Aicinām Rīgas Stradiņa universitātes (RSU) un Rīgas Tehniskās universitātes (RTU) zinātniekus, studējošos un inovāciju entuziastus uz dizaina domāšanas sesiju, kuras mērķis ir radīt jaunas, starpdisciplināras un ilgtspējīgas idejas BioPhoT pētniecības un inovāciju projektu (PIP) pieteikumiem.

Sesiju vadīs Dr. Elīna Miķelsone, RTU Dizaina fabrikas vadītāja, un palīdzēs strukturēt radošo darbu, veicinās sadarbību un virzīs komandas uz skaidri definētām projektu iecerēm. Pasākums norisināsies angļu valodā.

Pasākuma laikā RSU un RTU iepazīstinās ar projektu idejām, kurā tiek meklētas sadarbības iespējas ar otru universitāti. 

Projektu idejas

System for Estimating Fingers' Apparent Stiffness

Author: Dr Valters Āboliņš

This challenge focuses on the need for a system that can estimate the apparent stiffness of an individual finger or multiple fingers of one hand. Each finger functions as an effector driven by two opposing muscle groups - agonists and antagonists. When the finger is involuntarily displaced, these muscles are either stretched or shortened, and their stretch reflexes produce changes in force output even if the underlying neural command remains unchanged. This behavior is conceptually similar to Hooke’s law: when the system of both muscle groups are stretched or shortened, the resulting force changes proportionally. In humans, this relationship defines apparent stiffness – a spring-like property of the neuromuscular system. Unlike a physical spring with fixed stiffness, apparent stiffness varies continuously because it depends on dynamic motor signals, including muscle co-activation and reflex modulation.

Because apparent stiffness captures both mechanical and neural contributions, it can provide a sensitive indicator of how the neuromuscular system regulates stability, force production, and voluntary action. A system for estimating this parameter would preferably operate in dynamic conditions or, if static, be sufficiently compact and mobile to allow flexible use in laboratory or clinical environments.

To develop such a system, additional competencies from RTU are needed, particularly in precision sensing, signal acquisition, and data-processing methodologies that can extract apparent stiffness from controlled perturbations and measured force responses. RTU’s capabilities in engineering and prototyping would enable the transformation of the physiological concept of apparent stiffness into a functional, reliable, and mobile research instrument.

Biomechanical Stability and Motor Control-Based Analysis of Running Technique Using Wearable Sensors and AI Tools

Author: Dr Edgars Bernāns

Running technique stability reflects the runner’s ability to adapt movement execution to changing environmental or physical conditions while maintaining performance efficiency. One promising way to assess this adaptability is through movement variability analysis combined with principles of motor control. Current wearable technologies already allow the collection of basic kinematic parameters such as stride length and stride time with sufficient precision, making them suitable not only for research but also for potential large-scale end-user applications.
A scientifically grounded approach to analyse movement variability is the Goal-Equivalent Manifold (GEM) framework, which allows distinguishing two meaningful components of variability: (1) goal-direction variability that supports adaptability and does not disrupt the movement goal, and (2) orthogonal variability that reflects destabilising fluctuations. This type of analysis provides deeper insight into the organisation and control of running technique than traditional average-based biomechanical metrics.

Despite its potential, GEM-based stability analytics are not yet widely implemented or applied in real-time systems. Expanding this methodology could enable the development of new personalised feedback tools for wearable devices, artificial intelligence–based performance monitoring, and smart training recommendations based on real-time race or training data.

To move forward, computational tools and machine learning models are needed to integrate sensor-collected motion data with GEM-based analytics and make results interpretable for users. Collaboration with RTU would provide essential engineering and digital innovation expertise, including algorithm development, sensor signal processing, AI-based modeling, data visualisation interfaces, and software integration.

This interdisciplinary project would contribute to cutting-edge sports science innovation, expanding the field of biomechanical motor control research and enabling scalable practical applications for athletes, coaches, and health-oriented runners.

