SLStephan Johann Lehmler

Stephan Johann Lehmler

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Stephan Johann Lehmler has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Distributional Properties of ReLU-Activations in Artificial Neural Networks that Learn by Memorization

    by Stephan Johann Lehmler, Muhammad Saif-Ur-Rehman, Tobias Glasmachers, Ioannis Iossifidis

  • Open Access

    Comprehensive perceptions at the interface between health and environment: Applications models with a citizen science tool

    by Frauke Nees, Karina Janson, Stephan Lehmler, Sarah Böttger, Mira Tschorn, Philipp Hummer, Gunter Schumann, Michael Rapp, Sebastian Siehl, Nathalie Holz, et al.

    Background: Lived experience – how individuals perceive and interact with their environment – plays a central role in understanding mental health. Yet, insights into this first-person perspective, including subjective thoughts, emotions, and socio-contextual influences, remain limited in current research approaches. Methods: To address this gap, we developed StreetMind , a scalable, secure, and user-friendly digital citizen science platform grounded in a psycho-sociogeographic framework. The platform collects self-reported data on individuals’ activity spaces through a mobile app and web interface, capturing location visits, travel routes, and daily experiences. These subjective reports are combined with objective real-time health, environmental, and sociocultural data to generate integrated community “footprints.” Results: Initial usage data ( N = 1,010 for location and route entries; N = 509 for daily experiential data) demonstrate the platform’s structural robustness and functional feasibility. StreetMind enables classification of daily experiences by linking personal perceptions with contextual environmental data. This integration facilitates the identification and quantification of key environmental and psychosocial factors associated with mental well-being. Conclusion: StreetMind offers a novel, data-rich mapping of health–environment interactions by merging individual lived experience with environmental metrics. This approach supports the creation of dynamic “health–environment spaces” and holds promise for informing public health strategies and advancing precision mental health care.


  • Open Access

    An Ecological Momentary Assessment Approach of Environmental Triggers in the Role of Daily Affect, Rumination, and Movement Patterns in Early Alcohol Use Among Healthy Adolescents: Exploratory Study

    by Maren Prignitz, Stella Guldner, Stephan Johann Lehmler, Pascal-M Aggensteiner, Frauke Nees, IMAC-Mind Consortium

    Background: Adolescence is a period characterized by an increased susceptibility to developing risky alcohol consumption habits. This susceptibility can be influenced by social and situational factors encountered in daily life, which, in conjunction with emotions and thoughts, contribute to behavioral patterns related to alcohol use even in the early stages of alcohol experimentation, when initial experiences with alcohol are formed, and regular consumption is still evolving. Objective: This study aimed to examine the association between detailed behavioral and movement patterns, along with emotional and cognitive factors, and the early onset of alcohol use in the everyday lives of adolescents. Methods: A total of 65 healthy adolescents (33 male, twenty-nine 14-year-olds, and thirty-six 16-year-olds) underwent mobile-based ecological momentary assessments on alcohol (once a day at 9 AM, assessing alcohol use the day before), positive and negative affect, craving, rumination, and social context (6 prompts/day at 9 AM, 11 AM, 2 PM, 4 PM, 6 PM and 8 PM), type of day (weekdays or weekends, with weekend including Fridays, Saturdays, and Sundays), and using geospatial measures (specifically roaming entropy and number and type of trigger points for alcohol use met) over 14 days. After adjusting for a compliance rate of at least 50%, 52 participants (26 male and twenty-four 14-year-olds) were included in the analyses. Results: Generalized linear multilevel models revealed that higher positive affect (b=0.685, P=.007), higher rumination (b=0.586, P=.02), and a larger movement radius (roaming entropy) (b=8.126, P=.02) were positively associated with alcohol use on the same day. However, social context (b=–0.076, P=.90), negative affect (b=–0.077, P=.80), or potential trigger points (all P>.05) did not show significant associations. Alcohol use varied depending on the type of day, with more alcohol use on weekends (b=1.082, P<.001) and age (t50=–2.910, P=.005), with 16-year-olds (mean 1.61, SD 1.66) reporting more days of alcohol consumption than 14-year-olds (mean 0.548, SD 0.72). Conclusions: Our findings support previously identified factors as significant contributors to very early and low levels of alcohol consumption through fine-grained analysis of daily behaviors. These factors include positive affect, rumination, weekend days, and age. In addition, we emphasize that exploratory environmental movement behavior (roaming entropy) is also significantly associated with adolescent alcohol use, highlighting its importance as an additional factor.


