Parkinson's Disease is a progressive neurodegenerative disorder afflicting almost 12 million people. Increased understanding of its complex and heterogenous disease pathology, etiology and symptom manifestations has resulted in the need to design, capture and interrogate substantial clinical datasets. Herein we advocate how advances in the deployment of artificial intelligence models for Federated Data Analysis and Federated Learning can help spearhead coordinated and sustainable approaches to address this grand challenge.

Abstract Introduction Sleep pathology is increasingly being recognized in many neuropsychiatric conditions, including those not traditionally associated with sleep medicine. Improved therapeutic strategies will benefit from EEG-based sleep analytics, but EEG is limited by cost and availability. Here we evaluate the performance and usability of a user operated dry-EEG electrode-based sleep monitoring headband (the Dreem 3S, D3S), which is intended to provide detailed brain EEG activity with automated sleep staging and longitudinal recording ability from the home environment. Methods Two prospective studies included 60 subjects who underwent either one overnight in-lab traditional polysomnography (PSG) while concomitantly wearing the D3S, or three nights at home with the D3S. At home, subjects self-applied the D3S. Device usability and automated sleep staging performance were evaluated. Results Overall agreement between D3S and PSG scoring was 85.6% across sleep stages. Intra Class Correlation (ICC) for Total Sleep Time (TST), Sleep Efficiency (SE), Latency to Persistent Sleep (LPS), and Wake After Sleep Onset (WASO) exceeded 90%, indicating excellent agreement between D3S and PSG. ICC ranged from 65% to 86% for automated machine learning interpretation of time spent in N1, N2, N3, and REM, comparable or superior to other devices and individual human performance. Usability evaluation demonstrated safe and comfortable use of the device in the home setting, with a System Usability Score (SUS) score of ≥ 68. Subjects were able to operate and record 2-3 nights of high-quality data without additional support or technologist intervention. Furthermore, 96.6% of each record was deemed to be of sufficient EEG signal quality to enable manual expert review. Conclusion The Dreem 3S provided accurate sleep-staging and sleep metrics from EEG data acquired in conjunction with in-lab PSG or multiple nights in the home setting. Results were comparable to in-lab PSG results. D3S therefore offers objective EEG-based sleep investigation that is not feasible with traditional PSG, such as longitudinal data acquisition in a patient’s home environment without need for expert technologist applications. Such technology enables expansion of EEG-based sleep analysis for novel conditions. Support (if any) This work was supported by Beacon Biosignals, which is a private company with ownership of the D3S.

Artificial intelligence (AI) methods have been applied to medical imaging for several decades, but in the last few years, the number of publications and the number of AI-enabled medical devices coming on the market have significantly increased. While some AI-enabled approaches are proving very valuable, systematic reviews of the AI imaging field identify significant weaknesses in a significant proportion of the literature. Medical device regulators have recently become more proactive in publishing guidance documents and recognizing standards that will require that the development and validation of AI-enabled medical devices need to be more rigorous than required for tradition "rule-based" software. In particular, developers are required to better identify and mitigate risks (such as bias) that arise in AI-enabled devices, and to ensure that the devices are validated in a realistic clinical setting to ensure their output is clinically meaningful. While this evolving regulatory landscape will mean that device developers will take longer to bring novel AI-based medical imaging devices to market, such additional rigour is necessary to address existing weaknesses in the field and ensure that patients and healthcare professionals can trust AI-enabled devices. There would also be benefits in the academic community taking into account this regulatory framework, to improve the quality of the literature and make it easier for academically developed AI tools to make the transition to medical devices that impact healthcare.

