Dataset Information

Can education be personalised using pupils' genetic data?

ABSTRACT: The increasing predictive power of polygenic scores for education has led to their promotion by some as a potential tool for genetically informed policy. How accurately polygenic scores predict an individual pupil's educational performance conditional on other phenotypic data is however not well understood. Using data from a UK cohort study with data linkage to national schooling records, we investigated how accurately polygenic scores for education predicted pupils' test score achievement. We also assessed the performance of polygenic scores over and above phenotypic data that are available to schools. Across our sample, there was high overlap between the polygenic score and achievement distributions, leading to poor predictive accuracy at the individual level. Prediction of educational outcomes from polygenic scores were inferior to those from parental socioeconomic factors. Conditional on prior achievement, polygenic scores failed to accurately predict later achievement. Our results suggest that while polygenic scores can be informative for identifying group level differences, they currently have limited use for accurately predicting individual educational performance or for personalised education.

SUBMITTER: Morris TT

PROVIDER: S-EPMC7064332 | biostudies-literature | 2020 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Can education be personalised using pupils' genetic data?

Morris Tim T TT Davies Neil M NM Davey Smith George G

eLife 20200310

The increasing predictive power of polygenic scores for education has led to their promotion by some as a potential tool for genetically informed policy. How accurately polygenic scores predict an individual pupil's educational performance conditional on other phenotypic data is however not well understood. Using data from a UK cohort study with data linkage to national schooling records, we investigated how accurately polygenic scores for education predicted pupils' test score achievement. We a ...[more]

PMID: 32151313

Similar Datasets

Project description:IntroductionDigital health interventions are an effective way to treat depression, but it is still largely unclear how patients' individual symptoms evolve dynamically during such treatments. Data-driven forecasts of depressive symptoms would allow to greatly improve the personalisation of treatments. In current forecasting approaches, models are often trained on an entire population, resulting in a general model that works overall, but does not translate well to each individual in clinically heterogeneous, real-world populations. Model fairness across patient subgroups is also frequently overlooked. Personalised models tailored to the individual patient may therefore be promising.MethodsWe investigate different personalisation strategies using transfer learning, subgroup models, as well as subject-dependent standardisation on a newly-collected, longitudinal dataset of depression patients undergoing treatment with a digital intervention ( N=65 patients recruited). Both passive mobile sensor data as well as ecological momentary assessments were available for modelling. We evaluated the models' ability to predict symptoms of depression (Patient Health Questionnaire-2; PHQ-2) at the end of each day, and to forecast symptoms of the next day.ResultsIn our experiments, we achieve a best mean-absolute-error (MAE) of 0.801 (25% improvement) for predicting PHQ-2 values at the end of the day with subject-dependent standardisation compared to a non-personalised baseline ( MAE=1.062 ). For one day ahead-forecasting, we can improve the baseline of 1.539 by 12% to a MAE of 1.349 using a transfer learning approach with shared common layers. In addition, personalisation leads to fairer models at group-level.DiscussionOur results suggest that personalisation using subject-dependent standardisation and transfer learning can improve predictions and forecasts, respectively, of depressive symptoms in participants of a digital depression intervention. We discuss technical and clinical limitations of this approach, avenues for future investigations, and how personalised machine learning architectures may be implemented to improve existing digital interventions for depression.

Project description:While earlier research in human-robot interaction pre-dominantly uses rule-based architectures for natural language interaction, these approaches are not flexible enough for long-term interactions in the real world due to the large variation in user utterances. In contrast, data-driven approaches map the user input to the agent output directly, hence, provide more flexibility with these variations without requiring any set of rules. However, data-driven approaches are generally applied to single dialogue exchanges with a user and do not build up a memory over long-term conversation with different users, whereas long-term interactions require remembering users and their preferences incrementally and continuously and recalling previous interactions with users to adapt and personalise the interactions, known as the lifelong learning problem. In addition, it is desirable to learn user preferences from a few samples of interactions (i.e., few-shot learning). These are known to be challenging problems in machine learning, while they are trivial for rule-based approaches, creating a trade-off between flexibility and robustness. Correspondingly, in this work, we present the text-based Barista Datasets generated to evaluate the potential of data-driven approaches in generic and personalised long-term human-robot interactions with simulated real-world problems, such as recognition errors, incorrect recalls and changes to the user preferences. Based on these datasets, we explore the performance and the underlying inaccuracies of the state-of-the-art data-driven dialogue models that are strong baselines in other domains of personalisation in single interactions, namely Supervised Embeddings, Sequence-to-Sequence, End-to-End Memory Network, Key-Value Memory Network, and Generative Profile Memory Network. The experiments show that while data-driven approaches are suitable for generic task-oriented dialogue and real-time interactions, no model performs sufficiently well to be deployed in personalised long-term interactions in the real world, because of their inability to learn and use new identities, and their poor performance in recalling user-related data.

