Political Attacks in 280 Characters or Less: A New Tool for the Automated Classification of Campaign Negativity on Social Media

Negativity in election campaign matters. To what extent can the content of social media posts provide a reliable indicator of candidates' campaign negativity? We introduce and critically assess an automated classification procedure that we trained to annotate more than 16,000 tweets of candidates competing in the 2018 Senate Midterms. The algorithm is able to identify the presence of political attacks (both in general, and specifically for character and policy attacks) and incivility. Due to the novel nature of the instrument, the article discusses the external and convergent validity of these measures. Results suggest that automated classifications are able to provide reliable measurements of campaign negativity. Triangulations with independent data show that our automatic classification is strongly associated with the experts’ perceptions of the candidates’ campaign. Furthermore, variations in our measures of negativity can be explained by theoretically relevant factors at the candidate and context levels (e.g., incumbency status and candidate gender); theoretically meaningful trends are also found when replicating the analysis using tweets for the 2020 Senate election, coded using the automated classifier developed for 2018. The implications of such results for the automated coding of campaign negativity in social media are discussed.

New method to measure interbreath intervals in infants for the assessment of apnoea and respiration

BackgroundRespiratory disorders, including apnoea, are common in preterm infants due to their immature respiratory control compared with term-born infants. However, our inability to accurately measure respiratory rate in hospitalised infants results in unreported episodes of apnoea and an incomplete picture of respiratory activity.MethodsWe develop, validate and use a novel algorithm to identify interbreath intervals (IBIs) and apnoeas in preterm infants. In 42 preterm infants (1600 hours of recordings), we assess IBIs from the chest electrical impedance pneumograph using an adaptive amplitude threshold for the detection of breaths. The algorithm is refined by comparing its accuracy with clinically observed breaths and pauses in breathing. We develop an automated classifier to differentiate periods of true apnoea from artefactually low amplitude signal. We assess the performance of this algorithm in the detection of morphine-induced respiratory depression. Finally, we use the algorithm to investigate whether retinopathy of prematurity (ROP) screening alters the IBI distribution.ResultsIndividual breaths were detected with a false-positive rate of 13% and a false-negative rate of 12%. The classifier identified true apnoeas with an accuracy of 93%. As expected, morphine caused a significant shift in the IBI distribution towards longer IBIs. Following ROP screening, there was a significant increase in pauses in breathing that lasted more than 10 s (t-statistic=1.82, p=0.023). This was not reflected by changes in the monitor-derived respiratory rate and no episodes of apnoea were recorded in the medical records.ConclusionsWe show that our algorithm offers an improved method for the identification of IBIs and apnoeas in preterm infants. Following ROP screening, increased respiratory instability can occur even in the absence of clinically significant apnoeas. Accurate assessment of infant respiratory activity is essential to inform clinical practice.

Glider Observations of Thermohaline Staircases in the Tropical North Atlantic Using an Automated Classifier

Thermohaline staircases are stepped structures of alternating thick mixed layers and thin high gradient interfaces. These structures can be up to several tens of metres thick and are associated with double-diffusive mixing. Thermohaline staircases occur across broad swathes of the Arctic and tropical/subtropical oceans and can increase rates of diapycnal mixing by up to five times the background rate, driving substantial nutrient fluxes to the upper ocean. In this study, we present an improved classification algorithm to detect thermohaline staircases in ocean glider profiles. We use a dataset of 1162 glider profiles from the tropical North Atlantic collected in early 2020 at the edge of a known thermohaline staircase region. The algorithm identifies thermohaline staircases in 97.7 % of profiles that extend deeper than 300 m. We validate our algorithm against previous results obtained from algorithmic classification of Argo float profiles. Using fine resolution temperature data from a fast-response thermistor on one of the gliders, we explore the effect of varying vertical bin sizes on detected thermohaline staircases. Our algorithm builds on previous work with improved flexibility and the ability to classify staircases from profiles with poor salinity data. Using our results, we propose that the incidence of thermohaline staircases is limited by strong background vertical gradients in conservative temperature and absolute salinity.

