Tübingen
SNS2024

Systems
Neuroscience
Symposium

26–27 September 2024

Welcome to the 2024 Tübingen Systems Neuroscience Symposium

The 2024 Tübingen Systems Neuroscience Symposium (SNS 2024) brings together leading international researchers in systems neuroscience. The symposium features plenary talks, poster sessions, lab visit and a conference dinner. Join us in Tübingen to learn about the latest advances.

September 26–27, 2024

Download Program (PDF)

Speakers

  • Wael Asaad
  • Pascal Belin
  • Eduardo Blanco-Hernandez
  • Assaf Breska
  • Sebastian Bruijns
  • Sahiti Chebolu
  • Valentin Dragoi
  • Antonino Greco
  • Tobias Hauser
  • Laurence Hunt
  • Janneke Jehee
  • Zhaoping Li
  • Romy Lorenz
  • Viola Priesemann
  • Daniela Vallentin
  • Michael Wibral
  • Tong Zhang

Scientific Program

↓ Download Program as PDF

Thursday (26 September 2024)
8:30 - 9:15 Registration & Coffee
9:15 - 9:30 Welcome address
9:30 - 10:00 Neural mechanisms of vocal learning and production in songbirds
Daniela Vallentin, München
10:00 - 10:30 On cerebral processing of voice information in primates
Pascal Belin, Marseille
10:30 - 10:50 Break
10:50 - 11:20 Predictive learning shapes the representational geometry of the human brain
Antonino Greco, Tübingen
11:20 - 11:50 Deciding when to decide: neural mechanisms underlying biased information gathering in mental health
Tobias Hauser, Tübingen
11:50 - 14:40 Lunch & Posters
14:40 - 15:10 Neurophysiology and neuromodulation of memory & attention
Wael Asaad, Providence
15:10 - 15:40 Layer-specific processing in the prefrontal cortex during working memory
Romy Lorenz, Tübingen
15:40 - 16:00 Break
16:00 - 16:30 Computations and mechanisms in neural reinforcement learning
Sahiti Chebolu & Sebastian Bruijns, Tübingen
16:30 - 17:00 Uncertainty in perceptual decision making
Janneke Jehee, Nijmegen
19:30 Dinner at Japengo
Friday (27 September 2024)
9:30 - 10:00 Neurocomputational mechanisms of dynamic predictions: cortical oscillations and cerebellar control
Assaf Breska, Tübingen
10:00 - 10:30 Learning in living networks with dendritic predictive coding
Viola Priesemann, Göttingen
10:30 - 10:50 Break
10:50 - 11:20 Information theory for the age of neural networks
Michael Wibral, Göttingen
11:20 - 11:50 Decision-making in dynamic, continuously evolving environments
Laurence Hunt, Oxford
11:50 - 14:40 Lunch & Posters
14:40 - 15:10 Sensory and brain state modulation of head direction cells
Eduardo Blanco-Hernandez, Tübingen
15:10 - 15:40 The VBC framework for how vision works in primate brain
Zhaoping Li, Tübingen
15:40 - 16:00 Break
16:00 - 16:30 From the fovea to the periphery and back: mechanisms of trans-saccadic visual information transfer in the superior colliculus
Tong Zhang, Tübingen
16:30 - 17:00 Cortical circuits for information processing and decision making
Valentin Dragoi, Houston

Symposium Venue

The symposium takes place at the University Hospital Tübingen. The lectures will be held in Lecture Hall R210, CRONA, Level 4, Hoppe-Seyler-Str. 3

Symposium Dinner

The symposium dinner takes place at Japengo, Schaffhausenstraße 113, 72072 Tübingen

Talks

Neurophysiology and Neuromodulation of Memory & Attention

Wael Asaad, Brown University, USA

On Cerebral Processing of Voice Information in Primates

Pascal Belin, Aix-Marseille University, France

The human voice carries speech but is also an ‘auditory face’ which human listeners are expert at decoding during social interactions – like many non-human primates. I will present results from behavioural, fMRI and electrophysiological studies in human and non-human primates that suggest the existence of a ‘primate voice patch system’ with a long evolutionary history.

