Machine Learning-Based Prediction Models for Different Clinical Risks in Different Hospitals: Evaluation of Live Performance.

BACKGROUND: Machine learning algorithms are currently used in a wide array of clinical domains to produce models that can predict clinical risk events. Most models are developed and evaluated with retrospective data, very few are evaluated in a clinical workflow, and even fewer report performances in different hospitals. In this study, we provide detailed evaluations of clinical risk prediction models in live clinical workflows for three different use cases in three different hospitals. OBJECTIVE: The main objective of this study was to evaluate clinical risk prediction models in live clinical workflows and compare their performance in these setting with their performance when using retrospective data. We also aimed at generalizing the results by applying our investigation to three different use cases in three different hospitals. METHODS: We trained clinical risk prediction models for three use cases (ie, delirium, sepsis, and acute kidney injury) in three different hospitals with retrospective data. We used machine learning and, specifically, deep learning to train models that were based on the Transformer model. The models were trained using a calibration tool that is common for all hospitals and use cases. The models had a common design but were calibrated using each hospital's specific data. The models were deployed in these three hospitals and used in daily clinical practice. The predictions made by these models were logged and correlated with the diagnosis at discharge. We compared their performance with evaluations on retrospective data and conducted cross-hospital evaluations. RESULTS: The performance of the prediction models with data from live clinical workflows was similar to the performance with retrospective data. The average value of the area under the receiver operating characteristic curve (AUROC) decreased slightly by 0.6 percentage points (from 94.8% to 94.2% at discharge). The cross-hospital evaluations exhibited severely reduced performance: the average AUROC decreased by 8 percentage points (from 94.2% to 86.3% at discharge), which indicates the importance of model calibration with data from the deployment hospital. CONCLUSIONS: Calibrating the prediction model with data from different deployment hospitals led to good performance in live settings. The performance degradation in the cross-hospital evaluation identified limitations in developing a generic model for different hospitals. Designing a generic process for model development to generate specialized prediction models for each hospital guarantees model performance in different hospitals.

Exploring Novel Funding Strategies for Innovative Medical Research: The HORAO Crowdfunding Campaign.

BACKGROUND: The rise of the internet and social media has boosted online crowdfunding as a novel strategy to raise funds for kick-starting projects, but it is rarely used in science. OBJECTIVE: We report on an online crowdfunding campaign launched in the context of the neuroscience project HORAO. The aim of HORAO was to develop a noninvasive real-time method to visualize neuronal fiber tracts during brain surgery in order to better delineate tumors and to identify crucial cerebral landmarks. The revenue from the crowdfunding campaign was to be used to sponsor a crowdsourcing campaign for the HORAO project. METHODS: We ran a 7-week reward-based crowdfunding campaign on a national crowdfunding platform, offering optional material and experiential rewards in return for a contribution toward raising our target of Swiss francs (CHF) 50,000 in financial support (roughly equivalent to US $50,000 at the time of the campaign). We used various owned media (websites and social media), as well as earned media (press releases and news articles) to raise awareness about our project. RESULTS: The production of an explanatory video took 60 hours, and 31 posts were published on social media (Facebook, Instagram, and Twitter). The campaign raised a total of CHF 69,109. Approximately half of all donations came from donors who forwent a reward (CHF 28,786, 48.74%); the other half came from donors who chose experiential and material rewards in similar proportions (CHF 14,958, 25.33% and CHF 15,315.69, 25.93%, respectively). Of those with an identifiable relationship to the crowdfunding team, patients and their relatives contributed the largest sum (CHF 17,820, 30.17%), followed by friends and family (CHF 9288, 15.73%) and work colleagues (CHF 6028, 10.21%), while 43.89% of funds came from donors who were either anonymous or had an unknown relationship to the crowdfunding team. Patients and their relatives made the largest donations, with a median value of CHF 200 (IQR 90). CONCLUSIONS: Crowdfunding proved to be a successful strategy to fund a neuroscience project and to raise awareness of a specific clinical problem. Focusing on potential donors with a personal interest in the issue, such as patients and their relatives in our project, is likely to increase funding success. Compared with traditional grant applications, new skills are needed to explain medical challenges to the crowd through video messages and social media.

Chemo- and optogenetic activation of hypothalamic Foxb1-expressing neurons and their terminal endings in the rostral-dorsolateral PAG leads to tachypnea, bradycardia, and immobility.