In the field of smart healthcare, wearable and sensing devices are connected to the Internet of Things (IoT) to assess patients within their own homes. Fatigue is a multidimensional experience that can be characterised by exhaustion and reduced physical performance. Monitoring people within their homes to detect slower pace could be a promising method to objectively measure aspects of human fatigue. This paper makes use of both a wearable bracelet and Radio Frequency (RF) sensing to detect simulated fatigue from human activity monitoring. Different activities are collected at a normal pace to represent no fatigue and then repeated at a slower pace to represent fatigue. Artificial intelligence (AI) is used to detect if there is fatigue present or not regardless of activity taking place as well as identifying which activity took place and if fatigue is present in said activity. When using the data from both the bracelet and RF sensing, Random Forest and ResNet algorithms achieved 100 % in detecting fatigue as opposed to non-fatigue using algorithms. When using only the bracelet, only the Random Forest algorithm was able to achieve 100 % accuracy. Using only RF data, 94.80 % accuracy was achieved with a Convolutional Neural Network (CNN). When detecting individual activities with fatigue and no fatigue, the Random Forest algorithm achieved an accuracy score of 97.40 % using both the bracelet and RF sensing and with only the bracelet data. CNN was again the best algorithm for RF sensing only with an accuracy score of 89.84 %.

Introduction Fatigue is prevalent across a wide range of medical conditions and can be debilitating and distressing. It is likely that fatigue is experienced differently according to the underlying aetiology, but this is poorly understood. Digital health technologies present a promising approach to give new insights into fatigue.The aim of this study is to use digital health technologies, real-time self-reports and qualitative interview data to investigate how fatigue is experienced over time in participants with myeloma, long COVID, heart failure and in controls without problematic fatigue. Objectives are to understand which sensed parameters add value to the characterisation of fatigue and to determine whether study processes are feasible, acceptable and scalable.Methods and analysis An ecological momentary assessment study will be carried out over 2 or 4 weeks (participant defined). Individuals with fatigue relating to myeloma (n=10), heart failure (n=10), long COVID (n=10) and controls without problematic fatigue or a study condition (n=10) will be recruited. ECG patches will measure heart rate variability, respiratory rate, body temperature, activity and posture. A wearable bracelet accompanied by environment beacons will measure physical activity, sleep and room location within the home. Self-reports of mental and physical fatigue will be collected via smartphone app four times daily and on-demand. Validated fatigue and affect questionnaires will be completed at baseline and at 2 weeks. End-of-study interviews will investigate experiences of fatigue and study participation. A feedback session will be offered to participants to discuss their data.Data will be analysed using multilevel modelling and machine learning. Interviews and feedback sessions will be analysed using content or thematic analyses.Ethics and dissemination This study was approved by the East of England—Cambridge East Research Ethics Committee (22/EE/0261). The results will be disseminated in peer-reviewed journals and at international conferences.Trial registration number NCT05622669.

Digital health technologies have the potential to measure how a patient feels and functions with low patient burden e.g.: to provide real-world evidence of the benefit of a novel treatment. Despite this potential, traditional wearable devices, e.g. actigraphy, do not provide location-activity information which may improve the validity of these measures relative to established clinical measures. We describe a system that combines a wearable bracelet with Bluetooth-low-energy environment beacons to help localize and provide environmental context to that wearable data, and therefore provide a more clinically relevant measure of function. We describe an initial validation of the accuracy of the location information provided by this system in a study of 5 different real living environments of different size and layout, each collecting data over multiple days and recording their actual location in their home every 15 minutes in a diary. The results are presented in a confusion matrix. Mean overall accuracy was 94.0% (range 88.8-98.8%), which is sufficient to enable construction of more meaningful outcomes for patients than activity alone. For example, to determine if someone is moving around their home more, if they are getting outside more, if they are spending more time in bed, etc. It may be possible to improve location accuracy further with more sophisticated analysis of the beacon data.

The growing HCI agenda on health has focused on different chronic conditions but less so on Long Covid, despite its severe impact on the quality of life. We report findings from 2 workshops with 13 people living with Long Covid, indicating the challenges of making sense of their physical, cognitive, and emotional symptoms, and of monitoring the triggers of post-exertional malaise. While most participants engage in pacing activities for the self-management of fatigue, only a few are aware of the importance of planning all their daily activities and routines in order to avoid post-exertional malaise. We conclude with design implications to support lightweight tracking and sensemaking of fatigue symptoms, novel data analytics for monitoring the triggers of post-exertional malaise and the worsening of symptoms, and support for self-management in order to prevent post-exertional malaise.

