Input Function Research Articles

Comparison of delay correction methods to estimate 18F‐florbetaben kinetics in the brain using an image derived input function on total‐body EXPLORER PET

AbstractBackgroundAbsolute quantification of β‐amyloid binding from 18F‐florbetaben PET requires compartment kinetic modeling and an input function. This study uses non‐invasive Image Derived Input functions (IDIFs), derived using a new total‐body EXPLORER PET/CT scanner (Spencer et al., 2021). For accurate kinetic modelling, it is imperative to correct for time delay from the IDIF location to tracer arrival to the tissue of interest. The standard approach of Joint Estimation (JE) fits for delay time in tandem with other kinetic parameters and is computationally extensive (Feng et al., 2021). This study investigates the utility of Leading Edge (LE) ‐a novel and computationally efficient pulse timing method to estimate delay (Li et al., 2021) compared to JE in estimating 18F‐florbetaben kinetics in the brain grey matter regions, known to be vulnerable in Alzheimer’s Disease (AD).MethodThe study cohort included 14 individuals (9 cognitively unimpaired, 2 MCI, and 3AD; age = 66‐86 years) of the UC Davis Alzheimer’s Disease Research Center. Dynamic total‐body 18F‐florbetaben PET images were acquired for 110min using the uEXPLORER, and reconstructed (resolution = 2.344 mm isotropic voxels, framing protocol:30×2s, 12×10s, 7×60s, 20×300s).The dynamic brain‐cropped images were motion corrected (Jenkinson et al., 2002), and linearly registered to individual subject’s T1W (Jenkinson & Smith, 2001). Individual T1W were segmented using DKT atlas to obtain cerebral grey matter ROIs (Desikan et al., 2006). Dynamic time activity curves for each ROI were fit to the two‐tissue compartmental model, using a subject‐specific IDIF (population metabolite‐corrected) derived from the descending aorta (Figure1). Kinetic parameters were estimated and compared between JE and LE methods.ResultStrong associations were observed between JE and LE estimated 18F‐florbetaben kinetic parameters, including vb(r = 0.84, p<0.001), K1(r = 1, p<0.001), k2(r = 0.9, p<0.001), k3(r = 0.93, p<0.001), and k4(r = 0.86, p<0.001), in all the brain regions (Figure2). Quantitative kinetic results averaged across the selected grey matter regions (N = 11) are reported in Figure3.ConclusionThis study demonstrated the utility of IDIF extracted from total‐body dynamic PET to estimate 18F‐florbetaben tracer kinetics in the brain grey matter regions. Furthermore, the results show that LE can be an efficient surrogate for JE based approach in grey matter focused PET kinetic modelling of 18F‐florbetaben.

A cluster of neuropeptide S neurons regulates breathing and arousal

Neuropeptide S (NPS) is a highly conserved peptide found in all tetrapods that functions in the brain to promote heightened arousal; however, the subpopulations mediating these phenomena remain unknown. We generated mice expressing Cre recombinase from the Nps gene locus (NpsCre) and examined populations of NPS+ neurons in the lateral parabrachial area (LPBA), the peri-locus coeruleus (peri-LC) region of the pons, and the dorsomedial thalamus (DMT). We performed brain-wide mapping of input and output regions of NPS+ clusters and characterized expression patterns of the NPS receptor 1 (NPSR1). While the activity of all three NPS+ subpopulations tracked with vigilance state, only NPS+ neurons of the LPBA exhibited both increased activity prior to wakefulness and decreased activity during REM sleep, similar to the behavioral phenotype observed upon NPSR1 activation. Accordingly, we found that activation of the LPBA but not the peri-LC NPS+ neurons increased wake and reduced REM sleep. Furthermore, given the extended role of the LPBA in respiration and the link between behavioral arousal and breathing rate, we demonstrated that the LPBA but not the peri-LC NPS+ neuronal activation increased respiratory rate. Together, our data suggest that NPS+ neurons of the LPBA represent an unexplored subpopulation regulating breathing, and they are sufficient to recapitulate the sleep/wake phenotypes observed with broad NPS system activation.

