Krešimir Friganović

Wearable inertial sensors quantify gait markers relevant to mobility and fall risk, but real-world measurements are often interpreted without environmental context in which walking occurs. We examined whether publicly available built-environment data can contextualize wearable-derived gait during habitual neighbourhood walking. Seventeen community-dwelling adults at elevated mobility risk were enrolled in the study. We captured 3,092 GPS-anchored outdoor 2-minute walking windows. IMU-derived gait outcomes were expressed as deviations from each participant’s supervised baseline walk and linked to five built-environment composites: thermal/weather, road exposure, greenery/shade, pedestrian shelter, and crossings/intersections. Composites were calculated by standardizing features, aligning their direction so higher values represent more adverse walking context, and averaging features within each composite domain. Per-participant Spearman correlations were calculated between gait outcomes and derived composites. CAMINO demonstrates how outdoor walking context can expose person-specific fall-risk patterns that pooled or context-free analyses would miss, advancing behavioral, environmental, and public health research.

A neighborhood’s built environment can challenge the mobility of older mobility-challenged adults (due to knee osteoarthritis and falls), reducing their participation and quality of life. The Built Environment in Falls and arthrITis (BE-FIT) study aims to understand the neighborhood influence on the mobility, participation, and psychosocial health of older adults with knee osteoarthritis and/or falls. BE-FIT comprises four work packages (WPs). WP1 quantitatively explores relationships among environmental-, social-, and person-related factors and participation outcomes of its intended population. WP2 employs qualitative methods to comprehend the relationships among WP1’s variables. Via a combination of wearable sensor technology and qualitative geospatial methods, WP3 aims to characterize its population’s movement behavior, mobility, functional activity of daily living, and lived experiences of residing in a mature neighborhood. Finally, WP4 engages crucial stakeholders to co-develop evidence-based recommendations to inform public health, urban planning, and aging policies and implementation. BE-FIT could benefit societies with rising incidence of knee osteoarthritis and falls by improving neighborhoods and lives of older mobility-challenged residents.

A significant fraction of the microbial diversity remains unclassified, hindering our understanding of microbial roles in health and ecosystems. State-of-the-art methods like Kraken 2 perform well for taxa that are present in the database. However, their accuracy drops significantly when classifying taxa that are not included. While deep learning has advanced many fields, its applications in metagenomics remain limited, and its full potential has yet to be realized. Here, we present Metaxa, a transformer-based deep learning model designed for the taxonomic classification of long-read Nanopore sequences. Metaxa leverages the sequential context of Nanopore reads, enabling robust classification beyond fixed k-mer profiles. Our results show that Metaxa matches Kraken 2 on in-sample data at both the species and genus levels, and significantly outperforms both Kraken 2 and MetageNN at the genus level on out-of-sample datasets where the species genome is absent from the reference database but a different species from the same genus is present. Furthermore, Metaxa demonstrates strong generalization across different Nanopore chemistries (R9.4.1 and R10.4.1). This work highlights the potential of deep learning models to improve metagenomic classification accuracy, especially in complex or underexplored environments where traditional tools fall short.

Gait variability is a critical functional indicator of dynamic balance and neurocognitive decline in health. Its translation into clinical practice is, however, challenged by a lack of age-related normative trajectories and reference values under real-world ecological settings. Furthermore, the conventional metrics used to estimate gait variability (Coefficient of Variation, CV ; Standard Deviation, SD ) have a fundamental methodological flaw: the inherent sensitivity of conventional metrics to the statistical outliers and environmental noise in real-world walking. In this study, we mitigate this factor by applying a robust statistical framework to quantify gait variability. Analysing a large-scale cohort of community-dwelling older adults (n=2,193), we first demonstrate that freeliving gait data follows a heavy-tailed distribution, necessitating the use of robust estimators like the Robust Coefficient of Variation ( RCV _MAD ) and Median Absolute Deviation ( MAD ). Leveraging these metrics, we established the normative trajectory and reference values of real-world gait variability across the ageing lifespan, revealing a distinct, age-dependent increase in spatio-temporal fluctuations, indicating a decline in rhythmicity and steadiness with age. We further demonstrated the clinical utility of these robust metrics: RCV _MAD consistently yielded larger effect sizes than conventional CV in discriminating between fallers and non-fallers across all gait parameters. Furthermore, we illustrate the potential of long-term unsupervised monitoring to capture intrinsic variability during real-world walking. Validated for consistency and reliability, this robust framework provides the necessary ecological validity to transform gait variability into a standardised, rapid clinical metric for assessing functional decline at an early timepoint.

Nanopore sequencing enables real-time, long-read analysis by processing raw signals as they are produced. A key step, segmentation of signals into events,

Nanopore sequencing enables real-time, long-read analysis by processing raw signals as they are produced. A key step, segmentation of signals into events, is typically handled algorithmically, struggling in noisy regions. We present Campolina, a first deep-learning frame-work for accurate segmentation of raw nanopore signals. Campolina uses a convolutional model to identify event boundaries and significantly outperforms the traditional Scrappie algorithm on R9.4.1 and R10.4.1 datasets. We introduce a comprehensive evaluation pipeline and show that Campolina aligns better with reference-guided ground-truth segmentation. We show that integrating Campolina segmentation into real-time frameworks, Sigmoni and RawHash2, improves their performance while maintaining time efficiency.

Objective detection of a driver’s drowsiness is important for improving driving safety, and the most prominent indicator of drowsiness is changes in electroencephalographic (EEG) activity. Despite extensively documented behavioral differences between male and female drivers, previous studies have not differentiated drowsiness detection models based on drivers’ sex. Therefore, the overall aim of this study is to demonstrate that drowsiness detection can be improved with the use of drivers’ sex information, either as a feature or as separate sex-dependent datasets. Additionally, we aim to provide a reliable EEG-based sex classification model. The used dataset consists of 17 male and 17 female drivers which were evaluated during alert and drowsy sessions. Frequency-domain and recurrence quantification analysis EEG features were used. Four classification algorithms and three feature selection methods were applied to build the models. The accuracy of drowsiness detection based on sex-dependent datasets is 84% for male drivers and 88% for female drivers, which is 3% and 7% better, respectively, than the classification without information about driver’s sex (81%). The model for sex classification based on EEG achieved high accuracy: 97% correctly identified participants in alert sessions and 96% in drowsy sessions. All participants were correctly classified after the application of majority voting on five algorithm runs. The results suggest that sex-dependent datasets improve the accuracy of drowsiness models, which may be relevant to a variety of drowsiness detection systems currently being developed in the field.