Sensory and brain state modulation of head direction cells

Eduardo Blanco-Hernandez, University of Tübingen, Germany

Head-direction (HD) neurons are thought to contribute to hippocampal navigation and memory by providing a compass-like signal, which exclusively represents the animal's direction in space. We recorded from identified neurons in the anterior thalamus of awake mice. We found that HD neurons reliably respond to auditory and somatosensory sensory stimuli with high temporal precision, unlike non-HD neurons. The activity of HD cells, but not that of non-HD neurons, was also tightly correlated to spontaneous brain state fluctuations, showing precise coupling to pupil and whisker-pad motion. Collectively, our data show that sensory information and behavioral state modulate the HD representation pointing to a potential new role for the internal compass in relaying sensory and brain state changes into the para hippocampal memory system.

Neurocomputational mechanisms of dynamic predictions: cortical oscillations and cerebellar control

Assaf Breska, Max Planck Institute for Biological Cybernetics, Germany

The brain proactively adjusts perception, attention and action based on predicting the future unfolding of events. In dense and continuous sensory environments, it is critical to predict not only “what” will happen but also “when”. Such predictions were mostly attributed to cortical circuits, with temporal predictions specifically associated with alignment of neural oscillations. I will present work from our lab that used computational modelling, psychophysics and analysis of neural phase dynamics in healthy individuals and neurodegenerative patients to test the constraints on these models. The findings reveal the impact of oscillatory and non-oscillatory processes on low-level visual processing, and how they are utilized depending on the temporal structure and regularity of the environment. I will also show that the cerebellum is involved in sensory predictions both of when but also of what, specifically high-level semantic prediction. Together, these findings shed new light on the functional organization and neural architecture of temporal and non-temporal prediction.

Computations and mechanisms in neural reinforcement learning

Sebastian Bruijns and Sahiti Chebolu, Max Planck Institute for Biological Cybernetics, Germany

Neural reinforcement learning covers ethological/computational, psychological/algorithmic and neurobiological aspects of affective decision-making. We will illustrate the power and possibilities of this approach from two perspectives. The first (Sahiti Chebolu) provides a computational treatment of procrastination - when people fail to perform tasks in due time, to their detriment. We use reinforcement learning to provide a taxonomy of possible sources of problems that lead to such undesirable delays, and use the taxonomy to unpick the details of a substantial, real-world, dataset of students procrastinating over the course of a semester. The second perspective (Sebastian Bruijns) concerns the progressive replacement of the components of classical decision-making models with bespoke artificial neural networks in order to span the anticorrelated spectra of powerful versus interpretable accounts of animal behaviour. We apply this framework to a large dataset of mice performing a perceptual decision-making task with a hidden state inference component, detailing the sources of increased explanatory power in the increasingly flexible models.

Cortical Circuits for Information Processing and Decision Making

Valentin Dragoi, Rice University, USA

The long-range goal of my lab is to understand the mechanisms underlying state and experience-dependent changes in the function of cortical populations and how the coordination of distributed networks of neurons influences behavior. To accomplish these goals, we combine electrophysiological (multi-electrode recording in restrained and freely moving non-human primates), optogenetic and electrical stimulation, behavioral approaches, and computational methods. Our basic strategy is to help develop new tools for modulating and recording population activity across cortical circuits in restrained and unrestrained animals and then apply these techniques to examine the neural computations and coding principles across cortical circuits. My seminar will focus on recent work from my lab involving multi-electrode recordings of population activity in visual and prefrontal cortex in restrained and freely moving macaques to examine the neural network underpinnings of natural behavior. This includes information processing and coding principles in cortical networks underlying changes in behavioral performance after sleep and during complex behavior, such as foraging and social interactions.

