Neurodesk: an accessible, flexible and portable data analysis environment for reproducible neuroimaging.

Neuroimaging research requires purpose-built analysis software, which is challenging to install and may produce different results across computing environments. The community-oriented, open-source Neurodesk platform ( https://www.neurodesk.org/ ) harnesses a comprehensive and growing suite of neuroimaging software containers. Neurodesk includes a browser-accessible virtual desktop, command-line interface and computational notebook compatibility, allowing for accessible, flexible, portable and fully reproducible neuroimaging analysis on personal workstations, high-performance computers and the cloud.

Open and reusable deep learning for pathology with WSInfer and QuPath.

Digital pathology has seen a proliferation of deep learning models in recent years, but many models are not readily reusable. To address this challenge, we developed WSInfer: an open-source software ecosystem designed to streamline the sharing and reuse of deep learning models for digital pathology. The increased access to trained models can augment research on the diagnostic, prognostic, and predictive capabilities of digital pathology.

ChampKit: A framework for rapid evaluation of deep neural networks for patch-based histopathology classification.

BACKGROUND AND OBJECTIVE: Histopathology is the gold standard for diagnosis of many cancers. Recent advances in computer vision, specifically deep learning, have facilitated the analysis of histopathology images for many tasks, including the detection of immune cells and microsatellite instability. However, it remains difficult to identify optimal models and training configurations for different histopathology classification tasks due to the abundance of available architectures and the lack of systematic evaluations. Our objective in this work is to present a software tool that addresses this need and enables robust, systematic evaluation of neural network models for patch classification in histology in a light-weight, easy-to-use package for both algorithm developers and biomedical researchers. METHODS: Here we present ChampKit (Comprehensive Histopathology Assessment of Model Predictions toolKit): an extensible, fully reproducible evaluation toolkit that is a one-stop-shop to train and evaluate deep neural networks for patch classification. ChampKit curates a broad range of public datasets. It enables training and evaluation of models supported by timm directly from the command line, without the need for users to write any code. External models are enabled through a straightforward API and minimal coding. As a result, Champkit facilitates the evaluation of existing and new models and deep learning architectures on pathology datasets, making it more accessible to the broader scientific community. To demonstrate the utility of ChampKit, we establish baseline performance for a subset of possible models that could be employed with ChampKit, focusing on several popular deep learning models, namely ResNet18, ResNet50, and R26-ViT, a hybrid vision transformer. In addition, we compare each model trained either from random weight initialization or with transfer learning from ImageNet pretrained models. For ResNet18, we also consider transfer learning from a self-supervised pretrained model. RESULTS: The main result of this paper is the ChampKit software. Using ChampKit, we were able to systemically evaluate multiple neural networks across six datasets. We observed mixed results when evaluating the benefits of pretraining versus random intialization, with no clear benefit except in the low data regime, where transfer learning was found to be beneficial. Surprisingly, we found that transfer learning from self-supervised weights rarely improved performance, which is counter to other areas of computer vision. CONCLUSIONS: Choosing the right model for a given digital pathology dataset is nontrivial. ChampKit provides a valuable tool to fill this gap by enabling the evaluation of hundreds of existing (or user-defined) deep learning models across a variety of pathology tasks. Source code and data for the tool are freely accessible at https://github.com/SBU-BMI/champkit.

Neurodesk: An accessible, flexible, and portable data analysis environment for reproducible neuroimaging.

Neuroimaging data analysis often requires purpose-built software, which can be challenging to install and may produce different results across computing environments. Beyond being a roadblock to neuroscientists, these issues of accessibility and portability can hamper the reproducibility of neuroimaging data analysis pipelines. Here, we introduce the Neurodesk platform, which harnesses software containers to support a comprehensive and growing suite of neuroimaging software (https://www.neurodesk.org/). Neurodesk includes a browser-accessible virtual desktop environment and a command line interface, mediating access to containerized neuroimaging software libraries on various computing platforms, including personal and high-performance computers, cloud computing and Jupyter Notebooks. This community-oriented, open-source platform enables a paradigm shift for neuroimaging data analysis, allowing for accessible, flexible, fully reproducible, and portable data analysis pipelines.

Deep neural networks allow expert-level brain meningioma segmentation and present potential for improvement of clinical practice.

