Biology subjects

Ruge, H.

Publications and source records attributed to Ruge, H..

8 recordsLinked to original sources

bioRxiv2026

Rapid connectivity alterations of thalamic nuclei during initial learning of goal-directed behaviour

The thalamus is essential for learning, dynamically engaging with other subcortical and cerebral cortex regions throughout the learning process. Here, the thalamus serves as a critical connector hub and synchroniser within the thalamocortical system of the brain. However, whilst higher order thalamic nuclei are known to be particularly important for this process, the exact contributions of individual higher order and first order thalamic nuclei, alongside their individual involvement with cortical networks and subcortical regions, remains unexplored within the initial phase of learning. In light of this, we analysed fMRI data obtained within a paradigm which is designed to examine initial learning processes within feedback-driven stimulus-response learning, in order to explore thalamic contributions. We investigated dynamic learning-related functional connectivity alterations between various thalamic nuclei with other subcortical regions and cortical networks. Our results show that the initial phase of learning was associated with: (1) decreasing functional connectivity between thalamic nuclei and frontoparietal and cingulo-opercular networks, (2) increasing functional connectivity between thalamic nuclei with default mode and salience networks, (3) decreasing functional connectivity between thalamic nuclei and the putamen, and (4) decreasing functional connectivity amongst higher order thalamic nuclei. Furthermore (5) these dynamic alterations were associated primarily by mediodorsal thalamus. Altogether, these results indicate that higher order thalamic nuclei play a crucial role within initial learning and in the generation of novel goal-directed behaviour. This was demonstrated through enhanced functional connectivity with selected cortical networks which drive goal-directed behaviour, alongside decreased functional connectivity with striatal regions which drive motor selectivity.

neuroscience↗

bioRxiv2025

Encoding neural representations of time-continuous stimulus-response transformations in the human brain with advanced deep neural networks

Human behavior arises from the continuous transformation of sensory input into goal-directed actions. While existing analytical methods often break time into discrete events, the stages and underlying representations involved in stimulus-response (S-R) transformations within time-continuous, complex environments remain only partially understood. Encoding models, combined with deep neural networks (DNNs) for feature generation, offer a promising framework for capturing these neural processes. While DNNs continue to improve in performance, it remains unclear whether these advances translate into more accurate models of brain activity. To address this, we collected fMRI data from participants (N = 23) as they played arcade-style video games and applied DNN-based encoding models to predict voxel-level brain activity. We compared the prediction accuracy of features from three DNNs at different stages of development within our encoding model. We show that the most advanced DNN provides the most predictive feature space for neural responses, while also exhibiting a closer hierarchical alignment between its internal representations and the brains functional organization. These results enable a more fine-grained characterization of time-continuous S-R transformations in high-dimensional visuomotor tasks, progressing along the dorsal visual stream and extending into motor-related regions. This approach highlights the potential of machine learning to advance cognitive neuroscience by enhancing the ecological validity of experimental tasks.

neuroscience↗

bioRxiv2025

Advances in deep reinforcement learning enable better predictions of human behavior in time-continuous tasks

