Faculty Bibliography

PURPOSE/OBJECTIVE: METHODS:Like many other MRI simulators, ours discretizes magnetic fields in space. However, we extend the MR signal simulation at each grid point from the 0th-order approximation, which assumes piecewise constant fields, to a 1st-order approximation, which assumes piecewise linear fields. We solve the signal equation by analytically integrating over each grid cube, assuming linear field variations, and then summing over all cubes. We provide analytical integrals for several pulse sequences. RESULTS:The 1st-order approximation captures strongly varying fields and associated intravoxel dephasing more accurately, avoiding severe "ringing" artifacts present in the usual 0th-order simulations. This enables simulations on a much coarser grid, facilitating computational feasibility. CONCLUSION/CONCLUSIONS:The first-order simulator enables the evaluation of unconventional scanner designs with strongly varying magnetic fields.

OBJECTIVES/OBJECTIVE:Despite its high negative predictive value (NPV) for clinically significant prostate cancer (csPCa), MRI suffers from a substantial number of false positives, especially for intermediate-risk cases. In this work, we determine whether a deep learning model trained with PI-RADS-guided representation learning can disambiguate the PI-RADS 3 classification, detect csPCa from bi-parametric prostate MR images, and avoid unnecessary benign biopsies. MATERIALS AND METHODS/METHODS:This study included 28,263 MR examinations and radiology reports from 21,938 men imaged for known or suspected prostate cancer between 2015 and 2023 at our institution (21 imaging locations with 34 readers), with 6352 subsequent biopsies. We trained a deep learning model, a representation learner (RL), to learn how radiologists interpret conventionally acquired T2-weighted and diffusion-weighted MR images, using exams in which the radiologists are confident in their risk assessments (PI-RADS 1 and 2 for the absence of csPCa vs. PI-RADS 4 and 5 for the presence of csPCa, n=21,465). We then trained biopsy-decision models to detect csPCa (Gleason score ≥7) using these learned image representations, and compared them to the performance of radiologists, and of models trained on other clinical variables (age, prostate volume, PSA, and PSA density) for treatment-naïve test cohorts consisting of only PI-RADS 3 (n=253, csPCa=103) and all PI-RADS (n=531, csPCa=300) cases. RESULTS:On the 2 test cohorts (PI-RADS-3-only, all-PI-RADS), RL-based biopsy-decision models consistently yielded higher AUCs in detecting csPCa (AUC=0.73 [0.66, 0.79], 0.88 [0.85, 0.91]) compared with radiologists (equivocal, AUC=0.79 [0.75, 0.83]) and the clinical model (AUCs=0.69 [0.62, 0.75], 0.78 [0.74, 0.82]). In the PIRADS-3-only cohort, all of whom would be biopsied using our institution's standard of care, the RL decision model avoided 41% (62/150) of benign biopsies compared with the clinical model (26%, P<0.001), and improved biopsy yield by 10% compared with the PI-RADS ≥3 decision strategy (0.50 vs. 0.40). Furthermore, on the all-PI-RADS cohort, RL decision model avoided 27% of additional benign biopsies (138/231) compared to radiologists (33%, P<0.001) with comparable sensitivity (93% vs. 92%), higher NPV (0.87 vs. 0.77), and biopsy yield (0.75 vs. 0.64). The combination of clinical and RL decision models further avoided benign biopsies (46% in PI-RADS-3-only and 62% in all-PI-RADS) while improving NPV (0.82, 0.88) and biopsy yields (0.52, 0.76) across the 2 test cohorts. CONCLUSIONS:Our PI-RADS-guided deep learning RL model learns summary representations from bi-parametric prostate MR images that can provide additional information to disambiguate intermediate-risk PI-RADS 3 assessments. The resulting RL-based biopsy decision models also outperformed radiologists in avoiding benign biopsies while maintaining comparable sensitivity to csPCa for the all-PI-RADS cohort. Such AI models can easily be integrated into clinical practice to supplement radiologists' reads in general and improve biopsy yield for any equivocal decisions.

Astrocytes differentiate and mature during postnatal development, but the molecular mechanisms linking their maturation to neuronal function remain unclear. We investigated the role of Wnt/β-catenin signaling and its effector, the transcription factor TCF7L2, in postnatal astrocytes using single-nucleus RNA sequencing, imaging, morphometric analysis, microdialysis, and electrophysiology in Tcf7l2 conditional knockout (cKO) mice. Loss of Tcf7l2 caused widespread transcriptional dysregulation in astrocytes, particularly in genes related to amino acid and ion transport, as well as membrane potential regulation. These mice showed disrupted amino acid homeostasis, astrocyte swelling, and impaired extracellular potassium clearance in the somatosensory cortex. These astrocytic changes were accompanied by altered gene expression in cortical pyramidal neurons, reduced excitability, and a hyperpolarized resting membrane potential. Our results suggest that astrocytic TCF7L2 is crucial in coordinating ion and amino acid transport in adulthood, thereby contributing to maintaining extracellular homeostasis and supporting neuronal function. This study identifies TCF7L2 as a key regulator of astrocyte-mediated neurophysiological support and underscores the importance of its role in astrocyte maturation during postnatal development.