  • Open Access

    An Ecological Momentary Assessment Approach of Environmental Triggers in the Role of Daily Affect, Rumination, and Movement Patterns in Early Alcohol Use Among Healthy Adolescents: Exploratory Study (Preprint)

    by Maren Prignitz, Stella Guldner, Stephan Johann Lehmler, Pascal-M Aggensteiner, Frauke Nees

    BACKGROUND Adolescence is a period characterized by an increased susceptibility to developing risky alcohol consumption habits. This susceptibility can be influenced by social and situational factors encountered in daily life, which, in conjunction with emotions and thoughts, contribute to behavioral patterns related to alcohol use even in the early stages of alcohol experimentation, when initial experiences with alcohol are formed, and regular consumption is still evolving. OBJECTIVE This study aimed to examine the association between detailed behavioral and movement patterns, along with emotional and cognitive factors, and the early onset of alcohol use in the everyday lives of adolescents. METHODS A total of 65 healthy adolescents (33 male, twenty-nine 14-year-olds, and thirty-six 16-year-olds) underwent mobile-based ecological momentary assessments on alcohol (once a day at 9 AM, assessing alcohol use the day before), positive and negative affect, craving, rumination, and social context (6 prompts/day at 9 AM, 11 AM, 2 PM, 4 PM, 6 PM and 8 PM), type of day (weekdays or weekends, with weekend including Fridays, Saturdays, and Sundays), and using geospatial measures (specifically roaming entropy and number and type of trigger points for alcohol use met) over 14 days. After adjusting for a compliance rate of at least 50%, 52 participants (26 male and twenty-four 14-year-olds) were included in the analyses. RESULTS Generalized linear multilevel models revealed that higher positive affect (b=0.685, P=.007), higher rumination (b=0.586, P=.02), and a larger movement radius (roaming entropy) (b=8.126, P=.02) were positively associated with alcohol use on the same day. However, social context (b=–0.076, P=.90), negative affect (b=–0.077, P=.80), or potential trigger points (all P>.05) did not show significant associations. Alcohol use varied depending on the type of day, with more alcohol use on weekends (b=1.082, P<.001) and age (t50=–2.910, P=.005), with 16-year-olds (mean 1.61, SD 1.66) reporting more days of alcohol consumption than 14-year-olds (mean 0.548, SD 0.72). CONCLUSIONS Our findings support previously identified factors as significant contributors to very early and low levels of alcohol consumption through fine-grained analysis of daily behaviors. These factors include positive affect, rumination, weekend days, and age. In addition, we emphasize that exploratory environmental movement behavior (roaming entropy) is also significantly associated with adolescent alcohol use, highlighting its importance as an additional factor. CLINICALTRIAL

  • Deep transfer learning compared to subject-specific models for sEMG decoders

    by Stephan Johann Lehmler, Muhammad Saif-Ur-Rehman, Glasmachers Tobias, Ioannis Iossifidis

    Objective . Accurate decoding of surface electromyography (sEMG) is pivotal for muscle-to-machine-interfaces and their application e.g. rehabilitation therapy. sEMG signals have high inter-subject variability, due to various factors, including skin thickness, body fat percentage, and electrode placement. Deep learning algorithms require long training time and tend to overfit if only few samples are available. In this study, we aim to investigate methods to calibrate deep learning models to a new user when only a limited amount of training data is available. Approach . Two methods are commonly used in the literature, subject-specific modeling and transfer learning. In this study, we investigate the effectiveness of transfer learning using weight initialization for recalibration of two different pretrained deep learning models on new subjects data and compare their performance to subject-specific models. We evaluate two models on three publicly available databases (non invasive adaptive prosthetics database 2–4) and compare the performance of both calibration schemes in terms of accuracy, required training data, and calibration time. Main results . On average over all settings, our transfer learning approach improves 5%-points on the pretrained models without fine-tuning, and 12%-points on the subject-specific models, while being trained for 22% fewer epochs on average. Our results indicate that transfer learning enables faster learning on fewer training samples than user-specific models. Significance . To the best of our knowledge, this is the first comparison of subject-specific modeling and transfer learning. These approaches are ubiquitously used in the field of sEMG decoding. But the lack of comparative studies until now made it difficult for scientists to assess appropriate calibration schemes. Our results guide engineers evaluating similar use cases.