An artificial neural network for automated behavioral state classification in rats

Accurate behavioral state classification is critical for many research applications. Researchers typically rely upon manual identification of behavioral state through visual inspection of electrophysiological signals, but this approach is time intensive and subject to low inter-rater reliability. To overcome these limitations, a diverse set of algorithmic approaches have been put forth to automate the classification process. Recently, novel machine learning approaches have been detailed that produce rapid and highly accurate classifications. These approaches however, are often computationally expensive, require significant expertise to implement, and/or require proprietary software that limits broader adoption. Here we detail a novel artificial neural network that uses electrophysiological features to automatically classify behavioral state in rats with high accuracy, sensitivity, and specificity. Common parameters of interest to sleep scientists, including state-dependent power spectra and homeostatic non-REM slow wave activity, did not significantly differ when using this automated classifier as compared to manual scoring. Flexible options enable researchers to further increase classification accuracy through manual rescoring of a small subset of time intervals with low model prediction certainty or further decrease researcher time by generalizing trained networks across multiple recording days. The algorithm is fully open-source and coded within a popular, and freely available, software platform to increase access to this research tool and provide additional flexibility for future researchers. In sum, we have developed a readily implementable, efficient, and effective approach for automated behavioral state classification in rats.

Facial expressions can detect Parkinson’s disease: preliminary evidence from videos collected online

A prevalent symptom of Parkinson’s disease (PD) is hypomimia — reduced facial expressions. In this paper, we present a method for diagnosing PD that utilizes the study of micro-expressions. We analyzed the facial action units (AU) from 1812 videos of 604 individuals (61 with PD and 543 without PD, with a mean age 63.9 y/o, sd. 7.8) collected online through a web-based tool (www.parktest.net). In these videos, participants were asked to make three facial expressions (a smiling, disgusted, and surprised face) followed by a neutral face. Using techniques from computer vision and machine learning, we objectively measured the variance of the facial muscle movements and used it to distinguish between individuals with and without PD. The prediction accuracy using the facial micro-expressions was comparable to methodologies that utilize motor symptoms. Logistic regression analysis revealed that participants with PD had less variance in AU6 (cheek raiser), AU12 (lip corner puller), and AU4 (brow lowerer) than non-PD individuals. An automated classifier using Support Vector Machine was trained on the variances and achieved 95.6% accuracy. Using facial expressions as a future digital biomarker for PD could be potentially transformative for patients in need of remote diagnoses due to physical separation (e.g., due to COVID) or immobility.

Electrical Brain Stimulation and Continuous Behavioral State Tracking in Ambulatory Humans

Electrical brain stimulation (EBS) is an established treatment for patients with drug-resistant epilepsy. Sleep disorders are common in people with epilepsy and EBS therapies may actually further disturb normal sleep patterns and sleep quality. Novel devices capable of EBS and continuous intracranial EEG (iEEG) telemetry enable detailed assessments of therapy efficacy and tracking of sleep and comorbidities. Here, we investigate the feasibility of automated sleep classification using continuous iEEG data recorded from Papez's circuit in four patients with drug resistant mesial temporal lobe epilepsy using an investigational implantable sensing and stimulation device with electrodes implanted in bilateral hippocampus (HPC) and anterior nucleus of thalamus (ANT). The iEEG recorded from HPC are used to classify sleep during concurent ANT stimulation. Simultaneous polysomnography and HPC sensing was were used to train, validate and test an automated classifier for a range of ANT EBS frequencies (2 Hz, 7Hz, 100Hz, and 145 Hz). We show that it is possible to build a patient specific automated sleep staging classifier using power in band features extracted from one HPC sensing channel. The patient specific classifiers performed well under all thalamic EBS frequencies with an average F1-score 0.894, and provided viable classification into major sleep categories (Awake, NREM, REM). Within this project, we retrospectively analyzed classification performance with gold-standard polysomnography annotations, and then prospectively deployed the classifier on chronic continuous iEEG data spanning multiple months to characterize sleep patterns in ambulatory patients living in their home environment. The ability to continuously track behavioral state and fully characterize sleep should prove useful for optimizing EBS for epilepsy and associated sleep, cognitive and mood comorbidities.