Predictive learning shapes the representational geometry of the human brain

Antonino Greco, University of Tübingen, Germany

Predictive coding theories propose that the brain constantly updates its internal models of the world to minimize prediction errors and optimize sensory processing. However, the neural mechanisms that link the encoding of prediction errors and optimization of sensory representations remain unclear. Here, we provide direct evidence how predictive learning shapes the representational geometry of the human brain. We recorded magnetoencephalography (MEG) in human participants listening to acoustic sequences with different levels of regularity. Representational similarity analysis revealed how, through learning, the brain aligned its representational geometry to match the statistical structure of thesensory inputs, by clustering the representations of temporally contiguous and predictable stimuli. Crucially, we found that in sensory areas the magnitude of the representational shift correlated with the encoding strength of prediction errors. Furthermore, using partial information decomposition we found that, prediction errors were processed by a synergistic network of high-level associative and sensory areas. Importantly, the strength of synergistic encoding of precition errors predicted the magnitude of representational alignment during learning. Our findings provide evidence that large-scale neural interactions engaged in predictive processing modulate the representational content of sensory areas, which may enhance the efficiency of perceptual processing in response to the statistical regularities of the environment.

Deciding when to decide: neural mechanisms underlying biased information gathering in mental health

Tobias Hauser, University of Tübingen, Germany

Decisions are difficult, because in most situations, we only hold limited information about choice options, meaning that we never quite know what we may end up with. Often we can reduce this uncertainty by gathering more information. Information gathering however, also poses a challenge because collecting more information comes at expenditure of time and energy - it therefore needs careful balancing of how much information gathering is right. This is particularly problematic in mental disorders, like obsessive-compulsive disorder (OCD), where a pervasive indecisiveness is paralysing subjects' decision making. In this study, we investigate the neural and computational mechanisms underlying (biased) information gathering. Using a large smartphone-based sample (N>5000) as well as a big, clinical magnetoencephalography (MEG) sample (N>100), we show that information gathering is subject to recency-biased information integration, where most recent information is over-represented. We further show that this over-weighting of information is attenuated in OCD and along an OC spectrum, driving their indecisiveness. Using MEG decoding analyses, we show this effect to be present simultaneous in the brain and behaviour. Our finding thus show reveal the mechanisms underlying information gathering and their biases in an OC spectrum.

Decision-making in dynamic, continuously evolving environments

Laurence Hunt, University of Oxford, UK

Huge progress has been made in developing theoretical models of choice that jointly account for both behavioural and neural data. However, certain aspects of decision making that arise in naturalistic environments have been given less attention by cognitive neuroscience to date. For example, decision making in real-world settings is rarely confined to discrete trials; it contains many intermediate ‘information sampling’ decisions about what to attend to as the decision unfolds; and it often requires the decision-maker to actively navigate and explore the environment. These features all introduce unique challenges for the design and analysis of cognitive neuroscience experiments, but once they are accounted for, they can provide a new perspective on the neural representations that support adaptive behaviour. I will discuss recent neurophysiological experiments that address some of these challenges.

Uncertainty in Perceptual Decision Making

Janneke Jehee, Donders Institute, the Netherlands

Whether we are judging traffic in foggy conditions, estimating a ball’s trajectory when playing tennis, or interpreting radiological images to diagnose and treat disease – virtually every choice we make is based on uncertain evidence. How do we infer that information is more or less reliable when making these decisions? How does the brain represent knowledge of this uncertainty? In this talk, I will present recent neuroimaging data combined with novel analysis tools to address these questions. Our results indicate that sensory uncertainty can reliably be estimated from the human visual cortex on a trial-by-trial basis, and moreover that observers appear to use this uncertainty in their perceptual decision-making.