Augmented Reality-Based Feedback and Feedforward System for Running Technique Optimization and Research Applications

Author: Dr Edgars Bernāns

Movement technique optimisation in running and walking is essential not only for performance enhancement but also for injury prevention and rehabilitation. While wearable sensors provide valuable biomechanical data, there remains a gap in translating these measurements into intuitive, user-friendly interventions that can be applied in real-time during movement. Augmented reality (AR) technologies offer a unique opportunity to deliver personalised, interactive feedback and feedforward cues during gait and running activities, supporting both skill acquisition and controlled experimental manipulation.

The proposed idea is to develop an augmented reality–based tool compatible with AR glasses that can be used in research settings as well as real-world training environments. Previously measured biomechanical parameters such as cadence, stride length, vertical oscillation or stability metrics (from motor control ans biomechanical analyses) could be integrated into the system, allowing the user to perform predefined movement tasks. The tool would enable precise dosing of interventions, such as maintaining a target step frequency or adjusting stride mechanics, while receiving multimodal feedback (visual, auditory, or haptic), depending on available technological capabilities and user needs.

In future iterations, integration with real-time wearable sensor data would allow the system to adapt feedback dynamically based on the runner’s ongoing performance, creating a closed-loop learning environment. Such a tool would provide researchers with controlled experimental manipulation options (feedforward vs. feedback conditions, perturbations, task constraints), while practitioners could apply it for guided training, rehabilitation, and personalised performance development.

RTU expertise would be essential in areas such as AR interface design, sensor integration, signal processing, AI-driven adaptive algorithms, and development of user-centred interaction systems. This collaboration would enable the creation of a scalable, interdisciplinary solution bridging sports science, biomechanics, cognitive motor learning, and emerging technology.

Development of a Finger-Specific Force Measurement System for Motor Control Research

Author: Dr Edgars Bernāns

Precise assessment of finger-generated force during gripping movements is essential for advancing the understanding of fine motor control, movement organisation, and neuromuscular coordination. Currently available laboratory tools offer only limited capability to measure finger-specific force outputs, are often not ergonomically adaptable, and typically do not support experimental conditions beyond simple static tasks. As a result, research in hand motor control, particularly finger coordination, lacks a robust and versatile measurement platform.

The proposed solution is the development of a modular, ergonomically adaptable force-sensing system that allows force measurement for each finger individually during grip tasks. The system should be adjustable to different hand morphologies to ensure usability across diverse populations and age groups. While initial implementation may focus on isometric force production, which represents the most commonly studied condition in this field, an ideal long-term design would also support dynamic gripping tasks to broaden the ecological validity and application range.

The companion software should allow real-time acquisition of force and timing parameters and support structured testing protocols. Additional functionality, such as feedback and feedforward task modules (target matching, motor learning tasks, variability assessments), would significantly increase the research value of the system. This would enable studying various aspects of motor control, including coordination strategies, adaptation, learning curves, and performance variability.

Although the primary use case is research rather than everyday clinical or consumer application, a successful and validated solution would fill a global gap in the field. Due to its unique capabilities, such a system has the potential to attract interest from international neuroscience, rehabilitation, and motor learning laboratories.

Collaboration with RTU would be essential for designing the hardware, selecting and integrating force sensors, developing the signal processing pipeline, and creating a functional software platform tailored to research needs.

Development and Validation of Methodology for Evaluating Cross-Country Ski / Biathlon Glide Performance Using RTU Sliding Test Platform

Author: Dr Edgars Bernāns

Glide efficiency is a critical performance factor in cross-country skiing, significantly influencing speed, energy expenditure, and overall race effectiveness. However, systematic and controlled evaluation of ski glide performance remains technically challenging due to environmental variability, snow type differences, waxing conditions, and limited standardized research tools. The testing platform currently being developed at RTU for measuring sliding properties of luge, bobsled, and skeleton athletes presents a unique opportunity to expand its application toward the biomechanics and material evaluation needs of cross-country skiing.

The proposed challenge is to adapt this testing platform to enable precise evaluation of ski-specific glide parameters. This would require developing experimental protocols, measurement criteria, and calibration procedures suitable for cross-country skis. Potential outcome variables may include glide coefficient, friction dynamics under different loads, temperature effects, waxing conditions, and interactions between ski structure and surface material.