Validating noninvasive techniques for 11C‐ER176 PET quantification in controls and Alzheimer’s disease patients

AbstractBackgroundNoninvasive yet accurate PET quantification methods are crucial for safe yet meaningful PET studies. [11C]ER176 is a PET radioligand that images neuroinflammation via the 18‐kDa translocator protein (TSPO), which is upregulated in activated microglia and has previously been a PET target in Alzheimer’s disease research. Currently, the gold‐standard method for quantifying ER176 PET is kinetic modeling of the dynamic PET data using the arterial input function (AIF), which is invasive and expensive. The current study explores a pseudo‐reference region (RR) technique to noninvasively quantify ER176 scans to facilitate clinical PET studies.MethodDynamic 90‐min [11C]ER176 PET scans were acquired in 11 participants, 5 controls (5M/0F) and 6 patients (3 MCI‐3M/0F; 3 AD‐2M/1F). Arterial blood data were collected throughout the duration of the scan and were used along with the dynamic imaging data to determine regional total volume of distribution (V T) using the two‐tissue compartment model. Standardized Uptake Value Ratio (SUVR) was calculated by normalizing the radioactivity in each region by that of the reference region (cerebellar gray matter). We explored differences across methods (AIF‐derived V T vs. RR‐derived SUVR) and across groups (controls vs. patients).ResultWithin each group, the outcome measures from AIF and RR methods were well correlated (Controls: R2 = 0.986, p < 0.001; Patients: R2 = 0.983, p < 0.001; Figure 1). Both methods showed greater TSPO expression in patients compared to controls, with mean differences in the medial temporal cortex of 14.99% between groups for V T and 10.65% between groups for SUVR, but lower variance for SUVR compared to V T (Figure 2). Both quantification methods showed similar regional patterns of percent differences between groups, corresponding to tau progression along Braak staging (Figure 3).ConclusionResults suggest that using the cerebellum, which does not accumulate amyloid plaques or tau aggregates until end‐stage disease, as a reference region for [11C]ER176 may be a valid method for quantifying differences between AD patients and controls compared to gold‐standard methodology. Regional increases in neuroinflammation are more reliable for SUVR than V T in AD‐related regions. Ongoing work includes a larger sample size to further validate noninvasive PET quantification.

Inflammatory and synaptic PET imaging in behavioural variant frontotemporal dementia caused by a repeat expansion in C9orf72 (C9orf72‐FTD)

AbstractBackgroundPathophysiological mechanisms associated with neurodegeneration include chronic neuroinflammation and synaptic dysfunction/loss. Radiotracer [11C]PBR28 is a marker of inflammation which binds to a translocator protein expressed by activated microglia. [11C]UCB‐J is a marker of synaptic density which binds to synaptic vesicle protein 2A. We evaluated both tracers as biomarkers of C9orf72‐FTD.MethodThree individuals with C9orf72‐FTD (mean age 62.7 years) underwent two PET scans between 3 and 13 months apart using [11C]PBR28 and [11C]UCB‐J. Dynamic PET data were acquired for 90 minutes following injection of each tracer. Arterial plasma input functions were generated from arterial blood samples. Participants also had volumetric MRI scans for each PET scan. Age‐matched controls from different cohorts underwent the same procedures (n = 5 for [11C]PBR28; n = 17 for [11C]UCB‐J).Regions of interest (ROIs) were defined on co‐registered T1‐weighted MR images using the CIC anatomical atlas. [11C]PBR28 BPND values were generated using a simplified reference tissue model with a cerebellar reference region. [11C]UCB‐J DVR‐1 values were generated using regional VT normalised to the VT in the centrum semiovale. ROI volumes were expressed as a percentage of total intracranial volume. W‐scores computed from each control cohort, using age and gender as covariates, illustrated the magnitude of differences between cases and controls. W‐scores >1.65 were considered abnormal for [11C]PBR28 and ←1.65 for [11C]UCB‐J and MRI.Result[11C]PBR28 w‐scores indicated significant neuroinflammation in FTD but varied between cases. Differences in BPND were greater than in volume in frontal/temporal/cingulate cortices in the individual with the longest disease duration (DD) (16‐17 years). Inversely, in the individual with the shortest DD (1‐2 years), there were greater differences in volume than BPND in most ROIs.[11C]UCB‐J w‐scores indicated significant synaptic dysfunction/loss and were lower than w‐scores of volume in all cortical ROIs. In the individual with the longest DD, differences in [11C]PBR28 were greater than in [11C]UCB‐J in most ROIs. An inverse pattern again observed in the individual with the shortest DD suggests synaptic dysfunction/loss may precede significant neuroinflammation.ConclusionInflammatory and synaptic PET may provide more sensitive biomarkers of C9orf72‐FTD than MRI. Variability in PET signal within this genotype may be explained by DD.

