br Conflict of interest br Acknowledgments

Conflict of interest
Acknowledgments
Introduction Stress, one of the major causes of psychiatric disease, is a physiological, physical, and emotional response to an external or internal stimulus such as social pressure, life-threatening experiences, or any of numerous diseases [1]. It can initiate further mental health problems such as depression, loss of confidence, and apathy, among others [1,2]. Cortisol is a well-known biomarker of psychological stress [2–4]. In general, current rapid and accurate diagnostic tools for measuring stress use interviews, an electroencephalogram (EEG), an electrocardiogram (ECG), body temperature, and a self-questionnaire [2,5]. Although such methods have demonstrated a reasonable sensitivity and resolution for physiological responses to external stimuli [2], they are relatively complex and require bulky equipment to perform the test, making them less suitable for personal use in public settings or for point-of-care testing (POCT). There is no approach to point of care apoptosis testing more practical than a lateral flow assay (LFA). The naked eye can analyze the test qualitatively without any special equipment but cannot do so quantitatively [6]. The measurement of psychological stress should also be performed in real time. We were able to create a personalized cortisol testing platform based on a smartphone application. It utilizes a sensitive colorimetric LFA strip for specific detection and quantification of cortisol in human saliva (Fig. 1). A simple LED is used to light the test strip and quantify its intensity. Complementary metal oxide semiconductor (CMOS) image sensors are used to process health-related data through digital signals embedded in a smartphone. These cortisol LFA strip images are detected in real-time and data readings are taken by converting the red, green and blue signal data to hue (H) and brightness values. A curve-fitting method was used for quantification. In this paper, we describe a simple method exploiting a smartphone to achieve real-time measurement of human salivary cortisol in combination with a lateral flow immune-strip.
Materials and methods
Results and discussion The color signals for the assays correlate well with the analyte concentration, as shown in Fig. 3. Our strip measurement system was first calibrated by testing known concentrations of cortisol (0.01, 0.1, 1, 5, 10, 50, and 100ng/ml). The smartphone application was then calibrated with these samples. The data points and error bars in this figure represent the mean and relative standard deviation, respectively. The hue values were fit to an exponential decay curve of the cortisol concentration; the curve fit gave a confidence of determination (R2) of 0.7967 (Fig. 3A). The responses are linear between cortisol concentrations of 0.01–10ng/ml and the limit of detection is 1ng/ml. When brightness values are used to form the cortisol concentration function, the modified exponential decay curve-fitting of the cortisol concentration data gave coefficients of determination (R2) of 0.9837 (Fig. 3B). Therefore, we chose the calibration curve of brightness value as standard curve to validation the results. We also used Image J software (NIH, Bethesda, MD, USA) to measure the cortisol concentration at various concentrations (0.01, 0.1, 1, 5, 10, 50, 100ng/ml). The linear regression equation from these measurements was Y=−2.071ln(x)+108.69, which yielded an R2 coefficient value=0.9276 (Fig. 3C). The responses are linear between 0.01 and 10ng/ml cortisol, with a limit of detection of 0.01ng/ml cortisol. To further evaluate our device, we collected saliva from our laboratory members to measure the cortisol level in human saliva. The buffered solution was loaded into the strip. The time of detection was approximately 10min. We were then able to measure the cortisol concentration in our SLSM system. The data and error bars in this figure are the mean and relative standard deviation, respectively. Results from the human saliva samples were as follows: 1.2, 12.3, 3.3, 11.9, 3.8, 1.7, 1.2, 7.3 and 3.0ng/ml (Fig. 3D).