A Design Space Exploration (DSE) on Non-Invasive Sensing of Bladder Filling Using Near Infrared Spectroscopy (NIRS)

Urinary Incontinence (UI) is a widespread medical condition that affects one person from every three or four Americans. Near-Infrared Spectroscopy (NIRS) is a non-invasive under-study method for bladder filling sensation that can enhance the life quality of UI patients by finding the optimal voiding time. However, the application of NIRS to bladder volume sensing can be quite challenging due to three major obstacles: non-adequate traversal depth of NIR wavelengths, robustness and power efficiency requirements of the application, and low power transmission rate of NIR wavelengths. This work provides a Design Space Exploration (DSE) through the effect of various design parameters on NIRS applicability for bladder volume sensing. We investigate the impact of 7 different wavelengths from 650-950 nm, 16 possible detector-source distances, and 6 different sensation depths. The results of our work can be used as a guideline through optimal design and implementation of NIRS for bladder filling sensation.

💡 Research Summary

This paper investigates the feasibility of using Near‑Infrared Spectroscopy (NIRS) as a non‑invasive method for sensing bladder filling in patients with urinary incontinence (UI). UI affects roughly 25‑33 % of the U.S. population, and timely detection of bladder fullness could dramatically improve quality of life. While prior clinical studies have demonstrated that NIRS can detect bladder volume changes, three fundamental challenges have prevented practical deployment: (1) limited photon penetration depth due to strong absorption and scattering in abdominal tissues, especially in obese individuals; (2) the need for a robust, low‑power probe that tolerates mis‑placement and motion; and (3) the low transmission efficiency of NIR photons, which forces the system to operate near the minimum detectable power of the photodiode.

To address these issues, the authors performed a comprehensive Design Space Exploration (DSE) using Monte Carlo simulations. The exploration varied three key parameters: (i) wavelength (seven discrete values from 650 nm to 950 nm), (ii) source‑detector separation (sixteen distances ranging from 10 mm to 85 mm in 5 mm increments), and (iii) tissue thickness representing bladder depth (six values from 15 mm to 40 mm). For each combination, 500 million photons were launched, and photon trajectories were tracked through a multilayer tissue model whose optical properties (absorption coefficient μa, scattering coefficient μs, anisotropy g) were taken from published ex‑vivo measurements (Simpson et al., 1998).

A novel aspect of the simulation is the introduction of a Super‑Absorbing Layer (SAL) at the depth of interest. The SAL has an extremely high absorption coefficient and zero scattering, effectively “swallowing” any photon that reaches it. By comparing detector counts with and without the SAL, the authors directly measured the fraction of photons that actually penetrated to the target depth. This approach yields two complementary metrics: (a) the ratio of photons reaching the depth to the total number of input photons (overall transmission efficiency), and (b) the proportion of detected photons that originated from the target depth (a proxy for signal‑to‑noise ratio).

Results show a clear wavelength dependence. Shorter wavelengths (650‑700 nm) suffer from strong hemoglobin absorption and provide limited depth penetration. In the 800‑850 nm window, water and lipid absorption are minimized, allowing photons to reach depths of 30‑40 mm with the highest efficiency. At 900‑950 nm, water absorption rises again, reducing performance. Regarding source‑detector spacing, small separations (≤20 mm) yield high raw photon counts but most photons are back‑scattered from superficial layers, resulting in poor depth specificity. Larger separations (30‑45 mm) reduce total counts but dramatically increase the fraction of depth‑originated photons, improving the effective signal‑to‑noise ratio.

The authors translate these optical findings into power budgets. Using the energy relation E = N·hc/λ, they compute the energy transmission ratio η = N_detected/N_input for each configuration. Assuming a photodiode with a minimum sensible output power of ~10 nW (based on the QSB34 series datasheet), the required input optical power is back‑calculated. For the optimal configuration (≈800 nm wavelength, 35 mm source‑detector distance, 30 mm bladder depth), η is on the order of 1.2 × 10⁻⁴, implying an input power of roughly 0.5 mW. This level is well within the capabilities of a small rechargeable battery, suggesting that a wearable, battery‑operated NIRS probe could operate continuously for many hours.

The paper concludes that a probe employing an 800 nm wavelength source and a source‑detector separation of 30‑45 mm offers the best trade‑off between penetration depth, sensitivity, and power efficiency for bladder volume sensing. The SAL‑based depth‑measurement methodology provides a practical framework for future algorithm development that can estimate bladder fullness in real time. The authors recommend further work to validate the simulation results experimentally, to assess performance across a broader range of body habitus, and to evaluate mechanical robustness under motion. Their DSE methodology, however, establishes a solid foundation for the design of next‑generation, non‑invasive NIRS devices aimed at improving management of urinary incontinence.