Wearable Data From Subjects Playing Super Mario, Taking University Exams, or Performing Physical Exercise Help Detect Acute Mood Disorder Episodes via Self-Supervised Learning: Prospective, Exploratory, Observational Study.

BACKGROUND: Personal sensing, leveraging data passively and near-continuously collected with wearables from patients in their ecological environment, is a promising paradigm to monitor mood disorders (MDs), a major determinant of the worldwide disease burden. However, collecting and annotating wearable data is resource intensive. Studies of this kind can thus typically afford to recruit only a few dozen patients. This constitutes one of the major obstacles to applying modern supervised machine learning techniques to MD detection. OBJECTIVE: In this paper, we overcame this data bottleneck and advanced the detection of acute MD episodes from wearables' data on the back of recent advances in self-supervised learning (SSL). This approach leverages unlabeled data to learn representations during pretraining, subsequently exploited for a supervised task. METHODS: We collected open access data sets recording with the Empatica E4 wristband spanning different, unrelated to MD monitoring, personal sensing tasks-from emotion recognition in Super Mario players to stress detection in undergraduates-and devised a preprocessing pipeline performing on-/off-body detection, sleep/wake detection, segmentation, and (optionally) feature extraction. With 161 E4-recorded subjects, we introduced E4SelfLearning, the largest-to-date open access collection, and its preprocessing pipeline. We developed a novel E4-tailored transformer (E4mer) architecture, serving as the blueprint for both SSL and fully supervised learning; we assessed whether and under which conditions self-supervised pretraining led to an improvement over fully supervised baselines (ie, the fully supervised E4mer and pre-deep learning algorithms) in detecting acute MD episodes from recording segments taken in 64 (n=32, 50%, acute, n=32, 50%, stable) patients. RESULTS: SSL significantly outperformed fully supervised pipelines using either our novel E4mer or extreme gradient boosting (XGBoost): n=3353 (81.23%) against n=3110 (75.35%; E4mer) and n=2973 (72.02%; XGBoost) correctly classified recording segments from a total of 4128 segments. SSL performance was strongly associated with the specific surrogate task used for pretraining, as well as with unlabeled data availability. CONCLUSIONS: We showed that SSL, a paradigm where a model is pretrained on unlabeled data with no need for human annotations before deployment on the supervised target task of interest, helps overcome the annotation bottleneck; the choice of the pretraining surrogate task and the size of unlabeled data for pretraining are key determinants of SSL success. We introduced E4mer, which can be used for SSL, and shared the E4SelfLearning collection, along with its preprocessing pipeline, which can foster and expedite future research into SSL for personal sensing.

Sleep-wake variations of electrodermal activity in bipolar disorder.

BACKGROUND: Affective states influence the sympathetic nervous system, inducing variations in electrodermal activity (EDA), however, EDA association with bipolar disorder (BD) remains uncertain in real-world settings due to confounders like physical activity and temperature. We analysed EDA separately during sleep and wakefulness due to varying confounders and potential differences in mood state discrimination capacities. METHODS: We monitored EDA from 102 participants with BD including 35 manic, 29 depressive, 38 euthymic patients, and 38 healthy controls (HC), for 48 h. Fifteen EDA features were inferred by mixed-effect models for repeated measures considering sleep state, group and covariates. RESULTS: Thirteen EDA feature models were significantly influenced by sleep state, notably including phasic peaks (p < 0.001). During wakefulness, phasic peaks showed different values for mania (M [SD] = 6.49 [5.74, 7.23]), euthymia (5.89 [4.83, 6.94]), HC (3.04 [1.65, 4.42]), and depression (3.00 [2.07, 3.92]). Four phasic features during wakefulness better discriminated between HC and mania or euthymia, and between depression and euthymia or mania, compared to sleep. Mixed symptoms, average skin temperature, and anticholinergic medication affected the models, while sex and age did not. CONCLUSION: EDA measured from awake recordings better distinguished between BD states than sleep recordings, when controlled by confounders.

[Social determinants of the incidence of Covid-19 in Barcelona: a preliminary ecological study using public data.] / Determinantes sociales de la incidencia de la Covid-19 en Barcelona: un estudio ecológico preliminar usando datos públicos.

OBJECTIVE: Social determinants and health inequalities have a huge impact on health of populations. It is important to study their role in the management of the Covid-19 epidemic, especially in cities, as certain variables like the number of tests and the access to health system cannot be assumed as equal. The aim of this work was to determine the relation of social determinants in the incidence of Covid-19 in the city of Barcelona. METHODS: An observational retrospective ecological study was performed, with the neighbourhood as the population unit, based on data of cumulative incidence published at May 14th, 2020 by the Public Health Agency of Barcelona. Covid-19 incidence disparities depending on the income of the neighbourhoods, the Pearson linear correlation of the variables selected (age, sex, net density, immigrants, comorbidities, smokers, Body Mass Index [BMI] and Available Income per Family Index [AIFI]) with the incidence and the correlation with a multivariant Generalized Linear Model (GLM) were estimated. RESULTS: It was found that neighbourhoods belonging to the lowest quintile of income had a 42% more incidence than those belonging to the highest quintile: 942 cases per 100,000 inhabitants versus 545 per 100,000 inhabitants of the highest quintile. The Pearson correlation was statistically significative between the incidence of Covid-19 and the percentage of population over 75 (r=0.487), the percentage of immigration of the neighbourhood and the origin of the immigrants (r=-0.257), the AIFI (r=-0.462), the percentage of smokers (r=0.243) and the percentage of people with BMI over 25 (r=0.483). The GLM showed that the most correlated variables with the incidence are the percentage of people over 75 (Z-score=0.258), the percentage of people from Maghreb (Z-score=-0.206) and Latin America (Z-score=0.19) and the percentage of people with BMI over 25 (Z-score=0.334). The results of the GLM were significative. CONCLUSIONS: Social determinants are correlated with the modification of the incidence of Covid-19 in the neighbourhoods of Barcelona, with special relevance of the prevalence of BMI over 25 and the percentage of immigrants and its origin.

OBJETIVO: Los determinantes sociales tienen un gran impacto en la salud de las poblaciones. Es relevante estudiar su papel en la gestión de la epidemia de la Covid-19, especialmente en las ciudades, pues ciertas variables como el número de tests realizados o la disponibilidad de recursos sanitarios no se pueden asumir por igual. El objetivo de este trabajo fue estimar la relación de los determinantes sociales en la incidencia de la Covid-19 en Barcelona. METODOS: Se realizó un estudio ecológico, observacional retrospectivo, con el barrio como unidad de población, basado en los datos publicados a fecha de 14 de mayo de 2020 sobre incidencia acumulada de Covid-19 confirmada por PCR. Se estimó la diferencia de incidencia de la Covid-19 en función de la renta de los barrios, la correlación lineal de Pearson de las distintas variables seleccionadas (edad, sexo, densidad neta, inmigrantes, comorbilidades, tabaquismo, Índice de Masa Corporal [IMC] e Índice de Renta Familiar Disponible [IRFD]) con la incidencia acumulada y se llevó a cabo un análisis multivariante mediante un Modelo Lineal Generalizado (GLM). RESULTADOS: Los barrios del quintil de menor renta presentaban un 42% más de incidencia que aquellos del quintil con más renta: 942 casos por cada 100.000 habitantes frente a los 545 casos por cada 100.000 habitantes. La correlación de Pearson se mostró estadísticamente significativa entre la incidencia de la Covid-19 y el porcentaje de población mayor de 75 años (r=0,487), el porcentaje de inmigrantes (r=-0,257) y el origen de dichos inmigrantes, el IRFD (r=-0,462), el porcentaje de fumadores (r=0,243) y de personas con un IMC mayor de 25 (r=0,483). En GLM las variables que más correlación tenían con la incidencia entre barrios eran el porcentaje de población mayor de 75 años (Z-score=0,258), el porcentaje de inmigrantes latinoamericanos (Z-score=0,19) y magrebíes (Z-score=-0,206), y el porcentaje de personas con IMC>25 (Z-score=0,334). Los resultados del GLM fueron estadísticamente significativos. CONCLUSIONES: Los determinantes sociales se correlacionan con una modificación de la incidencia de la Covid-19 en los barrios de Barcelona, con especial relevancia de la prevalencia de IMC>25 y del porcentaje de inmigrantes y de su origen.