Abstract Introduction This study examined the long‐term influence of loneliness and social isolation on mental health outcomes in memory assessment service (MAS) attendees and their care partners, with a focus on interdependence and bidirectionality. Methods Longitudinal data from 95 clinic attendees with cognitive impairment, and their care partners (dyads), from four MAS in the North of England were analyzed. We applied the actor–partner interdependence model, seeking associations within the dyad. At baseline and 12‐month follow‐up, clinic attendees and care partners completed measures of loneliness and social isolation, depression, and anxiety. Results Social isolation at baseline was more prevalent in care partners compared to MAS attendees. Social isolation in MAS attendees was associated with higher anxiety symptoms (β = 0.28, 95% confidence intervals [CIs] = 0.11 to 0.45) in themselves at 12 months. We found significant positive actor and partner effects of loneliness on depression (actor effect: β = 0.36, 95% CIs = 0.19 to 0.53; partner effect: β = 0.23, 95% CIs = 0.06 to 0.40) and anxiety (actor effect: β = 0.39, 95% CIs = 0.23 to 0.55; partner effect: β = 0.22, 95% CIs = 0.05 to 0.39) among MAS attendees 1 year later. Loneliness scores of the care partners have a significant and positive association with depressive (β = 0.36, 95% CIs = 0.19 to 0.53) and anxiety symptoms (β = 0.32, 95% CIs = 0.22 to 0.55) in themselves at 12 months. Discussion Loneliness and social isolation in MAS clinic attendees had a downstream effect on their own and their care partners’ mental health. This highlights the importance of including care partners in assessments of mental health and social connectedness and expanding the remit of social prescribing in the MAS context.

Mobile health (mHealth) is becoming a prominent component of healthcare. As the border between wearable consumer devices and medical devices begins to thin, we extend the mHealth definition including sports, lifestyle, and wellbeing apps that may connect to smart bracelets and watches as well as medical device apps running on consumer platforms and dedicated connected medical devices. This trend raises security and privacy concerns, since these technologies collect data ubiquitously and continuously, both on the individual user and on the surroundings. Security issues include lack of authentication and authorization mechanisms, as well as insecure data transmission and storage. Privacy issues include users' lack of control on data flow, poor quality consent management, and limitations on the possibility to remain anonymous. In response to these threats, we propose an advanced reference platform, securing the use of wearables and mobile apps in the mHealth domains through citizens' active protection and information.

Smartphones and wearables are widely recognised as the foundation for novel Digital Health Technologies (DHTs) for the clinical assessment of Parkinson’s disease. Yet, only limited progress has been made towards their regulatory acceptability as effective drug development tools. A key barrier in achieving this goal relates to the influence of a wide range of sources of variability (SoVs) introduced by measurement processes incorporating DHTs, on their ability to detect relevant changes to PD. This paper introduces a conceptual framework to assist clinical research teams investigating a specific Concept of Interest within a particular Context of Use, to identify, characterise, and when possible, mitigate the influence of SoVs. We illustrate how this conceptual framework can be applied in practice through specific examples, including two data-driven case studies.

Sensor data from digital health technologies (DHTs) used in clinical trials provides a valuable source of information, because of the possibility to combine datasets from different studies, to combine it with other data types, and to reuse it multiple times for various purposes. To date, there exist no standards for capturing or storing DHT biosensor data applicable across modalities and disease areas, and which can also capture the clinical trial and environment-specific aspects, so-called metadata. In this perspectives paper, we propose a metadata framework that divides the DHT metadata into metadata that is independent of the therapeutic area or clinical trial design (concept of interest and context of use), and metadata that is dependent on these factors. We demonstrate how this framework can be applied to data collected with different types of DHTs deployed in the WATCH-PD clinical study of Parkinson’s disease. This framework provides a means to pre-specify and therefore standardize aspects of the use of DHTs, promoting comparability of DHTs across future studies.