Manipulation of Dietary DHA does not Alter DHA Brain Uptake Kinetics in Mice Despite Changing Brain and Plasma DHA Levelsvs

AbstractBackgroundIncreased brain docosahexaenoic acid (DHA) uptake reflects a compensatory mechanism for maintaining brain DHA homeostasis and can be captured through PET‐MRI scan uptake studies using radiolabeled DHA. It has been previously shown that APOE4 status and alcohol intake alter brain DHA uptake in mice. Whether a change in DHA intake affects the brain DHA incorporation coefficient (K*) however is not known.MethodFour‐month‐old C57BL/6J mice were fed either a diet rich in DHA (15 kcal% Fat, 1.7 g/kg ALA, and 7g/kg DHA) or deficient in DHA (15 kcal% Fat and 0.9 mg/kg ALA) for 4 months. At 8 months, anesthetized mice were subjected to dynamic PET‐MRI for 30 min following i.v. injection of 22‐[18F]fluorodocosahexaenoic acid ([18F]FDHA), prepared in‐house through radiosynthesis. Using an image‐derived cardiac input function, brain uptake was assessed by estimating [18F]FDHA K* values using both the Patlak model approach and the irreversible two‐tissue compartmental model (Irr2TCM) with and without correction of the brain radioactivity for radioactive spillover from the skull. Mice were then sacrificed, and the plasma and brains were analyzed using liquid chromatography‐mass spectrometry (LC‐MS) to evaluate differences in the unlabeled concentrations of fatty acids into phospholipids, triglycerides, and other lipids.ResultThere was no significant difference in the [18F]FDHA K* estimates (with either Patlak or Irr2TCM) between mice fed a DHA‐rich or a DHA‐deficient diet. LC‐MS based lipidomic analysis showed that the DHA‐rich diet significantly increased DHA concentration in triglycerides in plasma and in phosphatidylethanolamine in the brain.ConclusionThese results suggest that under physiological conditions, increased dietary intake of DHA in wild‐type mice increases DHA in brain phosphatidylethanolamine and in plasma triglycerides without changing the DHA K*. The net effect likely is increased brain DHA uptake, equal to the product of unesterified DHA plasma concentration and K*. Imaging brain DHA with dynamic [18F]FDHA PET‐MRI is a potentially useful tool not only to investigate the APOE4 phenotype, which promotes DHA dysregulation, but also to assess dietary and pharmacological interventions that affect brain DHA homeostasis as well as to guide future treatments.

Validating noninvasive techniques for 11C‐ER176 PET quantification in controls and Alzheimer’s disease patients

AbstractBackgroundNoninvasive yet accurate PET quantification methods are crucial for safe yet meaningful PET studies. [11C]ER176 is a PET radioligand that images neuroinflammation via the 18‐kDa translocator protein (TSPO), which is upregulated in activated microglia and has previously been a PET target in Alzheimer’s disease research. Currently, the gold‐standard method for quantifying ER176 PET is kinetic modeling of the dynamic PET data using the arterial input function (AIF), which is invasive and expensive. The current study explores a pseudo‐reference region (RR) technique to noninvasively quantify ER176 scans to facilitate clinical PET studies.MethodDynamic 90‐min [11C]ER176 PET scans were acquired in 11 participants, 5 controls (5M/0F) and 6 patients (3 MCI‐3M/0F; 3 AD‐2M/1F). Arterial blood data were collected throughout the duration of the scan and were used along with the dynamic imaging data to determine regional total volume of distribution (VT) using the two‐tissue compartment model. Standardized Uptake Value Ratio (SUVR) was calculated by normalizing the radioactivity in each region by that of the reference region (cerebellar gray matter). We explored differences across methods (AIF‐derived VT vs. RR‐derived SUVR) and across groups (controls vs. patients).ResultWithin each group, the outcome measures from AIF and RR methods were well correlated (Controls: R2 = 0.986, p < 0.001; Patients: R2 = 0.983, p < 0.001; Figure 1). Both methods showed greater TSPO expression in patients compared to controls, with mean differences in the medial temporal cortex of 14.99% between groups for VT and 10.65% between groups for SUVR, but lower variance for SUVR compared to VT (Figure 2). Both quantification methods showed similar regional patterns of percent differences between groups, corresponding to tau progression along Braak staging (Figure 3).ConclusionResults suggest that using the cerebellum, which does not accumulate amyloid plaques or tau aggregates until end‐stage disease, as a reference region for [11C]ER176 may be a valid method for quantifying differences between AD patients and controls compared to gold‐standard methodology. Regional increases in neuroinflammation are more reliable for SUVR than VT in AD‐related regions. Ongoing work includes a larger sample size to further validate noninvasive PET quantification.