Accurate brain meningioma segmentation and volumetric assessment are critical for serial patient follow-up, surgical planning and monitoring response to treatment. Current gold standard of manual labeling is a time-consuming process, subject to inter-user variability. Fully-automated algorithms for meningioma segmentation have the potential to bring volumetric analysis into clinical and research workflows by increasing accuracy and efficiency, reducing inter-user variability and saving time. Previous research has focused solely on segmentation tasks without assessment of impact and usability of deep learning solutions in clinical practice. Herein, we demonstrate a three-dimensional convolutional neural network (3D-CNN) that performs expert-level, automated meningioma segmentation and volume estimation on MRI scans. A 3D-CNN was initially trained by segmenting entire brain volumes using a dataset of 10,099 healthy brain MRIs. Using transfer learning, the network was then specifically trained on meningioma segmentation using 806 expert-labeled MRIs. The final model achieved a median performance of 88.2% reaching the spectrum of current inter-expert variability (82.6-91.6%). We demonstrate in a simulated clinical scenario that a deep learning approach to meningioma segmentation is feasible, highly accurate and has the potential to improve current clinical practice.

Classification of glioblastoma versus primary central nervous system lymphoma using convolutional neural networks.

A subset of primary central nervous system lymphomas (PCNSL) are difficult to distinguish from glioblastoma multiforme (GBM) on magnetic resonance imaging (MRI). We developed a convolutional neural network (CNN) to distinguish these tumors on contrast-enhanced T1-weighted images. Preoperative brain tumor MRIs were retrospectively collected among 320 patients with either GBM (n = 160) and PCNSL (n = 160) from two academic institutions. The individual images from these MRIs consisted of a training set (n = 1894 GBM and 1245 PCNSL), a validation set (n = 339 GBM; 202 PCNSL), and a testing set (99 GBM and 108 PCNSL). Three CNNs using the EfficientNetB4 architecture were evaluated. To increase the size of the training set and minimize overfitting, random flips and changes to color were performed on the training set. Our transfer learning approach (with image augmentation and 292 epochs) yielded an AUC of 0.94 (95% CI: 0.91-0.97) for GBM and an AUC of 0.95 (95% CI: 0.92-0.98) for PCNL. In the second case (not augmented and 137 epochs), the images were augmented prior to training. The area under the curve for GBM was 0.92 (95% CI: 0.88-0.96) for GBM and an AUC of 0.94 (95% CI: 0.91-0.97) for PCNSL. For the last case (augmented, Gaussian noise and 238 epochs) the AUC for GBM was 0.93 (95% CI: 0.89-0.96) and an AUC 0.93 (95% CI = 0.89-0.96) for PCNSL. Even with a relatively small dataset, our transfer learning approach demonstrated CNNs may provide accurate diagnostic information to assist radiologists in distinguishing PCNSL and GBM.

Resting cerebral oxygen metabolism exhibits archetypal network features.

Standard magnetic resonance imaging approaches offer high-resolution but indirect measures of neural activity, limiting understanding of the physiological processes associated with imaging findings. Here, we used calibrated functional magnetic resonance imaging during the resting state to recover low-frequency fluctuations of the cerebral metabolic rate of oxygen (CMRO2 ). We tested whether functional connections derived from these fluctuations exhibited organization properties similar to those established by previous standard functional and anatomical connectivity studies. Seventeen participants underwent 20 min of resting imaging during dual-echo, pseudocontinuous arterial spin labeling, and blood-oxygen-level dependent (BOLD) signal acquisition. Participants also underwent a 10 min normocapnic and hypercapnic procedure. Brain-wide, CMRO2 low-frequency fluctuations were subjected to graph-based and voxel-wise functional connectivity analyses. Results demonstrated that connections derived from resting CMRO2 fluctuations exhibited complex, small-world topological properties (i.e., high integration and segregation, cost efficiency) consistent with those observed in previous studies using functional and anatomical connectivity approaches. Voxel-wise CMRO2 connectivity also exhibited spatial patterns consistent with four targeted resting-state subnetworks: two association (i.e., frontoparietal and default mode) and two perceptual (i.e., auditory and occipital-visual). These are the first findings to support the use of calibration-derived CMRO2 low-frequency fluctuations for detecting brain-wide organizational properties typical of healthy participants. We discuss interpretations, advantages, and challenges in using calibration-derived oxygen metabolism signals for examining the intrinsic organization of the human brain.

