Automated Motor Tic Detection: A Machine Learning Approach.

BACKGROUND: Video-based tic detection and scoring is useful to independently and objectively assess tic frequency and severity in patients with Tourette syndrome. In trained raters, interrater reliability is good. However, video ratings are time-consuming and cumbersome, particularly in large-scale studies. Therefore, we developed two machine learning (ML) algorithms for automatic tic detection. OBJECTIVE: The aim of this study was to evaluate the performances of state-of-the-art ML approaches for automatic video-based tic detection in patients with Tourette syndrome. METHODS: We used 64 videos of n = 35 patients with Tourette syndrome. The data of six subjects (15 videos with ratings) were used as a validation set for hyperparameter optimization. For the binary classification task to distinguish between tic and no-tic segments, we established two different supervised learning approaches. First, we manually extracted features based on landmarks, which served as input for a Random Forest classifier (Random Forest). Second, a fully automated deep learning approach was used, where regions of interest in video snippets were input to a convolutional neural network (deep neural network). RESULTS: Tic detection F1 scores (and accuracy) were 82.0% (88.4%) in the Random Forest and 79.5% (88.5%) in the deep neural network approach. CONCLUSIONS: ML algorithms for automatic tic detection based on video recordings are feasible and reliable and could thus become a valuable assessment tool, for example, for objective tic measurements in clinical trials. ML algorithms might also be useful for the differential diagnosis of tics. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Routine stream monitoring data enables the unravelling of hydrological pathways and transfers of agricultural contaminants through catchments.

Catchment-scale understanding of water and contaminant fluxes through all pathways is essential to address land use and climate change impacts on freshwater. However, few options exist to obtain this understanding for the many catchments worldwide for which streamflow and low-frequency water chemistry, but little other data exists. We applied the Bayesian chemistry-assisted hydrograph separation and load partitioning model (BACH) to 47 catchments with widely differing characteristics. As BACH relies on concentration differences between pathways, chemodynamic behaviour of a water constituent indicates its likely suitability as tracer. Typical tracers (e.g. silica, chloride) were unavailable, but Electrical Conductivity and a few monitored nutrients proved chemodynamic in most catchments. Using one of two tracer combinations (Total Nitrogen + Electrical Conductivity, Total Nitrogen + Total Phosphorus) allowed in 85 % of the catchments to estimate streamflow contributions by near-surface (NS), shallow groundwater (SGW), and deep groundwater (DGW) pathways and pathway-specific tracer concentrations and yields with acceptable confidence. In 46 catchments, at least two pathways contributed ≥20 % of the streamflow, and all three ≥20 % in 12 catchments, cautioning against the notion of a single 'dominant' pathway. In contrast to hydrometric hydrograph separation, BACH allows differentiation between 'young' (NS + SGW) and 'old' (DGW) water, which is crucial for the understanding of pollution in catchments with strong temporal gradients in land use intensity. Consistent with generally increasing land use intensity, and groundwater denitrification occurring in some catchments, Total Nitrogen (TN) concentrations were in most catchments higher in NS and SGW compared to DGW. In most catchments, the greatest fraction of the TN yield was conveyed by SGW (≈ 40-90 %). Exceptions were wet and hilly catchments under bush, where the NS transferred most of the very low yields, and three young volcanic catchments where the DGW transferred the majority of the yield due to particularly high DGW flow contributions.

Nitrification inhibitor chlorate and nitrogen substrates differentially affect comammox Nitrospira in a grassland soil.

Introduction: Through the combined use of two nitrification inhibitors, Dicyandiamide (DCD) and chlorate with nitrogen amendment, this study aimed to investigate the contribution of comammox Nitrospira clade B, ammonia oxidizing bacteria (AOB) and archaea (AOA) to nitrification in a high fertility grassland soil, in a 90-day incubation study. Methods: The soil was treated with nitrogen (N) at three levels: 0 mg-N kg-1 soil, 50 mg-N kg-1 soil, and 700 mg-N kg-1 soil, with or without the two nitrification inhibitors. The abundance of comammox Nitrospira, AOA, AOB, and nitrite oxidising bacteria (NOB) was measured using qPCR. The comammox Nitrospira community structure was assessed using Illumina sequencing. Results and Discussion: The results showed that the application of chlorate inhibited the oxidation of both NH4+ and NO2- in all three nitrogen treatments. The application of chlorate significantly reduced the abundance of comammox Nitrospira amoA and nxrB genes across the 90-day experimental period. Chlorate also had a significant effect on the beta diversity (Bray-Curtis dissimilarity) of the comammox Nitrospira clade B community. Whilst AOB grew in response to the N substrate additions and were inhibited by both inhibitors, AOA showed litle or no response to either the N substrate or inhibitor treatments. In contrast, comammox Nitrospira clade B were inhibited by the high ammonium concentrations released from the urine substrates. These results demonstrate the differential and niche responses of the three ammonia oxidising communities to N substrate additions and nitrification inhibitor treatments. Further research is needed to investigate the specificity of the two inhibitors on the different ammonia oxidising communities.

Achieving unbiased predictions of national-scale groundwater redox conditions via data oversampling and statistical learning.

An important policy consideration for integrated land and water management is to understand the spatial distribution of nitrate attenuation in the groundwater system, for which redox condition is the key indicator. This paper proposes a methodology to accommodate the computational demands of large datasets, and presents national-scale predictions of groundwater redox class for New Zealand. Our approach applies statistical learning methods to relate the redox class determined on groundwater samples to spatially varying attributes. The trained model uses these spatial variables to predict redox status in areas without sample data. We assembled the groundwater sample data from regional authority databases, and assigned each sample a redox class. A key achievement was to overcome the influence of sample selection bias on model training via oversampling. We removed additional bias imposed by imbalances in the predictor variables by applying a conditional inference random forest classifier. The unbiased trained model uses eight predictors, and achieves a high validation performance (accuracy 0.81, kappa 0.71), providing good confidence in model predictions. National maps are provided for redox class and probability at specified depths. Feature importance rankings indicate that reducing conditions are associated with poorly-drained soils, and to a lesser extent, high hydrological variability, low elevation, and low-permeability lithology. These conditions are common in New Zealand's coastal and lowland plains, where artificial drainage is required to make land suitable for production. The spatial extent of reduced groundwater increases with depth, suggesting a shallow influence of soil infiltration or mobile organic carbon, and a deeper influence of lithological electron donors. Our model provides unbiased predictions at a scale relevant for environmental policy development and legislation. Identifying where the ecosystem service provided by denitrification can be utilised will enable spatially targeted interventions that can achieve the desired environmental outcome in a more cost-effective manner than non-targeted interventions.

Controlled Spacing between Nanopatterned Regions in Block Copolymer Films Obtained by Utilizing Substrate Topography for Local Film Thickness Differentiation.

Various types of devices require hierarchically nanopatterned substrates, where the spacing between patterned domains is controlled. Ultraconfined films exhibit extreme morphological sensitivity to slight variations in film thickness when the substrate is highly selective toward one of the blocks. Here, it is shown that using the substrate's topography as a thickness differentiating tool enables the creation of domains with different surface patterns in a fully controlled fashion from a single, unblended block copolymer. This approach is applicable to block copolymers of different compositions and to different topographical patterns and thus opens numerous possibilities for the hierarchical construction of multifunctional devices.