Posterior Primary Progressive Prosopagnosia: Imaging and Neuropathology

AbstractBackgroundProsopagnosia is common with right temporal lesions, often in the fusiform gyrus when caused by vascular pathology. When caused by frontotemporal dementia, the lesion typically begins in the temporal tip. We are reporting advanced imaging, including tau and inflammation, as well as neuropathology in a patient with pathology centered at the posterior extent of the fusiform gyrus.MethodA 69‐year‐old man with apperceptive prosopagnosia and a negative amyloid positron emission tomogram (PET) underwent MRI and PET using 11C‐PBR28, an inflammation tracer, and 18F‐flortaucipir. V T values for 11C‐PBR28 were calculated with the Logan plot and a metabolite‐corrected arterial input function. For 18F‐flortaucipir, the SUV ratio over the cerebellar gray matter was calculated for t = 80‐100 min. A neuropathological examination was conducted using standard procedures as well as phospho‐tau histochemistry.ResultAtrophy and 18F‐flortaucipir uptake involved a similar region at the posterior extent of the right fusiform gyrus and, minimally, a symmetrical region of the left fusiform gyrus and an area in right premotor cortex. Involving the same regions, the 11C‐PBR28 signal was however more extensive than atrophy or 18F‐flortaucipir signal. At autopsy, the patient had corticobasal degeneration (CBD), with severe 4R tau involvement of temporo‐occipital, inferior parietal, and frontal cortices, as well as limbic regions and subcortical nuclei. The location of CBD pathology in this patient was highly atypical, as shown by comparing the neuropathological findings with the findings in two typical CBD patients.ConclusionAlthough prosopagnosia from ischemic and other lesions in the right posterior temporo‐occipital region is common, primary progressive prosopagnosia in the absence of amyloid has been described mostly with involvement of the right anterior temporal region. As exemplified by our patient, it can rarely happen with posterior temporal involvement and in this case the pathology was not TDP‐43 type C, as present in most patients with anterior temporal involvement, but corticobasal degeneration, a 4R tauopathy.

Posterior Primary Progressive Prosopagnosia: Imaging and Neuropathology

AbstractBackgroundProsopagnosia is common with right temporal lesions, often in the fusiform gyrus when caused by vascular pathology. When caused by frontotemporal dementia, the lesion typically begins in the temporal tip. We are reporting advanced imaging, including tau and inflammation, as well as neuropathology in a patient with pathology centered at the posterior extent of the fusiform gyrus.MethodA 69‐year‐old man with apperceptive prosopagnosia and a negative amyloid positron emission tomogram (PET) underwent MRI and PET using 11C‐PBR28, an inflammation tracer, and 18F‐flortaucipir. VT values for 11C‐PBR28 were calculated with the Logan plot and a metabolite‐corrected arterial input function. For 18F‐flortaucipir, the SUV ratio over the cerebellar gray matter was calculated for t = 80‐100 min. A neuropathological examination was conducted using standard procedures as well as phospho‐tau histochemistry.ResultAtrophy and 18F‐flortaucipir uptake involved a similar region at the posterior extent of the right fusiform gyrus and, minimally, a symmetrical region of the left fusiform gyrus and an area in right premotor cortex. Involving the same regions, the 11C‐PBR28 signal was however more extensive than atrophy or 18F‐flortaucipir signal. At autopsy, the patient had corticobasal degeneration (CBD), with severe 4R tau involvement of temporo‐occipital, inferior parietal, and frontal cortices, as well as limbic regions and subcortical nuclei. The location of CBD pathology in this patient was highly atypical, as shown by comparing the neuropathological findings with the findings in two typical CBD patients.ConclusionAlthough prosopagnosia from ischemic and other lesions in the right posterior temporo‐occipital region is common, primary progressive prosopagnosia in the absence of amyloid has been described mostly with involvement of the right anterior temporal region. As exemplified by our patient, it can rarely happen with posterior temporal involvement and in this case the pathology was not TDP‐43 type C, as present in most patients with anterior temporal involvement, but corticobasal degeneration, a 4R tauopathy.