Humans have to respond to everyday tasks with goal-directed actions in complex and time-continuous environments. However, modeling human behavior in such environments has been challenging. Deep Q-networks (DQNs), an application of deep learning used in reinforcement learning (RL), enable the investigation of how humans transform high-dimensional, time-continuous visual stimuli into appropriate motor responses. While recent advances in DQNs have led to significant performance improvements, it has remained unclear whether these advancements translate into improved modeling of human behavior. Here, we recorded motor responses in human participants (N=23) while playing three distinct arcade games. We used stimulus features generated by a DQN as predictors for human data by fitting the DQNs response probabilities to human motor responses using a linear model. We hypothesized that advancements in RL models would lead to better prediction of human motor responses. Therefore, we used features from two recently developed DQN models (Ape-X and SEED) and a third baseline DQN to compare prediction accuracy. Compared to the baseline DQN, Ape-X and SEED involved additional structures, such as dueling and double Q-learning, and a long short-term memory, which considerably improved their performances when playing arcade games. Since the experimental tasks were time-continuous, we also analyzed the effect of temporal resolution on prediction accuracy by smoothing the model and human data to varying degrees. We found that all three models predict human behavior significantly above chance level. SEED, the most complex model, outperformed the others in prediction accuracy of human behavior across all three games. These results suggest that advances in deep reinforcement learning can improve our capability to model human behavior in complex, time-continuous experimental tasks at a fine-grained temporal scale, thereby opening an interesting avenue for future research that complements the conventional experimental approach, characterized by its trial structure and use of low-dimensional stimuli. Author summaryIn our complex environment, we are constantly responding to outside influences. Traditional trial-based psychological experiments have limitations because they cannot account for the continuous nature of the world. In this study, we combine psychological questions with advanced technology to investigate how humans behave over time. Artificial neural networks have not only proven they can outperform humans in complex tasks but also have the ability to generate features that can be used to model human behavior. We investigated whether technological advancements in deep neural networks would lead to higher accuracy in predicting human motor responses recorded from subjects while playing three arcade games. We compared the predictive performance of features generated by three neural networks of varying complexity. Our results provide evidence that all three models can predict human behavior, with the most advanced model achieving the highest prediction accuracy. These results suggest that advanced neural networks are suitable models for studying human behavior in a time-continuous context and serve as a complement to the trial-based study designs.

bioinformatics↗

bioRxiv2024

Progressive changes in functional connectivity between thalamic nuclei and cortical networks across learning

The thalamus is connected to cerebral cortex and subcortical regions, serving as a node within cognitive networks. It is a heterogeneous structure formed of functionally distinct nuclei with unique connectivity patterns. However, their contributions to cognitive functioning within networks is poorly understood. Recent animal research suggests that thalamic nuclei such as the mediodorsal nucleus play critical roles in goal-directed behaviour. Our aim was to investigate how functional integration of thalamic nuclei within cortical and subcortical networks changes whilst transitioning from more controlled goal-directed behaviour towards more automatic or habitual behaviour in humans. We analysed functional magnetic resonance imaging (fMRI) data from a stimulus-response learning study to investigate functional connectivity (FC) changes across learning between thalamic nuclei with cortical networks and subcortical structures in healthy subjects. We defined subcortical regions-of-interest (ROIs) individually in native space, segmenting the thalamus into 47 nuclei and segmenting 38 subregions within the basal ganglia and hippocampus. Additionally, we defined 12 cerebral cortex ROIs via maximum-probability network templates. Learning-related connectivity changes were examined via ROI-to-ROI functional network analysis. Our results showed that learning was associated with: 1) decreasing FC between the frontoparietal network and higher order thalamic nuclei; 2) increasing FC between the cingulo-opercular network and pulvinar nuclei, 3) decreasing FC between the default mode network (DMN) and right mediodorsal nuclei; 4) increasing FC between the DMN and left mediodorsal nuclei; 5) altered functional connectivity between thalamic nuclei and putamen subregions, and 6) increasing intrathalamic FC. Together, this suggests that several thalamic nuclei are involved in the learning-related transition from controlled to more automatic behaviour.

neuroscience↗

bioRxiv2024

Initial Learning in the Brain: From Rules to Action

We used fMRI to investigate the neural changes and representational dynamics associated with different learning modes during initial learning and subsequent implementation of previously acquired stimulus-response (S-R) associations. We compared instruction-based learning (INS) and trial-and-error learning (TE) via a third observation-based learning (OBS) condition. This was yoked to the TE condition and shared features with both, the INS and TE conditions. During learning, neural changes were observed in the Frontoparietal and Default Mode Networks across learning modes, consistent with a general decrease in cognitive control demand as learning progresses. INS and TE exhibited condition-specific signal changes, which we interpreted in the context of covert motor preparation during INS, and intentional action and increased cognitive control demand during early TE trials, respectively. Multivariate pattern analysis revealed individual rule information in bilateral prefrontal, premotor, and parietal cortices across learning modes. Most regions revealed consistent representations of individual S-R rules between the learning stage and subsequent implementation stage, regardless of the learning mode. This suggests that initially formed S-R rule representations guide task performance during S-R rule implementation, irrespective of how they are acquired. Finally, within the primary motor cortex, individual S-R rules were decodable during the learning stage not only when motor responses were overtly executed, as in TE, but also in the absence of overt motor execution, as in INS. This finding substantiates previous claims of covert motor preparatory mechanisms during INS.