Multiple Object Recognition Using OpenCV

For automatic vision systems used in agriculture, the project presents object characteristics analysis using image processing techniques. In agriculture science, automatic object characteristics identification is important for monitoring vast areas of crops, and it detects signs of object characteristics as soon as it occurs on plant leaves. Image content characterization and supervised classifier type neural network are used in the proposed deciding method. Pre-processing, image segmentation, and detection are some of the image processing methods used in this form of decision making. An image data will be rearranged and, if necessary, a region of interest will be selected during preparation. For network training and classification, colour and texture features are extracted from an input. Colour characteristics such as mean and variance in the HSV colour space, as well as texture characteristics such as energy, contrast, homogeneity, and correlation. The device will be trained to automatically identify test images in order to assess object characteristics. With some training samples of that type, an automated classifier NN could be used for classification supported learning in this method. The tangent sigmoid function is used as the kernel function in this network. Finally, the simulated results show that the used network classifier has a low error rate during training and higher classification accuracy. In the previous researches Object detection has been made possible, but in our current research we have attempted to do live Object Detection using OpenCV and also the techniques involved in it.

A new method to measure inter-breath intervals in infants for the assessment of apnoea and respiratory dynamics

Background Respiratory disorders, including apnoea, are common in preterm infants due to their immature respiratory control and function compared with term-born infants. However, our inability to accurately measure respiratory rate in hospitalised infants results in unreported episodes of apnoea and an incomplete picture of respiratory dynamics. Methods We develop, validate and use a novel algorithm to identify inter-breath intervals (IBIs) and apnoeas in infants. In 42 infants (a total of 1600 hours of recordings) we assess IBIs from the chest electrical impedance pneumograph using an adaptive amplitude threshold for the detection of individual breaths. The algorithm is refined by comparing its accuracy with clinically-observed breaths and pauses in breathing. We also develop an automated classifier to differentiate periods of true central apnoea from artefactually low amplitude signal. We use this algorithm to explore its ability to identify morphine-induced respiratory depression in 15 infants. Finally, in 22 infants we use the algorithm to investigate whether retinopathy of prematurity (ROP) screening alters the IBI distribution. Findings 88% of the central apnoeas identified using our algorithm were missed in the clinical notes. As expected, morphine caused a shift in the IBI distribution towards longer IBIs, with significant differences in all IBI metrics assessed. Following ROP screening, there was a shift in the IBI distribution with a significant increase in the proportion of pauses in breathing that lasted more than 10 seconds (t-statistic=1.82, p=0.023). This was not reflected by changes in the monitor- derived respiratory rate or episodes of apnoea recorded on clinical charts. Interpretation Better measurement of infant respiratory dynamics is essential to improve care for hospitalised infants. Use of the novel IBI algorithm demonstrates that following ROP screening increased instability in respiratory dynamics can be detected in the absence of clinically-significant apnoeas. Funding Wellcome Trust and Royal Society

Naive Bayes classifiers for verbal autopsies: comparison to physician-based classification for 21,000 child and adult deaths

Background Verbal autopsies (VA) are increasingly used in low- and middle-income countries where most causes of death (COD) occur at home without medical attention, and home deaths differ substantially from hospital deaths. Hence, there is no plausible “standard” against which VAs for home deaths may be validated. Previous studies have shown contradictory performance of automated methods compared to physician-based classification of CODs. We sought to compare the performance of the classic naive Bayes classifier (NBC) versus existing automated classifiers, using physician-based classification as the reference. Methods We compared the performance of NBC, an open-source Tariff Method (OTM), and InterVA-4 on three datasets covering about 21,000 child and adult deaths: the ongoing Million Death Study in India, and health and demographic surveillance sites in Agincourt, South Africa and Matlab, Bangladesh. We applied several training and testing splits of the data to quantify the sensitivity and specificity compared to physician coding for individual CODs and to test the cause-specific mortality fractions at the population level. Results The NBC achieved comparable sensitivity (median 0.51, range 0.48-0.58) to OTM (median 0.50, range 0.41-0.51), with InterVA-4 having lower sensitivity (median 0.43, range 0.36-0.47) in all three datasets, across all CODs. Consistency of CODs was comparable for NBC and InterVA-4 but lower for OTM. NBC and OTM achieved better performance when using a local rather than a non-local training dataset. At the population level, NBC scored the highest cause-specific mortality fraction accuracy across the datasets (median 0.88, range 0.87-0.93), followed by InterVA-4 (median 0.66, range 0.62-0.73) and OTM (median 0.57, range 0.42-0.58). Conclusions NBC outperforms current similar COD classifiers at the population level. Nevertheless, no current automated classifier adequately replicates physician classification for individual CODs. There is a need for further research on automated classifiers using local training and test data in diverse settings prior to recommending any replacement of physician-based classification of verbal autopsies.