Time-dependent Dynamics of the Corona

We present in this Letter the first global comparison between traditional line-tied steady-state magnetohydrodynamic models and a new, fully time-dependent thermodynamic magnetohydrodynamic simulation of the global corona. To approximate surface magnetic field distributions and magnitudes around solar minimum, we use the Lockheed Evolving Surface-Flux Assimilation Model to obtain input maps that incorporate flux emergence and surface flows over a full solar rotation, including differential rotation and meridional flows. Each time step evolves the previous state of the plasma with a new magnetic field input boundary condition, mimicking photospheric driving on the Sun. We find that this method produces a qualitatively different corona compared to steady-state models. The magnetic energy levels are higher in the time-dependent model, and coronal holes evolve more along the following edge than they do in steady-state models. Coronal changes, as illustrated with forward-modeled emission maps, evolve on longer timescales with time-dependent driving. We discuss implications for active and quiet Sun scenarios, solar wind formation, and widely used steady-state assumptions like potential field source surface calculations.

Single time point quantitation of cerebral glucose metabolism by FDG-PET without arterial sampling

BackgroundUntil recently, quantitation of the net influx of 2-[18F]fluorodeoxyglucose (FDG) to brain (Ki) and the cerebrometabolic rate for glucose (CMRglc) required serial arterial blood sampling in conjunction with dynamic positron emission tomography (PET) recordings. Recent technical innovations enable the identification of an image-derived input function (IDIF) from vascular structures, but are frequently still encumbered by the need for interrupted sequences or prolonged recordings that are seldom available outside of a research setting. In this study, we tested simplified methods for quantitation of FDG-Ki by linear graphic analysis relative to the descending aorta IDIF in oncology patients examined using a Biograph Vision 600 PET/CT with continuous bed motion (Aarhus) or using a recently installed Biograph Vision Quadra long-axial field-of-view (FOV) scanner (Bern).ResultsCorrelation analysis of the coefficients of a tri-exponential decomposition of the IDIFs measured during 67 min revealed strong relationships among the total area under the curve (AUC), the terminal normalized arterial integral (theta(52–67 min)), and the terminal image-derived arterial FDG concentration (Ca(52–67 min)). These relationships enabled estimation of the missing AUC from late recordings of the IDIF, from which we then calculated FDG-Ki in brain by two-point linear graphic analysis using a population mean ordinate intercept and the single late frame. Furthermore, certain aspects of the IDIF data from Aarhus showed a marked age-dependence, which was not hitherto reported for the case of FDG pharmacokinetics.ConclusionsThe observed interrelationships between pharmacokinetic parameters in the IDIF measured during the PET recording support quantitation of FDG-Ki in brain using a single averaged frame from the interval 52–67 min post-injection, with minimal error relative to calculation from the complete dynamic sequences.

Whole-brain mapping of monosynaptic afferent inputs to the CRH neurons in the medial prefrontal cortex of mice.

Corticotropin-releasing hormone (CRH) neurons are densely distributed in the medial prefrontal cortex (mPFC), which plays a crucial role in integrating and processing emotional and cognitive inputs from other brain regions. Therefore, it is important to know the neural afferent patterns of mPFCCRH neurons, which are still unclear. Here, we utilized a rabies virus-based monosynaptic retrograde tracing system to map the presynaptic afferents of the mPFCCRH neurons throughout the entire brain. The results show that the mPFCCRH neurons receive inputs from three main groups of brain regions: (1) the cortex, primarily the orbital cortex, somatomotor areas, and anterior cingulate cortex; (2) the thalamus, primarily the anteromedial nucleus, mediodorsal thalamic nucleus, and central medial thalamic nucleus; and (3) other brain regions, primarily the basolateral amygdala, hippocampus, and dorsal raphe nucleus. Taken together, our results are valuable for further investigations into the roles of the mPFCCRH neurons in normal and neurological disease states. These investigations can shed light on various aspects such as cognitive processing, emotional modulation, motivation, sociability, and pain.