Fixed and flexible: Dynamic prefrontal activations and working memory capacity relationships vary with memory demand.

Prefrontal cortex (PFC) activation during encoding of memoranda (proactive responses) is associated with better working memory (WM) compared to reactive/retrieval-based activation. This suggests that dynamic PFC activation patterns may be fixed, based upon one's WM ability, with individuals who have greater WM ability relying more on proactive processes and individuals with lesser WM ability relying more on reactive processes. We newly tested whether this heuristic applied when challenging an individual's WM capacity. Twenty-two participants (N = 22) underwent functional near-infrared spectroscopy (fNIRS) during a modified Sternberg WM paradigm. We tested whether the relationship between dynamic PFC activation patterns and WM capacity changed, as a function of WM demands (N = 14 after quality control). Here, higher-WM capacity was associated with more proactive PFC patterns, but only when WM capacity was overloaded. Lower-WM capacity was associated with these same patterns, but only when WM demand was low. Findings are inconsistent with a purely fixed view of dynamic PFC activation patterns and suggest higher- and lower-WM-capacity individuals flexibly engage PFC processes in a fundamentally different manner, dependent upon current WM demands.

Knowing What You Know in Brain Segmentation Using Bayesian Deep Neural Networks.

In this paper, we describe a Bayesian deep neural network (DNN) for predicting FreeSurfer segmentations of structural MRI volumes, in minutes rather than hours. The network was trained and evaluated on a large dataset (n = 11,480), obtained by combining data from more than a hundred different sites, and also evaluated on another completely held-out dataset (n = 418). The network was trained using a novel spike-and-slab dropout-based variational inference approach. We show that, on these datasets, the proposed Bayesian DNN outperforms previously proposed methods, in terms of the similarity between the segmentation predictions and the FreeSurfer labels, and the usefulness of the estimate uncertainty of these predictions. In particular, we demonstrated that the prediction uncertainty of this network at each voxel is a good indicator of whether the network has made an error and that the uncertainty across the whole brain can predict the manual quality control ratings of a scan. The proposed Bayesian DNN method should be applicable to any new network architecture for addressing the segmentation problem.

Sleep quality, duration, and consistency are associated with better academic performance in college students.

Although numerous survey studies have reported connections between sleep and cognitive function, there remains a lack of quantitative data using objective measures to directly assess the association between sleep and academic performance. In this study, wearable activity trackers were distributed to 100 students in an introductory college chemistry class (88 of whom completed the study), allowing for multiple sleep measures to be correlated with in-class performance on quizzes and midterm examinations. Overall, better quality, longer duration, and greater consistency of sleep correlated with better grades. However, there was no relation between sleep measures on the single night before a test and test performance; instead, sleep duration and quality for the month and the week before a test correlated with better grades. Sleep measures accounted for nearly 25% of the variance in academic performance. These findings provide quantitative, objective evidence that better quality, longer duration, and greater consistency of sleep are strongly associated with better academic performance in college. Gender differences are discussed.

LittleBrain: A gradient-based tool for the topographical interpretation of cerebellar neuroimaging findings.

Gradient-based approaches to brain function have recently unmasked fundamental properties of brain organization. Diffusion map embedding analysis of resting-state fMRI data revealed a primary-to-transmodal axis of cerebral cortical macroscale functional organization. The same method was recently used to analyze resting-state data within the cerebellum, revealing for the first time a sensorimotor-fugal macroscale organization principle of cerebellar function. Cerebellar gradient 1 extended from motor to non-motor task-unfocused (default-mode network) areas, and cerebellar gradient 2 isolated task-focused processing regions. Here we present a freely available and easily accessible tool that applies this new knowledge to the topographical interpretation of cerebellar neuroimaging findings. LittleBrain illustrates the relationship between cerebellar data (e.g., volumetric patient study clusters, task activation maps, etc.) and cerebellar gradients 1 and 2. Specifically, LittleBrain plots all voxels of the cerebellum in a two-dimensional scatterplot, with each axis corresponding to one of the two principal functional gradients of the cerebellum, and indicates the position of cerebellar neuroimaging data within these two dimensions. This novel method of data mapping provides alternative, gradual visualizations that complement discrete parcellation maps of cerebellar functional neuroanatomy. We present application examples to show that LittleBrain can also capture subtle, progressive aspects of cerebellar functional neuroanatomy that would be difficult to visualize using conventional mapping techniques. Download and use instructions can be found at https://xaviergp.github.io/littlebrain.

