Stephan Jonas

Background: Timely recognition of cognitive impairment and Alzheimer's disease (AD) is crucial for providing appropriate care to elderly individuals affected by these conditions. Digital short tests may help general practitioners (GPs) make more targeted referrals to specialists. Objective: We studied the real-world effectiveness of a referral pathway including the digital Montreal Cognitive Assessment (MoCA Duo), administered by medical assistants (MAs) in general practitioner (GP) offices, to identify mild or major neurocognitive impairment. Ancillary studies examined the procedure's feasibility, usability, and barriers in the care pathway. Methods: Ten GPs participated, one or two MAs per practice administered the MoCA Duo to patients with cognitive concerns. Patients with likely or questionable impairment were offered a referral to a memory clinic (MC) if deemed clinically appropriate. At the MC, patients received a comprehensive neuropsychological examination and a geriatric assessment. Results: Of the 299 patients tested in GP practices (mean age: 73; 65% female), 100 achieved MoCA Duo scores in the presumably unimpaired range. 93 patients (46.73%) scoring below 27 points were referred to the MC; of those, 66 attended an appointment. 40 patients had major cognitive impairment, including 9 diagnosed with AD or vascular dementia, and 15 had mild cognitive impairment (MCI). 83.33% of patients who screened positive on the MoCA Duo at the GP practice and were assessed at the MC had at least MCI. Conclusions: Digital tests administered by MAs can help to effectively screen for cognitive impairment in primary care, supporting targeted referrals of patients with likely cognitive impairment.

In human pose estimation, a comprehensive evaluation of state-of-the-art frameworks is necessary to advance both research and practical applications. This paper presents a thorough review of state-of-the-art 2D and 3D human pose estimation frameworks, analyzing 118 papers and four GitHub repositories, with a focus on frameworks made since 2019. The following frameworks are chosen based on predefined inclusion criteria: AlphaPose, Detectron2, MediaPipe, MeTRAbs, MHFormer, MMPose, MoveNet, OpenPifPaf, OpenPifPaf-vita, OpenPose, PoseFormerV2, rtmlib, StridedTransformer-Pose3D, ultralytics (YOLOv8), ViTPose, and YOLOv7. This paper evaluates these 16 frameworks on an existing, unpublished dataset consisting of exercise videos recorded with a monocular RGB camera and synchronized gold-standard motion capture data. The dataset includes videos of nine individuals performing eight exercises, recorded from two camera views with different planar angles. The analysis evaluates joint angle performance of the frameworks using weighted mean absolute error and weighted intraclass correlation coefficient as quantitative metrics. MeTRAbs emerged as the best overall framework, while AlphaPose, rtmlib, and YOLOv7 were the top 2D performers. The used code is available open source.¹.

BACKGROUND Psychiatric and neurological conditions are often accompanied by alterations in empathic resonance with emotional states of others, yet objective measures of affective empathy remain limited in clinical practice. Mobile and privacy-preserving facial expression analysis can provide a scalable approach for assessing spontaneous empathic reactivity in research and clinical practice. OBJECTIVE This paper aimed to evaluate an established iOS-based, computationally efficient and privacy-preserving facial expression assessment of empathy using mobile 3D face tracking during established static and dynamic empathy paradigms. METHODS We integrated Apple ARKit-based 3D facial landmark extraction from an iPad True-Depth assisted camera into a multidimensional empathy assessment in 32 healthy adults. Participants viewed affect-laden pictures from the Multifaceted Empathy Test-core-2 and emotional film clips while facial blend shapes with semantically meaningful labels (eg, cheekSquint, eyeSquint) were extracted in real time. Linear mixed-effects models tested valence-specific reactivity, temporal modulation, task-demand effects, and associations with subjective empathy ratings, trait-empathy questionnaires, age, and sex. Hierarchical clustering examined co-activation patterns among blend shapes. RESULTS Positive stimuli elicited valence-specific facial reactivity, including increased smile-related activations (eg, mouthSmile: β=.241, P<.001) and reduced negative-valence expressions (eg, mouthFrown: β=.131, P<.001) during positive trials, consistent with emotional contagion. Reactivity showed dynamic temporal patterns, and film clips elicited stronger responses than static images for several positive-expression blend shapes. Hierarchical clustering revealed coherent blend shape co-activations. Facial reactivity was associated with subjective affective empathy ratings, empathic-trait questionnaires, and demographic variables. CONCLUSIONS Mobile 3D face tracking captured subtle, spontaneous, and temporally dynamic facial reactivity during integrated empathy tasks while preserving privacy by avoiding raw video storage. This mHealth approach provides a scalable, clinically translatable framework for objective assessment of affective empathy and supports future implementation in neuropsychiatric research and mobile assessment.

Background: The European General Data Protection Regulation (GDPR) strictly regulates the processing of personal and health-related data, posing challenges for digital health research, especially when data are collected using participants’ own devices. Although scientific data can theoretically be anonymized, standard internet communication protocols inevitably expose transmission metadata, preventing true anonymization. Existing solutions, including virtual private networks, reverse proxies, and trust centers, improve confidentiality but do not technically or legally enable fully anonymized data collection. Consequently, large-scale digital health research often requires extensive organizational measures, complex consent procedures, and high regulatory overhead. Objective: This study aimed to develop a GDPR-compliant concept for fully anonymized scientific data collection, ensuring that no entity has simultaneous access to identifying information and donated data. We also implemented and evaluated this concept in a real-world public-private partnership. Methods: We designed a data donation architecture based on a blinded deidentification proxy that decouples identifying transmission metadata from encrypted user data at the time of donation. The concept combines symmetric (Advanced Encryption Standard-128 in Cipher Block Chaining) and asymmetric (Rivest-Shamir-Adleman with Optimal Asymmetric Encryption Padding) encryption, enabling end-to-end encrypted and anonymized data transfer without persistent identifiers. The system was integrated into the HerzFit app, a mobile lifestyle coach for cardiovascular disease prevention available in German-speaking countries, and evaluated for adoption, technical feasibility, and performance. Performance overhead was assessed using round-trip time benchmarks. Duplicate donations were identified and merged to estimate unique data donors. Results: The solution was integrated and tested in the HerzFit app with more than 200,000 downloads between April 2022 and December 2025. Since the introduction of the data donation feature, more than 13,000 donations have been received, translating to more than 9000 individual users contributing anonymized datasets. Proxy-based transmission resulted in an average round-trip time of 143 ms, compared to 58 ms for direct transfer, representing a modest overhead while maintaining usability. The operator of the donation database did not gain access to identifying information at any stage, demonstrating full technical anonymization. The approach can be operated reliably at scale with minimal server resources due to the stateless proxy design. Conclusions: This work introduces a novel system architecture enabling fully anonymized, GDPR-compliant data donation directly from participants’ devices. By decoupling identifying metadata from encrypted health data, the concept minimizes regulatory effort, strengthens privacy protection, and provides a practical framework for large-scale digital health research in research partnerships, for example, between a private company and a research institution. The real-world deployment in HerzFit demonstrates the feasibility, scalability, and scientific utility of this approach. The concept is broadly transferable to other mobile health apps and has the potential to substantially expand ethically and legally compliant data acquisition.

We present a low-cost, fully reproducible software and hardware protocol for smartphone sensor battery cost tests. Our pipeline combines a rigorous hardware checklist and light-sealed enclosure, a software checklist for iOS devices, and a BatteryTest app to control sensor configurations and log battery state during tests. Methodologically, we applied this standardized protocol in 30 independent analyzed test runs using six iPhone 14 Pro and three iPhone 13 Pro devices, and compared battery-life outcomes across predefined sensor conditions (idle, TrueDepth, GPS, accelerometer, pedometer, gyroscope, and rear camera), sampling rates, and sensor-specific settings. Key findings include: (i) Baseline battery life was approximately 10% higher on the 14 Pro versus the 13 Pro models under idle conditions. Sensor activation substantially reduced battery life, with GPS and camera usage exhibiting the strongest impact. (ii) Software parameters matter: the sampling rate change from 27 s to 3 s led to significantly decreased battery life in several scenarios, while reducing the location accuracy in GPS tests increased battery life by up to 20 h on the 13 Pro devices. (iii) Cross-device-generation consistency is heterogeneous. The iPhone 14 Pro lasts up to 50% longer on GPS tests, yet drains about an hour faster than the 13 Pro in camera tests. This work introduces the first standardized, and fully reproducible protocol for quantifying sensor-specific battery consumption on iPhones, enabling consistent, comparable, and low-cost energy benchmarking across device generations.

Speech alterations are early markers of neurodegenerative diseases. Transformer-based speech models such as Whisper have advanced automated speech assessment, but most systems rely on cloud-based computation, raising privacy concerns. On-device processing could offer a scalable and privacy-preserving alternative. This research’s objective was to evaluate whether a fully on-device speech analysis pipeline can achieve competitive accuracy in detecting Alzheimer’s disease and to quantify the contributions of acoustic, linguistic, and embedding features. Therefore, we developed an iOS application running all components, including acoustic analysis, two transformer-based speech-to-text modules (WhisperBase and quantized CrisperWhisper), linguistic feature extraction, and embedding generation, directly on the device. Using the ADReSS Challenge 2020 dataset (N = 156), we trained classical machine-learning classifiers across 20 configurations and evaluated them via a stratified 10-fold cross-validation. Area under the receiver operating curve (AUC), accuracy, precision, recall, and F1 scores were used as performance metrics. An ablation study examined the relevance of each feature group. The best-performing setup (Random Forest with CrisperWhisper transcription and Apple embeddings) achieved an accuracy of 85.4% and an AUC of 0.85. Performance was 5–7% below benchmark models relying on manual transcripts or server-based processing. Embedding features provided the strongest individual contribution, but the highest accuracy required combining acoustic, linguistic, and embedding information. A fully on-device pipeline for Alzheimer’s disease detection from speech is feasible and achieves competitive accuracy while maintaining strict data privacy. These findings highlight the potential of on-device transformer architectures for scalable, privacy-preserving digital screening. Future work should validate the approach in larger and more diverse cohorts.

Background: Scalable, process-sensitive tools are needed to address the gap between neuropathological onset and clinical recognition of cognitive decline. Tablet-based assessments can capture behavioral and temporal data beyond conventional test metrics, but require validation in target populations. Objective: To evaluate whether a tablet-based Symbol Digit Modalities Test (SDMT) and a response-time-based Performance Index capture clinically relevant cognitive performance in a memory-clinic population with mild cognitive impairment (MCI) or dementia. Methods: In a cross-sectional validation study, forty-one adults (mean age 73.7 years, 51% scoring below the CERAD cutoff for likely major cognitive impairment) completed a tablet-based SDMT alongside standard neuropsychological assessments (CERAD battery, MoCA). Digital response counts were compared with oral SDMT scores, the Performance Index was evaluated against clinical reference measures, and exploratory analyses examined within-task response-time patterns across repeated symbol exposures. Results: Digital and oral SDMT scores showed strong rank-order agreement ( ρ = 0 . 77 – 0 . 80 , p < . 001 ), with systematically lower absolute digital scores, indicating preserved inter-individual ranking but non-interchangeable modalities. The Performance Index and correct response count showed comparable associations with MoCA ( ρ = 0 . 59 and 0 . 62 ) and CERAD ( ρ = 0 . 49 for both; all p ≤ . 001 ); their convergence was attributable to near-ceiling response accuracy (mean 95.6%, excluding skipped items). Exploratory analyses revealed within-task response-time facilitation across repeated symbol exposures. Conclusion: The tablet-based SDMT preserves inter-individual ranking relative to the oral format, and the Performance Index generalizes to a memory-clinic population. Within-task response-time facilitation was observed even among participants with impaired declarative memory. Whether it reflects implicit or procedural learning rather than explicit mediation remains a hypothesis for controlled study.

Background: Digital short cognitive tests administered by medical assistants (MAs) in general practitioners’ (GPs) practices have great potential for the timely identification of patients with dementia, because they can lead to targeted specialist referrals or to immediate reassurance of patients regarding their perceived concerns. However, integration of this testing approach into clinical practice requires good usability for the test itself, especially for cognitively impaired older adults. Objective: In this implementation study, the digital version of the Montreal Cognitive Assessment (MoCA) Duo was conducted by MAs in general practice. We tested if the interaction with the test is associated with usability problems for the patients and aimed to find additional relevant constructs that should be considered for the potential implementation of such digital tests into clinical practice. We focused the study on subjective success, usability, and workload as well as their association with the result of the cognitive test to assess whether the MoCA Duo can be implemented into general practice. Methods: In total, 10 GPs took part in the study. Within their practices, 299 GP patients (aged 51-97 years) with cognitive concerns completed the MoCA Duo administered by MAs. Subsequently, patients and MAs completed digital questionnaires regarding the interaction with the app. Usability was measured using the adapted System Usability Scale, and perceived workload using the National Aeronautics and Space Administration Task Load Index. For the perceived workload, we included an assessment of the patient by the MA. Results of the MoCA Duo were supplied to the GPs for their consultation with the patient. Results: The results indicated good usability for the MoCA Duo. Self-assessment by the patients indicated that 64% (191/299) could perform in the test to the best of their ability, affected by their MoCA score. We found significant higher usability ratings by patients with better MoCA scores as well as by younger patients. Furthermore, the perceived workload showed overall medium workload. We found significant correlations between the subjective perceived workload of the patients and the assessment by MAs. Self-assessments as well as assessments by the MAs were significantly influenced by the MoCA scores and the age of the participants. Conclusions: The results indicate good usability of the digital MoCA within the sample, supporting the idea that the resulting scores are adequate to assess cognitive status without dependence on technological affinity. Furthermore, the results highlight the relevance of heterogenous samples for comparable evaluation studies, based on the significant effect of cognitive status and age on usability and workload.