A validated methodology would support both scientific and applied domains, enabling structured testing within research, elite sport preparation, and potential industry collaboration with ski manufacturers or waxing technology developers. The work could further integrate biomechanical factors, such as skier loading patterns or technique variability, to better reflect real-use conditions and improve ecological validity.

RTU expertise in mechanical engineering, materials science, sensor technology integration, and mechatronics would be essential for adapting the existing infrastructure, designing measurement interfaces, and developing software for automated acquisition and analysis. Collaboration would also support long-term development toward portable or semi-portable versions suitable for field use.

Although the project would be relevant after the completion of the current State Research Programme, this concept represents a promising future applied research direction with clear interdisciplinary value, international relevance, and commercialization potential in winter sports technology.

Virtual Immunohistochemical Staining from H&E Slides Using Deep Learning

Author: Edgars Edelmers

Immunohistochemistry (IHC) is essential for diagnosis and biomarker assessment but is costly, tissue-consuming and time-intensive. Recent advances in deep learning suggest that virtual IHC (vIHC) can be generated directly from routine hematoxylin–eosin (H&E) slides, potentially enabling targeted, cost-effective use of “wet” IHC.

Aim: To develop and validate a deep learning model that predicts virtual IHC stains from H&E whole-slide images for selected diagnostic and research biomarkers.

Objectives: 

• Assemble a paired H&E–IHC whole-slide dataset for 1–3 markers (e.g. Ki-67, HER2, CD117/ANO1);
• Train and optimise a generative model (e.g. conditional GAN / image-to-image transformer) for H&E → vIHC translation;
• Quantitatively and qualitatively validate vIHC against real IHC at slide-, region- and cell-level;
• Explore use of vIHC as a triage tool to prioritise cases for conventional IHC.

Methods: Retrospective cohort; WSIs digitised at high resolution. Paired registration of H&E and IHC sections, patch-based training. Evaluation using structural similarity, correlation of quantitative scores, and blinded pathologist review.

Expected results / impact: A validated vIHC pipeline capable of generating interpretable synthetic IHC maps from H&E, demonstrating potential reduction in unnecessary IHC requests and providing a foundation for integration into digital pathology workflows.

References: Lin, W., Hu, Y., Zhu, R., Wang, B., & Wang, L. (2025). Virtual staining for pathology: Challenges, limitations and perspectives. Intelligent Oncology, 1(2), 105–119. https://doi.org/10.1016/j.intonc.2025.03.005 

Non-invasive optical hormone monitoring

Author: Dr Dace Reihmane

Chronic stress and burnout are exacerbating global health crises, serving as precursors to cardiovascular and autoimmune diseases. Currently, medicine lacks a tool to observe the body’s biochemical response to stressors in real time. Clinical standards (blood/saliva tests) provide only single-point snapshots and cannot capture cortisol’s dynamic circadian rhythm. Meanwhile, consumer wearables (smartwatches) measure only secondary symptoms (heart-rate variability), which are non-specific and delayed. We need a “Holter monitor” for biomarkers – a way to continuously monitor a patient’s hormonal stress response throughout the day without invasive blood tests, similar to the glucose monitors already available on the market.

A similar approach could be applied to monitoring female sex hormones. Given their monthly fluctuations, this could enable personalised hormone-replacement therapy, predict optimal timing for conception, and, even more importantly, provide contextual data in any study (including clinical trials) to understand interactions with interventions (physical activity, medications, etc.). Continuous sex-hormone monitoring for women could be a major breakthrough in medicine and science.

Projekts Biomedicīnas un fotonikas pētniecības platforma inovatīvu produktu radīšanai (BioPhoT) tiek īstenots Ekonomikas ministrijas finansētās ilgtermiņa valsts pētījumu programmas Inovāciju fonds – ilgtermiņa pētījumu programma ietvaros. Platformas numurs: IVPP-EM-Inovācija-2024/1-0002.