Background: Tinnitus is experienced by up to 15% of the population and can lead to significant disability and distress. There is rarely a medical or surgical target and psychological therapies are recommended. We investigated whether mindfulness-based cognitive therapy (MBCT) could offer an effective new therapy for tinnitus. Methods: This single-site randomized controlled trial compared MBCT to intensive relaxation training (RT) for chronic, distressing tinnitus in adults. Both treatments involved 8 weekly, 120-min sessions focused on either relaxation (RT) or mindfulness meditation (MBCT). Assessments were completed at baseline and at treatment commencement 8 weeks later. The primary outcomes were tinnitus severity (Tinnitus Questionnaire) and psychological distress (Clinical Outcomes in Routine Evaluation - Non-Risk, CORE-NR), 16 weeks after baseline. The analysis utilized a modified intention-to-treat approach. Results: A total of 75 patients were randomly allocated to MBCT (n = 39) or RT (n = 36). Both groups showed significant reductions in tinnitus severity and loudness, psychological distress, anxiety, depression, and disability. MBCT led to a significantly greater reduction in tinnitus severity than RT, with a mean difference of 6.3 (95% CI 1.3-11.4, p = 0.016). Effects persisted 6 months later, with a mean difference of 7.2 (95% CI 2.1-2.3, p = 0.006) and a standardized effect size of 0.56 (95% CI 0.16-0.96). Treatment was effective regardless of initial tinnitus severity, duration, or hearing loss. Conclusions: MBCT is effective in reducing tinnitus severity in chronic tinnitus patients compared to intensive RT. It also reduces psychological distress and disability. Future studies should explore the generalizability of this approach and how outcome relates to different aspects of the intervention.

No abstract

BackgroundA significant proportion of patients with chronic tinnitus report clinical levels of sleep disturbance (insomnia). Despite the significant health and functioning implications of this, no rigorous trials have investigated treatments that target tinnitus-related insomnia. This is the first randomised controlled trial evaluating Cognitive Behavioural Therapy for insomnia (CBTi) in tinnitus compared with other psychological treatments.Methods/designThe study will test the efficacy of group CBTi as a treatment for tinnitus-related insomnia in a single-centre randomised controlled trial. Participants will be 102 patients with chronic, clinically significant tinnitus and insomnia in the absence of organic sleep disorders. Participants will be randomised to one of three intervention arms: six sessions of CBTi or six sessions of sleep support group or two sessions of audiologically based care. The primary outcomes will be changes in sleep as measured on the Insomnia Severity Index and key outcomes on a 2-week sleep diary (sleep efficiency and total sleep time). Outcomes will be collected 3, 10, 14 and 34 weeks post-randomisation. Secondary measures include sleep quality, sleep beliefs, tinnitus severity, psychological distress and quality of life. A sub-sample of participants will provide two weeks of actigraphy data at the same time points. Data on satisfaction and treatment experience will be collected at 10 and 34 weeks post-randomisation from all participants.DiscussionFindings from the study will be submitted to a peer-reviewed journal. It is anticipated that findings may inform future clinical practice in the treatment of tinnitus-related insomnia.Trial registrationClinicalTrials.gov, NCT03386123. Retrospectively registered on 29 December 2017.

Deep learning methods have recently shown great success in numerous fields, including finance, healthcare, linguistics, robotics and even cybersports. Unsupervised learning methods identify the dominant patterns of variability that shape a data set. Such patterns may correspond to well-understood processes, previously unknown clusters or anomalies. This paper presents a case study where a state-of-the-art family of unsupervised deep learning models called variational autoencoder (VAE) is applied to data accrued from a network of fibre-optic sensors installed within a composite steel–concrete half-through railway bridge. The goals were (a) to characterise automatically the behaviour of the bridge based on sensor measurements and, (b) based on this characterisation, to determine when a train passes across a bridge. Based on the VAE model, an algorithm is presented to identify automatically the ‘train event’ points in an unsupervised setting. Two architectures for the VAE model are compared with commonly used baselines. The architecture tailored for modelling sequential data is shown to outperform other methods considered, on both seen and unseen data. No special hyperparameter optimisation is required. This study illustrates how state-of-the-art deep learning methods can be applied to a civil infrastructure engineering problem without directly modelling the physics of the objects or performing tedious hyperparameter optimisation.

No abstract