Biology subjects

Boedecker, J.

Publications and source records attributed to Boedecker, J..

8 recordsLinked to original sources

bioRxiv2026

Evaluating Transferability and Robustness of Process-Guided Neural Networks in Forest Carbon Flux Modelling

Making robust and generalizable predictions within ecological systems such as forests remains challenging due to limited data availability and the slow pace of environmental change. To address this, we integrate a semi-empirical environmental process model (PRELES) to support deep learning approaches, specifically artificial neural networks (ANNs). We replicate and extend previous work on process-guided neural networks (PGNN) by introducing new model types and conducting a comprehensive hyperparameter optimisation within systematic nested cross-validation analyses in both data-thinning and extrapolative scenarios. Results show that both data-driven ANNs and PGNNs consistently outperform the stand-alone process model, while PGNNs provide additional advantages over ANNs in data-sparse settings and under transfer scenarios to unseen, changing climatic conditions. We further estimate the generalisation error for data-driven models as a function of the amount of training data, allowing for guidance on model suitability under different data availability. A variable importance analysis using accumulated local effects reveals that both PGNNs and ANNs learn simple, physically plausible relationships, whereas PRELES exhibits a strong bias toward boreal conditions and limited ability to predict unseen, climatically divergent sites. HighlightsO_LIProcess-guided, and plain neural networks outperform a calibrated process-based model (PRELES) in predicting forest ecosystem carbon fluxes. C_LIO_LIProcess-guided neural networks provide advantages over naive neural networks in sparse-data settings and show greater robustness under transferable scenarios with unseen changing climatic conditions. C_LIO_LIVariable-importance analyses using accumulated local effects show that both process-guided and naive neural networks learn simple yet physically plausible relationships between meteorological drivers and target responses, whereas the process model (PRELES) exhibits a better fit toward boreal conditions and limited ability to predict unseen, climatically divergent sites. C_LI

ecology↗

bioRxiv2025

Efficient Gaussian Process-based Motor Hotspot Hunting with Concurrent Optimization of TMS Coil Location and Orientation

BACKGROUNDTranscranial magnetic stimulation (TMS) is a widely used non-invasive brain stimulation technique in neuroscience research and clinical applications. TMS-based motor hotspot hunting, the process of identifying the optimal scalp location to elicit robust and reliable motor responses, is critical to ensure reproducibility and efficiency, as well as to determine safe and precise stimulation intensities in both healthy participants and patients. Typically, this process targets motor responses in contralateral short hand muscles. However, hotspot hunting remains challenging due to the vast parameter space and time constraints. OBJECTIVETo address this, we present an approach that concurrently optimizes both spatial and angular TMS parameters for hotspot hunting using Gaussian processes and Bayesian optimization. METHODSWe systematically evaluated five state-of-the-art acquisition functions on electromyographic TMS data from eight healthy individuals enhanced by simulated data from generative models. RESULTSOur results consistently demonstrate that optimizing spatial and angular TMS parameters simultaneously enhances the efficacy and spatial precision of hotspot hunting. Furthermore, we provide mechanistic insights into the acquisition function behavior and the impact of coil rotation constraints, revealing critical limitations in current hotspot-hunting strategies. Specifically, we show that arbitrary constraints on coil rotation angle are suboptimal, as they reduce flexibility and fail to account for individual variability. We further demonstrate that acquisition functions differ in sampling strategies and performance. Functions overly emphasizing exploitation tend to converge prematurely to local optima, whereas those balancing exploration and exploitation--particularly Thompson sampling--achieve superior performance. CONCLUSIONThese findings highlight the importance of acquisition function selection and the necessity of removing restrictive coil rotation constraints for effective hotspot hunting. Our work advances TMS-based hotspot identification, potentially reducing participant burden and improving safety in both research and clinical applications beyond the motor cortex.

neuroscience↗

bioRxiv2025

Efficient Prospective Electric Field-Informed Localization of Motor Cortical Targets of Transcranial Magnetic Stimulation

Transcranial magnetic stimulation (TMS) is a versatile non-invasive tool for brain mapping and neu-romodulation in both healthy individuals and patients. Effective TMS-based causal brain mapping relies on precise localization of cortical targets. Current state-of-the-art approaches use statistical methods to quantify the relationship between TMS-induced electric fields (E-fields) and motor evoked potential (MEP) amplitudes. However, this method typically relies on the random selection of coil configurations, which limits its efficacy. In this study, we present a novel optimization strategy for TMS-based motor mapping by prospectively selecting coil configurations based on their E-field characteristics using an iterative sampling algorithm called farthest point sampling (FPS). Through a combination of theoretical analysis, simulation and experimental validation including 10 healthy individuals, we systematically evaluated the performance of FPS against the random sampling approach. Our results demonstrate that FPS is twice as efficient as random sampling in reducing the number of trials required for estimating the motor map, while also being more robust across participants and less susceptible to noise. These findings highlight the potential of FPS to significantly enhance the efficiency of motor mapping, paving the way for the development of more effective TMS mapping algorithms.

neuroscience↗

bioRxiv2025

Sleep Resolves Competition Between Explicit and Implicit Memory Systems

Sleep supports stabilization of explicit, declarative memory and benefits implicit statistical learning. In addition, sleep may change the quality of memory representations. Explicit and implicit learning systems, usually linked to hippocampus and striatum, can compete during learning, but whether they continue to interact during offline periods remains unclear. Here, we investigate for feedback-driven classification learning, whether sleep integrates explicit exemplar memory and implicit information-integration. The negative relationship between implicit and explicit memory components was resolved over sleep, but not wakefulness. Additionally, sleep benefitted performance on a categorization task that allows the use of both explicit and implicit strategies, and participants who slept showed superior performance in generalizing their knowledge to unseen exemplars. A reinforcement learning model relates this to better transfer of the learned exemplar value representation after sleep. Together, this suggests that sleep allows flexible access to information learned by different routes to respond optimally to everyday life contingencies.

neuroscience↗

bioRxiv2024

Enhancement of network architecture alignment in comparative single-cell studies

1Animal data can provide meaningful context for human gene expression at the single-cell level. This context can improve cell-type or cell-state detection and clarify how well the animal models human biological processes. To achieve this, we propose a deep learning approach that identifies a unified latent space to map complex patterns between datasets. Specifically, we combine variational autoencoders with a data-level nearest neighbor search to align neural network architectures across species. We visualize commonalities by mapping cell samples into the latent space. The aligned latent representation facilitates information transfer in applications of liver, white adipose tissue, and glioblastoma cells from various animal models. We also identify genes that exhibit systematic differences and commonalities between species. The results are robust for small datasets and with large differences in the observed gene sets. Thus, we reliably uncover and exploit similarities between species to provide context for human single-cell data.