Project description:BackgroundEducation can create better opportunities for health, and vice versa. Using a so-called 'add-in' approach, school-based physical activity (PA) promotion and prevention of sedentary behaviours can increase pupils' wellbeing and learning and, on the longer term, reduce the risk of non-communicable diseases. A PA 'add-in' approach involves integrating PA into teachers' curricular obligations without being an extra burden as opposed to an 'add-on' approach which requires additional operational resources and include activities that do not explicitly contribute towards curricular targets making them less long-term acceptable in a school-based context. Previous studies investigating education outside the classroom (EOtC) show mutual benefits for both health and education outcomes among children and adolescents. However, the evidence is of mixed quality and questionable certainty, which calls for further investigation. The aim of this study protocol is to describe and discuss the study design and methods to investigate the efficacy and mechanisms of EOtC as a vehicle for health and education. The study investigates the intervention developed and conducted in the TEACHOUT study with updated and strengthened design and measures.MethodsThe efficacy of EOtC will be investigated in a cluster randomised waitlist design. Participants will be pupils in ~54 classes, grades 4-10 (ages 10-15 years) in ~30 Danish elementary schools. Fifteen schools will be randomised to the intervention: a two-day EOtC training course targeting teachers followed by the teachers implementing EOtC >5 hours weekly over the course of one school year. Pre- and post-measures of health (PA and wellbeing) and learning (school motivation and academic achievement) will be collected. Investigation of pedagogical and motivational mechanisms will be based on observations of EOtC.DiscussionThe updated randomised controlled design will provide firmer evidence for the efficacy and mechanisms of EOtC and provide knowledge about how mutual benefits of health and education can be obtained.Trial registrationRegistered with ClinicalTrials.gov (ID NCT05237674 ) [University of Copenhagen. MOVEOUT: a Cluster RCT of the Efficacy, Mechanisms, and Mediation of an Education Outside the Classroom Intervention on Adolescents' Physical Activity, 2023], February 14, 2022. Most recently updated on November 23, 2022 (Version 2).

Project description:Beyond subject matter learning, primary science teaching aims at encouraging positive attitudes toward, and lasting interest in, science. This study tested whether the learning and science commitment of 8- to 10-year old pupils was affected by extracurricular learning opportunities involving repeated interactions with free-living northern bald ibis (Geronticus eremita; an endangered bird species) and professional researchers. To examine the project's efficacy, pupils' learning progress was monitored by repeated measures of knowledge (i) about northern bald ibises in specific, and (ii) of bird diversity in general. In addition, the children's attitudes toward science were monitored and their dynamic mental models of northern bald ibis' morphology and size were assessed from drawings. A total of 55 pupils from two schools were tested for eight months (before, during and after the experience). Control groups went through regular (curricular) science lessons with similar content and time investment. The extracurricular experience produced a clear learning progress with large effect sizes, which was particularly evident on a long timescale. In addition to subject matter knowledge, the project group pupils could name more bird species and expressed their views about the importance of scientific research for society in a higher proportion than control groups. Differences between schools suggest that project participation also changed the teacher's own interest in northern bald ibises, which affected learning in the control group taught by the same teacher. Beyond the pupils' language and reading skills, learning progress was also visible by scoring ibis-typical features in drawings; their mental models of relative size were not changed due to project participation, however. The results are discussed in reference to inducing and maintaining pupils' attitude and interest in a topic. The study adds empirical evidence for the potentials of advancing primary science education e. g. by fostering the collaboration of schools with scientists.

Dataset Information

Can education be personalised using pupils' genetic data?

Publications

Can education be personalised using pupils' genetic data?

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets