Humans' voice offers the widest variety of motor phenomena of any human activity. However, its clinical evaluation in people with movement disorders such as Parkinson's disease lags behind current knowledge on advanced analytical automatic speech processing methodology. Here, we use deep learning-based speech processing to differentially analyze voice recordings in 14 people with PD before and after dopaminergic medication using personalized Convolutional Recurrent Neural Networks and Phone Attribute Codebooks. p-CRNN yields an accuracy of 82.35% in the binary classification of ON (...) and OFF motor states at a sensitivity/specificity of 0.86/0.78. The PAC-based approach's accuracy was slightly lower with 73.08% at a sensitivity/specificity of 0.69/0.77, but this method offers easier interpretation and understanding of the computational biomarkers. Both p-CRNN and PAC provide a differentiated view and novel insights into the distinctive components of the speech of persons with PD. Both methods detect voice qualities that are amenable to dopaminergic treatment, including active phonetic and prosodic features. Our findings may pave the way for quantitative measurements of speech in persons with PD. (shrink)

Complex‐trait genetics has advanced dramatically through methods to estimate the heritability tagged by SNPs, both genome‐wide and in genomic regions of interest such as those defined by functional annotations. The models underlying many of these analyses are inadequate, and consequently many SNP‐heritability results published to date are inaccurate. Here, we review the modelling issues, both for analyses based on individual genotype data and association test statistics, highlighting the role of a low‐dimensional model for the heritability of each SNP. We use (...) state‐of‐art models to present updated results about how heritability is distributed with respect to functional annotations in the human genome, and how it varies with allele frequency, which can reflect purifying selection. Our results give finer detail to the picture that has emerged in recent years of complex trait heritability widely dispersed across the genome. Confounding due to population structure remains a problem that summary statistic analyses cannot reliably overcome. Also see the video abstract here: https://youtu.be/WC2u03V65MQ. (shrink)

In 2018 and 2019 Cancer Research UK launched a controversial advertising campaign to inform the British public of obesity being a preventable cause of cancer. On each occasion the advertisements used were emotive and provoked frustration among the British public which was widely vocalised on social media. As well serving to educate the public of this association, the advertisements also had the secondary effect of acting as health promotion through social marketing, a form of advertising designed to influence behavioural changes. (...) As CRUK delivered a public health message through its campaign, the advertisements should be held according to the ethical principles which underpin healthcare in the UK. This article evaluates whether the advertisements used by CRUK in 2018 and 2019 fulfilled the ethical principles of beneficence, autonomy, non-maleficence and justice. It is found that while providing an important message, the oversimplification of obesity as being the result of personal decisions ignored the complex aetiology and served to stigmatise the target demographic, potentially disengaging them from the message. Additionally, posting cancer as the consequence of obesity invokes feelings of fear due to its connotations of suffering and premature death. Based on available evidence, the use of fear in social marketing does not create sustained behavioural change. This essay recommends that CRUK discontinue its use of such strategies in its future social marketing endeavours. (shrink)