Faculty Bibliography

Steady progress in time-domain diffuse optical tomography (TD-DOT) technology is allowing for the first time the design of low-cost, compact, and high-performance systems, thus promising more widespread clinical TD-DOT use, such as for recording brain tissue hemodynamics. TD-DOT is known to provide more accurate values of optical properties and physiological parameters compared to its frequency-domain or steady-state counterparts. However, achieving high temporal resolution is still difficult, as solving the inverse problem is computationally demanding, leading to relatively long reconstruction times. The runtime is further compromised by processes that involve ‘nontrivial’ empirical tuning of reconstruction parameters, which increases complexity and inefficiency. To address these challenges, we present a new reconstruction algorithm that combines a deep-learning approach with our previously introduced sensitivity-equation-based, non-iterative sparse optical reconstruction (SENSOR) code. The new algorithm (called SENSOR-NET) unfolds the iterations of SENSOR into a deep neural network. In this way, we achieve high-resolution sparse reconstruction using only learned parameters, thus eliminating the need to tune parameters prior to reconstruction empirically. Furthermore, once trained, the reconstruction time is not dependent on the number of sources or wavelengths used. We validate our method with numerical and experimental data and show that accurate reconstructions with 1 mm spatial resolution can be obtained in under 20 milliseconds regardless of the number of sources used in the setup. This opens the door for real-time brain monitoring and other high-speed DOT applications.

Steady progress in time-domain diffuse optical tomography (TD-DOT) technology is allowing for the first time the design of low-cost, compact, and high-performance systems, thus promising more widespread clinical TD-DOT use, such as for recording brain tissue hemodynamics. TD-DOT is known to provide more accurate values of optical properties and physiological parameters compared to its frequency-domain or steady-state counterparts. However, achieving high temporal resolution is still difficult, as solving the inverse problem is computationally demanding, leading to relatively long reconstruction times. The runtime is further compromised by processes that involve 'nontrivial' empirical tuning of reconstruction parameters, which increases complexity and inefficiency. To address these challenges, we present a new reconstruction algorithm that combines a deep-learning approach with our previously introduced sensitivity-equation-based, non-iterative sparse optical reconstruction (SENSOR) code. The new algorithm (called SENSOR-NET) unfolds the iterations of SENSOR into a deep neural network. In this way, we achieve high-resolution sparse reconstruction using only learned parameters, thus eliminating the need to tune parameters prior to reconstruction empirically. Furthermore, once trained, the reconstruction time is not dependent on the number of sources or wavelengths used. We validate our method with numerical and experimental data and show that accurate reconstructions with 1 mm spatial resolution can be obtained in under 20 milliseconds regardless of the number of sources used in the setup. This opens the door for real-time brain monitoring and other high-speed DOT applications.

BACKGROUND:The association between sleep and Alzheimer's disease (AD) biomarkers are well-established, but little is known about how they interact to change the course of AD. OBJECTIVE:To determine the potential interaction between sleep disturbance and Aβ, tau, and APOE4 on brain atrophy and cognitive decline. METHODS:Sample included 351 participants (mean age 72.01 ± 6.67, 50.4%female) who were followed for approximately 5 years as part of the Alzheimer's Disease Neuroimaging Initiative. Informant-reported sleep disturbance (IRSD) was measured using the Neuropsychiatric Inventory (NPI). Changes in magnetic resonance imaging (MRI)-measured AD signature brain regions and cognitive performance and IRSD's interaction with cerebrospinal fluid amyloid-β (Aβ42) and p-Tau depositions and APOE4 status were examined using the linear mixed models. RESULTS:Baseline IRSD was not significantly associated with the rate of atrophy after adjusting for covariates (age, sex, education, total NPI severity score, and sleep medications). However, there was a significant interaction between IRSD and AD biomarkers on faster atrophy rates in multiple brain regions, including the cortical and middle temporal volumes. Post-hoc analyses indicated that Aβ and p-Tau/Aβ predicted a faster decline in these regions/domains in IRSD, compared with biomarker-negative individuals with IRSD (ps≤0.001). There was a significant IRSD*APOE4 interaction for brain atrophy rate (ps≤0.02) but not for cognition. CONCLUSION:IRSD may increase the future risk of AD by contributing to faster brain atrophy and cognitive decline when combined with the presence of AD biomarkers and APOE4. Early intervention for sleep disturbance could help reduce the risk of developing AD.