Determinantes sociales de la incidencia de la Covid-19 en Barcelona: un estudio ecológico preliminar usando datos públicos / Social determinants of the incidence of Covid-19 in Barcelona: a preliminary ecological study using public data

OBJETIVO: Los determinantes sociales tienen un gran impacto en la salud de las poblaciones. Es relevante estudiar su papel en la gestión de la epidemia de la Covid-19, especialmente en las ciudades, pues ciertas variables como el número de tests realizados o la disponibilidad de recursos sanitarios no se pueden asumir por igual. El objetivo de este trabajo fue estimar la relación de los determinantes sociales en la incidencia de la Covid-19 en Barcelona. MÉTODOS: Se realizó un estudio ecológico, observacional retrospectivo, con el barrio como unidad de población, basado en los datos publicados a fecha de 14 de mayo de 2020 sobre incidencia acumulada de Covid-19 confirmada por PCR. Se estimó la diferencia de incidencia de la Covid-19 en función de la renta de los barrios, la correlación lineal de Pearson de las distintas variables seleccionadas (edad, sexo, densidad neta, inmigrantes, comorbilidades, tabaquismo, Índice de Masa Corporal [IMC] e Índice de Renta Familiar Disponible [IRFD]) con la incidencia acumulada y se llevó a cabo un análisis multivariante mediante un Modelo Lineal Generalizado (GLM). RESULTADOS: Los barrios del quintil de menor renta presentaban un 42% más de incidencia que aquellos del quintil con más renta: 942 casos por cada 100.000 habitantes frente a los 545 casos por cada 100.000 habitantes. La correlación de Pearson se mostró estadísticamente significativa entre la incidencia de la Covid-19 y el porcentaje de población mayor de 75 años (r=0,487), el porcentaje de inmigrantes (r=-0,257) y el origen de dichos inmigrantes, el IRFD (r=-0,462), el porcentaje de fumadores (r=0,243) y de personas con un IMC mayor de 25 (r=0,483). En GLM las variables que más correlación tenían con la incidencia entre barrios eran el porcentaje de población mayor de 75 años (Z-score=0,258), el porcentaje de inmigrantes latinoamericanos (Z-score=0,19) y magrebíes (Z-score=-0,206), y el porcentaje de personas con IMC>25 (Z-score=0,334). Los resultados del GLM fueron estadísticamente significativos. CONCLUSIONES: Los determinantes sociales se correlacionan con una modificación de la incidencia de la Covid-19 en los barrios de Barcelona, con especial relevancia de la prevalencia de IMC>25 y del porcentaje de inmigrantes y de su origen

OBJECTIVE: Social determinants and health inequalities have a huge impact on health of populations. It is important to study their role in the management of the Covid-19 epidemic, especially in cities, as certain variables like the number of tests and the access to health system cannot be assumed as equal. The aim of this work was to determine the relation of social determinants in the incidence of Covid-19 in the city of Barcelona. METHODS: An observational retrospective ecological study was performed, with the neighbourhood as the population unit, based on data of cumulative incidence published at May 14th, 2020 by the Public Health Agency of Barcelona. Covid-19 incidence disparities depending on the income of the neighbourhoods, the Pearson linear correlation of the variables selected (age, sex, net density, immigrants, comorbidities, smokers, Body Mass Index [BMI] and Available Income per Family Index [AIFI]) with the incidence and the correlation with a multivariant Generalized Linear Model (GLM) were estimated. RESULTS: It was found that neighbourhoods belonging to the lowest quintile of income had a 42% more incidence than those belonging to the highest quintile: 942 cases per 100,000 inhabitants versus 545 per 100,000 inhabitants of the highest quintile. The Pearson correlation was statistically significative between the incidence of Covid-19 and the percentage of population over 75 (r=0.487), the percentage of immigration of the neighbourhood and the origin of the immigrants (r=-0.257), the AIFI (r=-0.462), the percentage of smokers (r=0.243) and the percentage of people with BMI over 25 (r=0.483). The GLM showed that the most correlated variables with the incidence are the percentage of people over 75 (Z-score=0.258), the percentage of people from Maghreb (Z-score=-0.206) and Latin America (Z-score=0.19) and the percentage of people with BMI over 25 (Z-score=0.334). The results of the GLM were significative. CONCLUSIONS: Social determinants are correlated with the modification of the incidence of Covid-19 in the neighbourhoods of Barcelona, with special relevance of the prevalence of BMI over 25 and the percentage of immigrants and its origin