bioinformatics↗

bioRxiv2023

Mechanisms of Premotor-Motor Cortex Interactions during Goal Directed Behavior

Deciphering the neural code underlying goal-directed behavior is a long-term mission in neuroscience1,2. Neurons exhibiting preparation and movement-related activity are intermingled in the premotor and motor cortices3,4, thus concealing the neural code of planned movements. We employed a combination of electrophysiology, pathway-specific optogenetics, phototagging, and inverse reinforcement learning (RL) to elucidate the role of defined neuronal subpopulations in the rat rostral and caudal forelimb areas (RFA and CFA), which correspond to the premotor and motor cortical areas. The inverse RL enabled the functional dissection of spatially intermingled neuronal subpopulations, complementing our pathway-specific optogenetic manipulations and unveiling differential functions of the preparation and movement subpopulations projecting from RFA to CFA. Our results show that the projecting preparation subpopulation suppresses movements, whereas the projecting movement subpopulation promotes actions. We found the influence of RFA on CFA to be adaptable, with the projection either inhibiting or exciting neurons in the superficial and deep CFA layers, depending on context and task phase. These complex interactions between RFA and CFA likely involve the differential recruitment of inhibitory interneurons in the CFA, which is supported by our electron microscopy analysis of the connectivity between these regions. We provide here unprecedented mechanistic insights into how the premotor and primary motor cortices are functionally and structurally interlinked with the potential to advance neuroprosthetics. Graphical abstractThis study provides mechanistic insights into the interactions between the rostral forelimb area (RFA) and the caudal forelimb area (CFA). Specifically, we provide evidence for a differential impact of RFA on CFA depending on the task phase and the targeted CFA layers. RFA contains at least two spatially intermingled subpopulations - one related to movement preparation and one to movement execution. Both subpopulations project to CFA. Here we investigated the impact of these two subpopulations on the activity of the local CFA circuit as well as on the behavior in different contexts. When rats were not involved in a task, the effect of RFA was mainly excitatory in the deep CFA layers, while the superficial layers remained unaffected. This can be interpreted as a non-selective activation of the deep CFA neurons enabling a variety of spontaneous movements. During the preparation phase before a movement, the RFA had an opposite impact on the superficial and deep layers: while the superficial CFA layers were excited by RFA input, the deeper layers were mostly inhibited, minimizing movements and enabling continued holding of a lever. During the movement phase, the inhibitory effect on neurons in the deep CFA layers was counterbalanced by excitation, thus enabling a selected conduction of movements. The opposing effects during preparation and movement phase on CFA deep layers were correlated with increased firing rates of the RFA preparation and movement subpopulations, respectively, making it likely that the inhibition resulted from increased activities of these subpopulation specifically. With an electron microcopy approach we demonstrate that inhibitory and excitatory CFA neurons are directly targeted by RFA, thus providing a mechanism for the bidirectional control of CFA activity. Please note that the depicted impact of RFA on excitatory or inhibitory CFA neurons refers to net effects in this figure, not to the targeting of individual neurons. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=109 SRC="FIGDIR/small/524944v1_ufig1.gif" ALT="Figure 1"> View larger version (23K): org.highwire.dtl.DTLVardef@4de2d5org.highwire.dtl.DTLVardef@167164forg.highwire.dtl.DTLVardef@e9d298org.highwire.dtl.DTLVardef@101320b_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2022

Process-guidance improves predictive performance of neural networks for carbon turnover in ecosystems

Despite deep-learning being state-of-the-art for data-driven model predictions, it has not yet found frequent application in ecology. Given the low sample size typical in many environmental research fields, the default choice for the modelling of ecosystems and its functions remain process-based models. The process understanding coded in these models complements the sparse data and neural networks can detect hidden dynamics even in noisy data. Embedding the process model in the neural network adds information to learn from, improving interpretability and predictive performance of the combined model towards the data-only neural networks and the mechanism-only process model. At the example of carbon fluxes in forest ecosystems, we compare different approaches of guiding a neural network towards process model theory. Evaluation of the results under four classical prediction scenarios supports decision-making on the appropriate choice of a process-guided neural network. O_TEXTBOXSignificance StatementDeep-learning is the state-of-the-art for data-driven model predictions. Given the low sample size typical in many environmental research fields, these approaches can rarely be applied. When data are complemented by process understanding, as coded in physical or empirical models, both predictions and generalisations can be substantially improved over both data-only neural networks and mechanism-only process models. Comparing different approaches of such process-guidance helps decide on how to best combine process models and neural networks. C_TEXTBOX

ecology↗

bioRxiv2018

Deep Learning Based BCI Control of a Robotic Service Assistant Using Intelligent Goal Formulation

As autonomous service robots become more affordable and thus available for the general public, there is a growing need for user-friendly interfaces to control these systems. Control interfaces typically get more complicated with increasing complexity of the robotic tasks and the environment. Traditional control modalities as touch, speech or gesture commands are not necessarily suited for all users. While non-expert users can make the effort to familiarize themselves with a robotic system, paralyzed users may not be capable of controlling such systems even though they need robotic assistance most. In this paper, we present a novel framework, that allows these users to interact with a robotic service assistant in a closed-loop fashion, using only thoughts. The system is composed of several interacting components: non-invasive neuronal signal recording and co-adaptive deep learning which form the brain-computer interface (BCI), high-level task planning based on referring expressions, navigation and manipulation planning as well as environmental perception. We extensively evaluate the BCI in various tasks, determine the performance of the goal formulation user interface and investigate its intuitiveness in a user study. Furthermore, we demonstrate the applicability and robustness of the system in real world scenarios, considering fetch-and-carry tasks and tasks involving human-robot interaction. As our results show, the system is capable of adapting to frequent changes in the environment and reliably accomplishes given tasks within a reasonable amount of time. Combined with high-level planning using referring expressions and autonomous robotic systems, interesting new perspectives open up for non-invasive BCI-based human-robot interactions.

neuroscience↗