Parametric and non-parametric Poisson regression for modelling of the arterial input function in positron emission tomography

Full quantification of Positron Emission Tomography (PET) requires an arterial input function (AIF) for measurement of certain targets, or using particular radiotracers, or for the quantification of specific outcome measures. The AIF represents the measurement of radiotracer concentrations in the arterial blood plasma over the course of the PET examination. Measurement of the AIF is prone to error as it is a composite measure created from the combination of multiple measurements of different samples with different equipment, each of which can be sources of measurement error. Moreover, its measurement requires a high degree of temporal granularity for early time points, which necessitates a compromise between quality and quantity of recorded samples. For these reasons, it is often desirable to fit models to this data in order to improve its quality before using it for quantification of radiotracer binding in the tissue. The raw observations of radioactivity in arterial blood and plasma samples are derived from radioactive decay, which is measured as a number of recorded counts. Count data have several specific properties, including the fact that they cannot be negative as well as a particular mean-variance relationship. Poisson regression is the most principled modelling strategy for working with count data, as it both incorporates and exploits these properties. However, no previous studies to our knowledge have taken this approach, despite the advantages of greater efficiency and accuracy which result from using the appropriate distributional assumptions. Here, we implement a Poisson regression modelling approach for the AIF as proof-of-concept of its application. We applied both parametric and non-parametric models for the input function curve. We show that a negative binomial distribution is a more appropriate error distribution for handling overdispersion. Furthermore, we extend this approach to a hierarchical non-parametric model which is shown to be highly resilient to missing data. We thus demonstrate that Poisson regression is both feasible and effective when applied to AIF data, and propose that this is a promising strategy for modelling blood count data for PET in future.

Variationally mimetic operator networks

In recent years operator networks have emerged as promising deep learning tools for approximating the solution to partial differential equations (PDEs). These networks map input functions that describe material properties, forcing functions, and boundary data to the solution of a PDE. This work describes a new architecture for operator networks, called the variationally mimetic operator network (VarMiON ), that mimics the form of the numerical solution obtained from an approximate variational or weak formulation of the problem. Like the conventional Deep Operator Network (DeepONet) the VarMiON is also composed of a sub-network that constructs the basis functions for the output and another that constructs the coefficients for these basis functions. However, in contrast to the DeepONet, the architecture of these sub-networks in the VarMiON is precisely determined. An analysis of the error in the VarMiON solution reveals that it contains contributions from the error in the training data, the training error, the quadrature error in sampling input and output functions, and a “covering error” that measures the distance between the test input functions and the nearest functions in the training dataset. It also depends on the stability constants for the exact solution operator and its VarMiON approximation. The application of the VarMiON to a canonical elliptic PDE and a nonlinear PDE reveals that for approximately the same number of network parameters, on average the VarMiON incurs smaller errors than a standard DeepONet and a recently proposed multiple-input operator network (MIONet). Further, its performance is more robust to variations in input functions, the techniques used to sample the input and output functions, the techniques used to construct the basis functions, and the number of input functions. Moreover, it consistently outperforms baseline methods at various dataset sizes. The data and code accompanying this manuscript are publicly available at https://github.com/dhruvpatel108/VarMiON.

Motor Processing in Children With Cochlear Implants as Assessed by Functional Near-Infrared Spectroscopy.

Auditory-motor and visual-motor networks are often coupled in daily activities, such as when listening to music and dancing; but these networks are known to be highly malleable as a function of sensory input. Thus, congenital deafness may modify neural activities within the connections between the motor, auditory, and visual cortices. Here, we investigated whether the cortical responses of children with cochlear implants (CI) to a simple and repetitive motor task would differ from that of children with typical hearing (TH) and we sought to understand whether this response related to their language development. Participants were 75 school-aged children, including 50 with CI (with varying language abilities) and 25 controls with TH. We used functional near-infrared spectroscopy (fNIRS) to record cortical responses over the whole brain, as children squeezed the back triggers of a joystick that vibrated or not with the squeeze. Motor cortex activity was reflected by an increase in oxygenated hemoglobin concentration (HbO) and a decrease in deoxygenated hemoglobin concentration (HbR) in all children, irrespective of their hearing status. Unexpectedly, the visual cortex (supposedly an irrelevant region) was deactivated in this task, particularly for children with CI who had good language skills when compared to those with CI who had language delays. Presence or absence of vibrotactile feedback made no difference in cortical activation. These findings support the potential of fNIRS to examine cognitive functions related to language in children with CI.