Background: Psoriatic arthritis (PsA) has many consequences, reflecting musculoskeletal and skin inflammation, with the potential to adversely affect overall quality of life. Patient reported outcome measures (PROM) assess a holistic range of aspects of quality of life, including physical and mental components, and provide broad detailed information of the impact of disease. Biologic DMARDs (bDMARDs) targeting TNF have been used to treat PsA for over 10 years whereas inhibitors of IL-17, IL-12/23 and Janus kinases (JAK) have only been available more recently. They all target differing cytokines, including JAK inhibitors which inhibit IL-12 and IL-23 signaling but not TNF signaling. Their relative impact on PROMs is unknown. Objectives: To assess, in routine care, the relative impact in PsA of TNF inhibitors (TNFi) versus non-TNFi bDMARDs, targeting IL-17, IL-12/23 and JAK, on PROMs. Methods: We performed a cross section analysis of PsA patients with established disease treated with bDMARDs and JAKi, under routine care at St George’s University Hospital, London, UK. Patients completed the 12-item psoriatic arthritis impact of disease (PsAID) tool. The total PsAID score was calculated using the on-line EULAR toolkit (see reference). The PsAID total and individual domain scores were compared between TNFi and non-TNFi groups using the Mann Whitney U test. A total PsAID score below 4 out of 10 is considered a ‘patient-acceptable state’. Results: A total 95 patients (female n= 53, 56%) completed the PSAID; TNFi n=72 (female 50%, adalimumab n=41, Etanercept n= 24, Golimumab n=4, Infliximab n =2, Certolizumab n=1) and non-TNFi n= 23 (female 74%, Secukinumab n=9, Ixekizumab n=1, Ustekinumab n=9, Tofacitinib n=4). The mean age was 53.6 (TNFi 53.5, non-TNFi 53.7) years, and duration of time on bDMARD treatment was TNFi 49.5 (range 1- 141) months, non-TNFi 25.3 (range 4 -59) months. The total and individual domain PsAID scores are shown in the Table 1. A ‘patient acceptable state’ total score <4 was recorded in TNFi 36/72 (50%) and non-TNFi 11/23 (48%). There was no significant difference between TNFi and non-TNFi groups in the mean total PsAID score, or proportion achieving a patient acceptable state. Patients on TNFi had lower (better outcome) mean scores for all 12 domains except skin, and the differences, versus non-TNFi treated patients, were significant for pain, functional capacity, discomfort and depression. PSAID domain TNFi Non-TNFi P value Total score 3.31 4.64 N.S. Pain 3.67 5.43 0.02 Fatigue 4.04 5.65 N.S. Skin problems 3.11 2.78 N.S. Work/leisure activities 3.56 4.78 N.S. Functional capacity 3.29 5.04 0.02 Discomfort 3.88 5.65 0.02 Sleep disturbance 3.42 4.78 N.S. Coping 3.01 4.09 N.S. Anxiety, fear, uncertainty 2.64 4.17 N.S. Embarrassment/shame 2.28 3.39 N.S. Social participation 2.6 3.48 N.S. Depression 2.11 3.91 0.03 Conclusion: In PsA, TNFi appear to have a greater impact over non-TNFi bDMARDs on some aspects of quality of life, including pain and functional capacity. TNFi and non-TNFi were no different with respect to patients’ perspective on skin disease, embarrassment or shame, despite less good cutaneous responses in clinical trials from TNFi agents. Overall, the PsAID tool reveals an unmet burden on quality of life in PsA patients treated with all classes of bDMARDs and JAKi, as 50% fail to achieve a ‘patient acceptable state’. This should prompt scrutiny of the high scoring domains and utilization of additional treatment modalities to achieve better holistic outcomes for PsA patients in routine care. References: [1]PsAID tool: http://pitie-salpetriere.aphp.fr/psaid/raid_psaid_quest_home.php Disclosure of Interests: Atif Rauf: None declared, Catherine Hughes: None declared, Diane Hill: None declared, Patrick Kiely Speakers bureau: Abbvie.

PURPOSE:A standard MRI system phantom has been designed and fabricated to assess scanner performance, stability, comparability and assess the accuracy of quantitative relaxation time imaging. The phantom is unique in having traceability to the International System of Units, a high level of precision, and monitoring by a national metrology institute. Here, we describe the phantom design, construction, imaging protocols, and measurement of geometric distortion, resolution, slice profile, signal-to-noise ratio (SNR), proton-spin relaxation times, image uniformity and proton density. METHODS:The system phantom, designed by the International Society of Magnetic Resonance in Medicine ad hoc committee on Standards for Quantitative MR, is a 200 mm spherical structure that contains a 57-element fiducial array; two relaxation time arrays; a proton density/SNR array; resolution and slice-profile insets. Standard imaging protocols are presented, which provide rapid assessment of geometric distortion, image uniformity, T1 and T2 mapping, image resolution, slice profile, and SNR. RESULTS:Fiducial array analysis gives assessment of intrinsic geometric distortions, which can vary considerably between scanners and correction techniques. This analysis also measures scanner/coil image uniformity, spatial calibration accuracy, and local volume distortion. An advanced resolution analysis gives both scanner and protocol contributions. SNR analysis gives both temporal and spatial contributions. CONCLUSIONS:A standard system phantom is useful for characterization of scanner performance, monitoring a scanner over time, and to compare different scanners. This type of calibration structure is useful for quality assurance, benchmarking quantitative MRI protocols, and to transition MRI from a qualitative imaging technique to a precise metrology with documented accuracy and uncertainty.

Identifying predictors of cognitive ability and brain structure in later life is an important step towards understanding the mechanisms leading to cognitive decline and dementia. This study used ultra-performance liquid chromatography mass spectrometry (UPLC-MS) and nuclear magnetic resonance (NMR) to measure targeted and untargeted metabolites, mainly lipids and lipoproteins, in ∼600 members of the Lothian Birth Cohort 1936 (LBC1936) at aged ∼73 years. Penalized regression models (LASSO) were then used to identify sets of metabolites that predict variation in general cognitive ability and structural brain variables. UPLC-MS-POS measured lipids, together predicted 19% of the variance in total brain volume and 17% of the variance in both grey matter and normal appearing white matter volumes. Multiple subclasses of lipids were included in the predictor, but the best performing lipid was the sphingomyelin SM(d18:2/14:0) which occurred in 100% of iterations of all three significant models. No metabolite set predicted cognitive ability, or white matter hyperintensities or connectivity. Future studies should concentrate on identifying specific lipids as potential cognitive and brain-structural biomarkers in older individuals.### Competing Interest StatementThe authors have declared no competing interest.

Our goal was to assess the enrichment utility of hippocampal volume (HV) as an enrichment biomarker in amnestic mild cognitive impairment (aMCI) clinical trials, and, hence, develop an HV neuroimaging-informed clinical trial enrichment tool. Modeling of integrated longitudinal patient-level data came from open-access natural history studies in patients diagnosed with aMCI-the Alzheimer's Disease Neuroimaging Initiative (ADNI)-1 and ADNI-2-and indicated that a decrease of 1 cm(3) with respect to the analysis dataset median baseline intracranial volume-adjusted HV (ICV-HV; ~ 5 cm(3)) is associated with > 50% increase in disease progression rate as measured by the Clinical Dementia Rating Scale-Sum of Boxes. Clinical trial simulations showed that the inclusion of aMCI subjects with baseline ICV-HV below the 84th or 50th percentile allowed an approximate reduction in trial size of at least 26% and 55%, respectively. This clinical trial enrichment tool can help design more efficient and informative clinical trials.

The Dementias Platform UK Data Portal is a data repository facilitating access to data for 3 370 929 individuals in 42 cohorts. The Data Portal is an end-to-end data management solution providing a secure, fully auditable, remote access environment for the analysis of cohort data. All projects utilising the data are by default collaborations with the cohort research teams generating the data. The Data Portal uses UK Secure eResearch Platform infrastructure to provide three core utilities: data discovery, access, and analysis. These are delivered using a 7 layered architecture comprising: data ingestion, data curation, platform interoperability, data discovery, access brokerage, data analysis and knowledge preservation. Automated, streamlined, and standardised procedures reduce the administrative burden for all stakeholders, particularly for requests involving multiple independent datasets, where a single request may be forwarded to multiple data controllers. Researchers are provided with their own secure ‘lab’ using VMware which is accessed using two factor authentication. Over the last 2 years, 160 project proposals involving 579 individual cohort data access requests were received. These were received from 268 applicants spanning 72 institutions (56 academic, 13 commercial, 3 government) in 16 countries with 84 requests involving multiple cohorts. Projects are varied including multi-modal, machine learning, and Mendelian randomisation analyses. Data access is usually free at point of use although a small number of cohorts require a data access fee.