Transcranial AC stimulation (tACS) of the cerebellum can entrain spiking activity in the Purkinje cells (PCs) of the cerebellar cortex and, through their projections, the cells in the cerebellar nuclei (CN). In this paper, we investigated if the cells in the motor thalamus (Mthal) can also be modulated (i.e. spikes entrained) via the CN-Mthal projections in rodents. A total of 82 thalamic cells were found, presumably in the Mthal by their stereotaxic coordinates, that were modulated by tACS of the cerebellum. Out of the 346 cells isolated, the thalamic cells with shorter action potentials and regular firing patterns had a higher probability of modulation by cerebellar stimulation than the cells with wider action potentials. The modulation level had a tuning curve with a maximum around 100-200 Hz. Spike histograms over the stimulation cycle transitioned between unimodal and bimodal distributions depending on the frequency. Most cells had a unimodal distribution at low frequencies, a bimodal distribution for frequencies between 80-125 Hz, and then a unimodal one for frequencies above 150 Hz. In addition, tACS of the motor cortex (MC) was also tested in a subset of thalamic cells. Unlike cerebellar stimulation, modulation levels peaked at two distinct frequencies, presumably due to entrainment through multiple MC-Mthal pathways with different preferred frequencies. The results demonstrate the feasibility of modulating a deep brain structure such as the thalamus through multi-synaptic pathways by stimulation of the cerebellar cortex (and the motor cortex) using a non-invasive neuromodulation method.

The autonomic nervous system (ANS) plays a vital role in health care for both acute care and chronic diseases. The traditional view of the ANS is to divide it into individual organ systems and study the separate components with a reductionist approach, which has been proven insufficient. Here, we argue that a holistic network-level view of the ANS is critical for generating new insights and deepening our understanding of its complex and dynamic functions. In this review, we treat the ANS as such a coordinated and dynamic network. We advocate for studying its interactions with major organ systems and the central nervous system using continuous and longitudinal monitoring in ambulatory and at-home settings rather than clinic-based snapshots. We first briefly review ANS physiology, then outline our network perspective, and finally highlight cutting-edge research directions and emerging engineering innovations in ANS monitoring, modeling, and modulation that benefit from this network-level view.

The locus coeruleus (LC) plays important roles in sleep/wake regulation and cognitive functions. LC neurons may be particularly sensitive to neural injury and serve as an early site of accumulation pathological tau in Alzheimer's disease. Obstructive sleep apnea (OSA) creates both chronic intermittent hypoxia and sleep fragmentation as potential insults to differentially sensitive neural populations including the locus coeruleus (LC). Using high field 7T imaging in cognitively normal older adults, we demonstrate that time spent with an oxygen saturation below 90% (T90), a measure of OSA's hypoxic burden, inversely correlates with LC structural integrity and explains significant variance in LC structural integrity after controlling for age, sex, and BMI. In contrast, other sleep variables such as the apnea-hypopnea index (AHI), total sleep time, and sleep efficiency did not contribute significant variance in LC structural integrity in this model. Thus, in the diagnosis of OSA, attention to hypoxic burden variables may be important in risk stratification for LC neural injury. This observation may inform future work determining whether mitigation of the hypoxemic burden from OSA can slow deterioration in LC integrity.

BACKGROUND:Familial dysautonomia (FD) is a hereditary neurodevelopmental disorder caused by aberrant splicing of the ELP1 gene, leading to a tissue-specific reduction in ELP1 protein expression. Preclinical models indicate that increasing ELP1 levels can mitigate disease manifestations. A blood-based ELP-1 protein assay may provide a reliable way to monitor gene target engagement. DESIGN AND METHODS/METHODS:Using a newly developed radioimmunoassay, we quantified ELP1 protein levels in peripheral blood samples collected from 59 homozygous FD patients carrying the IVS20 + 6T>C mutation and 66 heterozygous carriers. To assess the reproducibility of the measurement, replicate samples were collected in 43 participants. Longitudinal variability was evaluated in 22 participants who underwent repeat sampling 1 year later. RESULTS: = 0.827, p < 0.001). An ELP1 threshold of 492 pg/mL yielded a sensitivity of 80.2% (CI of 70.6 to 87.2%) and a specificity of 98.2% (95% CI of 90%-99%) with a positive likelihood ratio of 46.5, indicating that individuals with FD were over 46 times more likely to have ELP1 levels below this threshold compared to non-affected carriers. CONCLUSION/CONCLUSIONS:Blood ELP1 levels are robust and reproducible, with concentrations below 492 pg/mL strongly indicative of disease. Moreover, given their longitudinal stability, ELP1 can serve as a marker of target engagement to evaluate the efficacy of gene-targeted therapies aimed at correcting ELP1 gene splicing and protein production.