Deep neural operators can predict the real-time response of floating offshore structures under irregular waves

The utilization of neural operators in a digital twin model of an offshore floating structure holds the potential for a significant shift in the prediction of structural responses and health monitoring, offering valuable real-time control insights. In this work, we investigate the effectiveness of three neural operators, namely the deep operator network (DeepONet), the Fourier neural operator (FNO), and the Wavelet neural operator (WNO), to accurately capture the responses of a floating structure under six different sea state codes (3−8) based on the wave characteristics described by the World Meteorological Organization (WMO). To further enhance the accuracy of the vanilla architecture of the neural operators, novel extensions, such as wavelet-DeepONet and self-adaptive WNO, are proposed in this paper. The results demonstrate that these high-precision neural operators can deliver structural responses more efficiently, up to two orders of magnitude faster than a dynamic analysis using conventional numerical solvers. Additionally, compared to gated recurrent units (GRUs), a commonly used recurrent neural network for time-series estimation, neural operators are both more accurate and efficient, especially in situations with limited data availability. Taken together, our study shows that FNO outperforms all other operators for approximating the mapping of one input functional space to the output space as well as for responses that have small bandwidth of the frequency spectrum. Conversely, DeepONet, with historical states, proves most accurate in learning the mapping of multiple input functions to the output space and capturing responses within a broad frequency spectrum.

Pulmonary Transit Time Derived from First-Pass Perfusion Cardiac MR Imaging: A Potential New Marker for Cardiac Involvement and Prognosis in Light-Chain Amyloidosis.

First-pass perfusion cardiac MR imaging could reflect pulmonary hemodynamics. However, the clinical value of pulmonary transit time (PTT) derived from first-pass perfusion MRI in light-chain (AL) amyloidosis requires further evaluation. To assess the clinical and prognostic value of PTT in patients with AL amyloidosis. Prospective observational study. 226 biopsy-proven systemic AL amyloidosis patients (age 58.62 ± 10.10 years, 135 males) and 43 healthy controls (age 42 ± 16.2 years, 20 males). SSFP cine and phase sensitive inversion recovery late gadolinium enhancement (LGE) sequences, and multislice first-pass myocardial perfusion imaging with a saturation recovery turbo fast low-angle shot (SR-TurboFLASH) pulse sequence at 3.0T. PTT was measured as the time interval between the peaks of right and left ventricular cavity arterial input function curves on first-pass perfusion MR images. Independent-sample t test, Mann-Whitney U test, Chi-square test, Fisher's exact test, analysis of variance, or Kruskal-Wallis test, as appropriate; univariable and multivariable Cox proportional hazards models and Kaplan-Meier curves, area under receiver operating characteristic curve were used to determine statistical significance. PTT could differentiate AL amyloidosis patients with (N = 188) and without (N = 38) cardiac involvement (area under the curve [AUC] = 0.839). During a median follow-up of 35 months, 160 patients (70.8%) demonstrated all-cause mortality. After adjustments for clinical (Hazard ratio [HR] 1.061, confidence interval [CI]: 1.021-1.102), biochemical (HR 1.055, CI: 1.014-1.097), cardiac MRI-derived (HR 1.077, CI: 1.034-1.123), and therapeutic (HR 1.063, CI: 1.024-1.103) factors, PTT predicted mortality independently in patients with AL amyloidosis. Finally, PTT could identify worse outcomes in patients demonstrating New York Heart Association class III, Mayo 2004 stage III, and transmural LGE pattern. PTT may serve as a new imaging predictor of cardiac involvement and prognosis in AL amyloidosis. 2 TECHNICAL EFFICACY: Stage 2.

A patient-specific timing protocol for improved CT pulmonary angiography

Rationale and objectivesTo improve the image quality of CT pulmonary angiography (CTPA) using a patient-specific timing protocol. Material and methodsA total of 24 swine (48.5 ± 14.3 kg) underwent continuous contrast-enhanced dynamic CT acquisition over 30 s to capture the pulmonary arterial input function (AIF). Multiple contrast injections were made under different cardiac outputs (1.4–5.1 L/min), resulting in a total of 154 AIF curves. The volume scans with maximal enhancement in these AIF curves were retrospectively selected as the reference standard (group A). Two prospective CTPA protocols with bolus-tracking were then simulated using these AIF curves: one used a fixed delay of 5 s between triggering and CTPA acquisition (group B), while the other used a specific delay based on one-half of the contrast injection duration (group C). The mean attenuation, signal-to-noise (SNR) and contrast-to-noise ratios (CNR) between the three groups were then compared using independent sample t-test. Subjective image quality scores were also compared using Wilcoxon-Mann-Whitney test. ResultsThe mean attenuation of pulmonary arteries for group A, B and C (expressed in [HU]) were 870.1 ± 242.5 HU, 761.1 ± 246.7 HU and 825.2 ± 236.8 HU, respectively. The differences in the mean SNR and CNR between Group A and Group C were not significant (SNR: 65.2 vs. 62.4, CNR: 59.6 vs. 56.4, both p > 0.05), while Group B was significantly lower than Group A (p < 0.05). ConclusionThe image quality of CT pulmonary angiography is significantly improved with a timing protocol determined using contrast injection delivery time, as compared with a standard timing protocol with a fixed delay between bolus triggering and image acquisition.