Enhancing workplace digital learning by use of the science of learning.

Digital learning is becoming the most commonly used portal for workplace learning, but its effectiveness is not clearly understood. We studied 99 employees on-site in a large company as they watched an already used and required training video. Employees were randomly assigned to one of four conditions: (1) a baseline condition of watching the video as in current practice; (2) a spontaneous discussion condition in which participants discussed the video with colleagues immediately after the video without any guidelines; (3) a structured discussion condition in which participants discussed the video with colleagues immediately after the video with an instructor guiding discussion topics; and (4) a testing condition in which test questions were interpolated throughout the video. Memory for the content of the video was measured on a recognition memory test completed 20-35 hours after watching the video. Employees who were in the interpolated-testing or structured discussion conditions had significantly superior memory for the video content (26% and 25% better respectively) relative to typical video viewing; spontaneous discussion did not enhance memory for content. These findings demonstrate that interpolated testing and structured discussion enhance information retention in the workplace and point to how learning science may accelerate workplace learning more generally.

Distributed Weight Consolidation: A Brain Segmentation Case Study.

Collecting the large datasets needed to train deep neural networks can be very difficult, particularly for the many applications for which sharing and pooling data is complicated by practical, ethical, or legal concerns. However, it may be the case that derivative datasets or predictive models developed within individual sites can be shared and combined with fewer restrictions. Training on distributed data and combining the resulting networks is often viewed as continual learning, but these methods require networks to be trained sequentially. In this paper, we introduce distributed weight consolidation (DWC), a continual learning method to consolidate the weights of separate neural networks, each trained on an independent dataset. We evaluated DWC with a brain segmentation case study, where we consolidated dilated convolutional neural networks trained on independent structural magnetic resonance imaging (sMRI) datasets from different sites. We found that DWC led to increased performance on test sets from the different sites, while maintaining generalization performance for a very large and completely independent multi-site dataset, compared to an ensemble baseline.

Disruption of hippocampal synaptic transmission and long-term potentiation by psychoactive synthetic cannabinoid 'Spice' compounds: comparison with Δ9 -tetrahydrocannabinol.

There has been a marked increase in the availability of synthetic drugs designed to mimic the effects of marijuana. These cannabimimetic drugs, sold illicitly as 'Spice' and related products, are associated with serious medical complications in some users. In vitro studies suggest that synthetic cannabinoids in these preparations are potent agonists at central cannabinoid CB1 receptors (CB1Rs), but few investigations have delineated their cellular effects, particularly in comparison with the psychoactive component of marijuana, Δ9 -tetrahydrocannabinol (Δ9 -THC). We compared the ability of three widely abused synthetic cannabinoids and Δ9 -THC to alter glutamate release and long-term potentiation in the mouse hippocampus. JWH-018 was the most potent inhibitor of hippocampal synaptic transmission (EC50 ~15 nM), whereas its fluoropentyl derivative, AM2201, inhibited synaptic transmission with slightly lower potency (EC50 ~60 nM). The newer synthetic cannabinoid, XLR-11, displayed much lower potency (EC50 ~900 nM) that was similar to Δ9 -THC (EC50 ~700 nM). The effects of all compounds occurred via activation of CB1Rs, as demonstrated by reversal with the selective antagonist/inverse agonist AM251 or the neutral CB1R antagonist PIMSR1. Moreover, AM2201 was without effect in the hippocampus of transgenic mice lacking the CB1R. Hippocampal slices exposed to either synthetic cannabinoids or Δ9 -THC exhibited significantly impaired long-term potentiation (LTP). We find that, compared with Δ9 -THC, the first-generation cannabinoids found in Spice preparations display higher potency, whereas a recent synthetic cannabinoid is roughly equipotent with Δ9 -THC. The disruption of synaptic function by these synthetic cannabinoids is likely to lead to profound impairments in cognitive and behavioral function.