Foxb1 -expressing neurons occur in the dorsal premammillary nucleus (PMd) and further rostrally in the parvafox nucleus, a longitudinal cluster of neurons in the lateral hypothalamus of rodents. The descending projection of these Foxb1+ neurons end in the dorsolateral part of the periaqueductal gray (dlPAG). The functional role of the Foxb1+ neuronal subpopulation in the PMd and the parvafox nucleus remains elusive. In this study, the activity of the Foxb1+ neurons and of their terminal endings in the dlPAG in mice was selectively altered by employing chemo- and optogenetic tools. Our results show that in whole-body barometric plethysmography, hM3Dq-mediated, global Foxb1+ neuron excitation activates respiration. Time-resolved optogenetic gain-of-function manipulation of the terminal endings of Foxb1+ neurons in the rostral third of the dlPAG leads to abrupt immobility and bradycardia. Chemogenetic activation of Foxb1+ cell bodies and ChR2-mediated excitation of their axonal endings in the dlPAG led to a phenotypical presentation congruent with a 'freezing-like' situation during innate defensive behavior.

Eliminating the VGlut2-Dependent Glutamatergic Transmission of Parvalbumin-Expressing Neurons Leads to Deficits in Locomotion and Vocalization, Decreased Pain Sensitivity, and Increased Dominance.

The calcium-binding protein parvalbumin (PV) is a recognized marker of short-axon GABA-ergic neurons in the cortex and the hippocampus. However in addition, PV is expressed by excitatory, glutamatergic neurons in various areas of the brain and spinal cord. Depending on the location of these neurons, loading of their synaptic vesicles with glutamate is mediated by either of three vesicular glutamate transporters (VGlut): VGlut1, VGlut2, or VGlut3. Driven by our interest in one of these glutamatergic/PV-expressing cell clusters-the lateral hypothalamic parvafox nucleus-we investigated the functions of this population of neurons by the selective deletion of VGlut2 expression in PV-expressing cells according to the Cre/Lox-approach. PV-Cre;VGlut2-Lox mutant mice are phenotypically characterized by deficits in locomotion and vocalization, by a decreased thermal nociception, and by an increased social dominance. We conducted a search of the Allen Brain Atlas for regions that might co-express the genes encoding PV and VGlut2, and that might thus contribute to the manifestation of the observed phenotypes. Our survey revealed several structures that could contribute to the deficits in locomotion and vocalization, such as the red, the subthalamic and the deep cerebellar nuclei. It also disclosed that a shift in the balance of afferental glutamatergic neurotransmission to the periaqueductal gray matter might be accountable for the decrease in sensitivity to pain and for the increase in social dominance. As a whole, this study broadens the state of knowledge about PV-expressing excitatory neurons.

Reduction in 50-kHz call-numbers and suppression of tickling-associated positive affective behaviour after lesioning of the lateral hypothalamic parvafox nucleus in rats.

The parvafox nucleus is located ventrolaterally in the lateral hypothalamic area (LHA). Its core and shell are composed of neurons expressing the calcium-binding protein parvalbumin (PV) and the transcription factor Foxb1, respectively. Given the known functions of the LHA and that the parvafox nucleus receives afferents from the lateral orbitofrontal cortex and projects to the periaqueductal gray matter, a functional role of this entity in the expression of positive emotions has been postulated. The purpose of the present study was to ascertain whether the deletion of neurons in the parvafox nucleus influenced the tickling-induced 50-kHz calls, which are thought to reflect positive affective states, in rats. To this end, tickling of the animals (heterospecific play) was combined with intracerebral injections of the excitotoxin kainic acid into the parvafox nucleus. The most pronounced surgery-associated reduction in 50-kHz call-numbers was observed in the group of rats in which, on the basis of PV-immunoreactive-cell counts in the parvafox nucleus, bilateral lesions had been successfully produced. Two other parameters that were implemented to quantify positive affective behaviour, namely, an approach towards and a following of the hand of the tickling experimenter, were likewise most markedly suppressed in the group of rats with bilaterally successful lesions. Furthermore, positive correlations were found between each of the investigated parameters. Our data afford evidence that the parvafox nucleus plays a role in the production of 50-kHz calls in rats, and, more generally, in the expression of positive emotions.