Innovative tools are urgently needed to accelerate the evaluation and subsequent approval of novel treatments that may slow, halt, or reverse the relentless progression of Parkinson disease (PD). Therapies that intervene early in the disease continuum are a priority for the many candidates in the drug development pipeline. There is a paucity of sensitive and objective, yet clinically interpretable, measures that can capture meaningful aspects of the disease. This poses a major challenge for the development of new therapies and is compounded by the considerable heterogeneity in clinical manifestations across patients and the fluctuating nature of many signs and symptoms of PD. Digital health technologies (DHT), such as smartphone applications, wearable sensors, and digital diaries, have the potential to address many of these gaps by enabling the objective, remote, and frequent measurement of PD signs and symptoms in natural living environments. The current climate of the COVID-19 pandemic creates a heightened sense of urgency for effective implementation of such strategies. In order for these technologies to be adopted in drug development studies, a regulatory-aligned consensus on best practices in implementing appropriate technologies, including the collection, processing, and interpretation of digital sensor data, is required. A growing number of collaborative initiatives are being launched to identify effective ways to advance the use of DHT in PD clinical trials. The Critical Path for Parkinson’s Consortium of the Critical Path Institute is highlighted as a case example where stakeholders collectively engaged regulatory agencies on the effective use of DHT in PD clinical trials. Global regulatory agencies, including the US Food and Drug Administration and the European Medicines Agency, are encouraging the efficiencies of data-driven engagements through multistakeholder consortia. To this end, we review how the advancement of DHT can be most effectively achieved by aligning knowledge, expertise, and data sharing in ways that maximize efficiencies.

The Critical Path for Parkinson's (CPP) Imaging Biomarker and Modeling and Simulation working groups aimed to achieve qualification opinion by the European Medicines Agency (EMA) Committee for Medical Products for Human Use (CHMP) for the use of baseline dopamine transporter neuroimaging for patient selection in early Parkinson's disease clinical trials. This paper describes the regulatory science strategy to achieve this goal. CPP is an international consortium of three Parkinson's charities and nine pharmaceutical partners, coordinated by the Critical Path Institute.

As therapeutic trials target early stages of Parkinson's disease (PD), appropriate patient selection based purely on clinical criteria poses significant challenges. Members of the Critical Path for Parkinson's Consortium formally submitted documentation to the European Medicines Agency (EMA) supporting the use of Dopamine Transporter (DAT) neuroimaging in early PD. Regulatory documents included a comprehensive literature review, a proposed analysis plan of both observational and clinical trial data, and an assessment of biomarker reproducibility and reliability. The research plan included longitudinal analysis of the Parkinson Research Examination of CEP-1347 Trial (PRECEPT) and the Parkinson's Progression Markers Initiative (PPMI) study to estimate the degree of enrichment achieved and impact on future trials in subjects with early motor PD. The presence of reduced striatal DAT binding based on visual reads of single photon emission tomography (SPECT) scans in early motor PD subjects was an independent predictor of faster decline in UPDRS Parts II and III as compared to subjects with scans without evidence of dopaminergic deficit (SWEDD) over 24 months. The EMA issued in 2018 a full Qualification Opinion for the use of DAT as an enrichment biomarker in PD trials targeting subjects with early motor symptoms. Exclusion of SWEDD subjects in future clinical trials targeting early motor PD subjects aims to enrich clinical trial populations with idiopathic PD patients, improve statistical power, and exclude subjects who are unlikely to progress clinically from being exposed to novel test therapeutics.