Marine diesel engine turbocharger fouling phenomenon risk assessment application by using fuzzy FMEA method

Turbocharger fouling phenomenon was analyzed from the risk assessment perspective in this study. The research employed exhaust system and turbocharger equipment of commercial ship that equipped with a Doosan-MAN B&amp;W 6S50 MC-C diesel engine was used as the main research materials, and utilized the fuzzy Failure Mode and Effects Analysis (FMEA) based on the expert system as a methodical approach. The experts revealed different types of turbocharger fouling failure modes (FMs), along with their respective causes and subsequent consequences. Following that, the specialists allocated an O, S, and D score to each FM. Within the framework of fuzzy logic, the process entails the establishment of input and output membership functions, as well as the construction of a fuzzy model incorporating an inference mechanism and a rule base. Based on the analysis findings, the three primary factors are as follows: low cylinder compression pressure with a Fuzzy Risk Priority Number (FRPN) score of 6.95, high main engine fuel oil consumption with a score of 6.92, and high CO, CO2, SOx emissions with a 6.45. The phenomenon of turbocharger fouling, being an inherent occurrence, has significant ramifications on the main engine, the vessel as a whole, and the ecological surroundings. The quantitative results presented in this study provide valuable insights into the risks associated with maritime endeavors. The data generated from this research can be used by stakeholders in the maritime industry to better understand this situation and take proactive measures to mitigate potential risks in the future. Furthermore, the findings of the research provide corroboration for the implementation of predictive maintenance procedures.

Evaluation of trans- and cis-4-[18F]Fluorogabapentin for Brain PET Imaging.

Gabapentin, a selective ligand for the α2δ subunit of voltage-dependent calcium channels, is an anticonvulsant medication used in the treatment of neuropathic pain, epilepsy, and other neurological conditions. We recently described two radiofluorinated derivatives of gabapentin (trans-4-[18F]fluorogabapentin, [18F]tGBP4F, and cis-4-[18F]fluorogabapentin, [18F]cGBP4F) and showed that these compounds accumulate in the injured nerves in a rodent model of neuropathic pain. Given the use of gabapentin in brain diseases, here we investigate whether these radiofluorinated derivatives of gabapentin can be used for imaging α2δ receptors in the brain. Specifically, we developed automated radiosynthesis methods for [18F]tGBP4F and [18F]cGBP4F and conducted dynamic PET imaging in adult rhesus macaques with and without preadministration of pharmacological doses of gabapentin. Both radiotracers showed very high metabolic stability, negligible plasma protein binding, and slow accumulation in the brain. [18F]tGBP4F, the isomer with higher binding affinity, showed low brain uptake and could not be displaced, whereas [18F]cGBP4F showed moderate brain uptake and could be partially displaced. Kinetic modeling of brain regional time-activity curves using a metabolite-corrected arterial input function shows that a one-tissue compartment model accurately fits the data. Graphical analysis using Logan or multilinear analysis 1 produced similar results as compartmental modeling, indicating robust quantification. This study advances our understanding of how gabapentinoids work and provides an important advancement toward imaging α2δ receptors in the brain.

Differential flatness-based adaptive robust tracking control for wheeled mobile robots with slippage disturbances

Wheeled mobile robots (WMRs) have a wide range of applications in logistics transportation and industrial productions, among which the motion control has always been one of the hot spots in the current WMR researches. However, most of previous designed controllers assumed that the WMR motion had no slippage. Ignoring the slippage factors usually results in a decrease in control performance and even leads to unstable motion. To address such a challenge, a kinematic model with differential flatness is established through dynamic feedback-linearization, which comprehensively considers the multidirectional slippage of mobile robot, including longitudinal and steering slippage. Subsequently, benefited from the one-to-one mapping of states and inputs to flat outputs in differential flat system, an adaptive robust control (ARC) method is proposed to stabilize the system. Different from previous robust control studies, even if the knowledge of the upper bound of system uncertainties is unknown in advance, the proposed adaptive robust controller can still achieve satisfying performance by adaptive estimation of the upper bound of system uncertainties. The effectiveness and feasibility of the proposed method are confirmed by comparative experiments on WMR with slippage disturbance.