neuroscience↗

bioRxiv2020

Fast estimation of L1-regularized linear models in the mass univariate setting

In certain modeling approaches, activation analyses of task-based fMRI data can involve a relatively large number of predictors. For example, in the encoding model approach, complex stimuli are represented in a high-dimensional feature space, resulting in design matrices with many predictors. Similarly, single-trial models and finite impulse response models may also encompass a large number of predictors. In settings where only few of those predictors are expected to be informative, a sparse model fit can be obtained via L1-regularization. However, estimating L1-regularized models requires an iterative fitting procedure, which considerably increases computation time compared to estimating unregularized or L2-regularized models, and complicates the application of L1-regularization on whole-brain data and large sample sizes. Here we provide several functions for estimating L1-regularized models that are optimized for the mass-univariate analysis approach. The package includes a parallel implementation of the coordinate descent algorithm for CPU-only systems and two implementations of the alternating direction method of multipliers algorithm requiring a GPU device. While the core algorithms are implemented in C++/CUDA, data input/output and parameter settings can be conveniently handled via Matlab. The CPU-based implementation is highly memory-efficient and provides considerable speed-up compared to the standard implementation not optimized for the mass-univariate approach. Further acceleration can be achieved on systems equipped with a CUDA-enabled GPU. Using the fastest GPU-based implementation, computation time for whole-brain estimates can be reduced from 9 hours to 5 minutes in an exemplary data setting. Overall, the provided package facilitates the use of L1-regularization for fMRI activation analyses and enables an efficient employment of L1-regularization on whole-brain data and large sample sizes.

neuroscience↗

bioRxiv2019

Initial neural representation of individual rules after first-time instruction

By following explicit instructions humans can instantaneously get the hang of tasks they have never performed before. Here, we used a specially calibrated multivariate analysis technique to uncover the elusive representational states following newly instructed arbitrary behavioural rules such as for coffee, press red button, while transitioning from knowing what to do to actually doing it. Subtle variation in distributed neural activity patterns reflected rule-specific representations within the ventrolateral prefrontal cortex (VLPFC), confined to instructed stimulus-response learning in contrast to incidental learning involving the same stimuli and responses. VLPFC representations were established right after first-time instruction and remained stable across early implementation trials. More and more fluent application of novel rule representations was channelled through increasing cooperation between VLPFC and anterior striatum. These findings inform representational theories on how the prefrontal cortex supports behavioural flexibility by enabling ad-hoc coding of novel task rules without recourse to familiar sub-routines

neuroscience↗

bioRxiv2018

Deterministic response strategies in trial-and-error learning

Trial-and-error learning is a universal strategy for establishing which actions are beneficial or harmful in new environments. However, learning stimulus-response associations solely via trial-and-error is often suboptimal, as in many settings dependencies among stimuli and responses can be exploited to increase learning efficiency. Previous studies have shown that in settings featuring such dependencies, humans typically engage high-level cognitive processes and employ advanced learning strategies to improve their learning efficiency. Here we analyze in detail the initial learning phase of a sample of human subjects (N = 85) performing a trial-and-error learning task with deterministic feedback and hidden stimulus-response dependencies. Using computational modeling, we find that the standard Q-learning model cannot sufficiently explain human learning strategies in this setting. Instead, newly introduced deterministic response models, which are theoretically optimal and transform stimulus sequences unambiguously into response sequences, provide the best explanation for 50.6% of the subjects. Most of the remaining subjects either show a tendency towards generic optimal learning (21.2%) or at least partially exploit stimulus-response dependencies (22.3%), while a few subjects (5.9%) show no clear preference for any of the employed models. After the initial learning phase, asymptotic learning performance during the subsequent practice phase is best explained by the standard Q-learning model. Our results show that human learning strategies in trial-and-error learning go beyond merely associating stimuli and responses via incremental reinforcement. Specifically during initial learning, high-level cognitive processes support sophisticated learning strategies that increase learning efficiency while keeping memory demands and computational efforts bounded. The good asymptotic fit of the Q-learning model indicates that these cognitive processes are successively replaced by the formation of stimulus-response associations over the course of learning.

animal behavior and cognition↗