof of concept and feasibility for’s technology.

Whitepaper Image 1.png

A linear regression model fit was demonstrated at R^2=0.9048, with a measurement range of 0-10ng/mL.

Whitepaper Image 2.png

A linear regression model fit was demonstrated at R^2=0.9574, with a measurement range of 0-5ng/mL.


The 0-5 range is 95% of cortisol measurements in healthy individuals.


The strong binning behavior of the data is due to the grouping in the X-axis,  reporting the ground truth values measured by ELISA. There’s a wider spread of measurements as the cortisol concentration increases due to variability in the LFA itself as the primary source, while the second source of variability is in the algorithm stability over repeated measurements and lighting conditions. 

Summary and Next Steps

​ demonstrates its proof of concept for its CVDL and LFA technology applied to cortisol-specific measurements. Additional testing involving end-to-end saliva collection and measurement is in progress and will be published at a later date. In addition, additional R&D efforts are needed to reduce the overall LFA variability and increase visual contrast for measurement using computer vision.

Scientific Advisory Board

Ioana A. Bina, M.D., Ph.D.

Integrative Medicine

Yale Trained Gastroenterologist

Ph.D. in Endocrinology

Read More


Prof. Phyllis Gardner, M.D.

Internal Medicine

Harvard Medical School

Stanford Professor of Medicine

Read More


Prof. Maria Hopman, M.D., Ph.D.


Radboud UMC

300+ Scientific Papers

Read More

Proof-of-Concept Whitepaper, Reports | January 2022 Technology for Real-Time Measurement of Salivary Cortisol

bottom of page