The VBC framework for how vision works in primate brain

Zhaoping Li, Max Planck Institute for Biological Cybernetics, Germany

The VBC framework is composed of: (1) The V1 Saliency Hypothesis (V1SH), (2) the attentional bottleneck, and (3) The Central-Peripheral Dichotomy theory (CPD). These three components motivate and work with each other to shape the framework for vision. The V1 Saliency Hypothesis (V1SH) holds that neural responses in primary visual cortex (V1) to visual inputs form a bottom-up saliency map of the visual field. V1SH has received convergent experimental support: e.g., V1 activity to a visual location is correlated with faster saccades to that location in monkeys (Yan, Zhaoping, Li 2018), and human gaze is strongly attracted to a location with a unique eye-of-origin of input which V1 responses would single out, even though it is not perceptually distinctive (Zhaoping 2008). Since the saliency map guides visual attention to center the attentional spotlight on the fovea, V1SH motivates the idea that the attentional bottleneck, which limits the extent of deeper processing of visual information, starts already at V1's output to downstream areas along the visual pathway. Together, V1SH and the bottleneck motivate the central-peripheral dichotomy (CPD) theory, which hypothesizes distinct roles for central and peripheral vision that should be supported by different algorithms and neural architecture (Zhaoping 2019): (1) peripheral vision is mainly for looking (guiding gaze/attentional shifts) whereas central vision is mainly for seeing (recognition); (2) top-down feedback from downstream to upstream regions along the visual pathway should mainly target central vision to aid seeing by querying for more information from upstream areas (e.g., V1). I will review recent evidence from neural, fMRI, and psychophysical data in support of this V1SH-Bottleneck-CPD (VBC) framework. I will highlight psychophysical findings from experiments that test two predictions of the VBC framework: (1) the novel reversed depth illusion, that is only, or more, visible in peripheral vision; and (2) this illusion nevertheless becomes visible in central vision when top-down feedback is compromised by backward masking. I will show how the VBC is related to but distinct from some classical and modern ideas, and how it could guide us for understanding vision by a hierarchical network of brain areas including V1, superior colliculus, and cortical areas beyond V1.

Layer-specific processing in the prefrontal cortex during working memory

Romy Lorenz, Max Planck Institute for Biological Cybernetics, Germany

Although working memory (WM) reliably activates the dorsolateral prefrontal cortex (dlPFC), the functional significance of its distinct cytoarchitectonic layers is not well understood in humans. A recent 7T fMRI study at ultrahigh-resolution provided the first evidence of layer-specific responses in the human dlPFC during WM, revealing that superficial layers were more active during the manipulation of WM, while deep layers exhibited increased activity during motor responses to the probe. To assess the replicability of these findings, we conducted a pre-registered replication of this study using a fully automated and reproducible analysis pipeline. While we confirmed the preferential involvement of superficial layers during WM manipulation, our results did not replicate the original findings of deep layer activation during motor responses. In a subsequent study, we demonstrated that the superficial layer effect extends to increases in WM load, potentially pointing towards a lamina-specific activation of the frontoparietal network to heightened task demands more generally. Additionally, multivariate analyses revealed that superficial layers of the dlPFC are involved in various WM subprocesses by dynamically adapting to current task demands. This suggests altered network coupling in the superficial layers as WM demands shift, aligning with their established intra-prefrontal and cortico-cortical connections. Finally, we replicated our earlier null result, showing that both layers in the dlPFC are non-differentially involved in motor responses. Our work underscores the importance of methodological rigor and reproducibility in layer-specific fMRI research, highlights the more complex role of deep layers in WM than previously understood, and contributes to a more nuanced understanding of dlPFC function during WM.

Learning in Living Networks with Dendritic Predictive Coding

Viola Priesemann, Max Planck Institute for Dynamics and Self-Organization, Germany

Living systems hinge on efficient information processing. Importantly, efficient representations can be learned through local, unsupervised learning rules. We show how such learning can already be implemented by dendritic branches. By deriving learning rules a priori, we demonstrate that on the dendritic branch, the deviation from excitation-inhibition balance equals the encoding error, and that this error signal can directly determine synaptic plasticity. Furthermore, we illustrate the versatile role of homeostatic plasticity in tuning network performance. Overall, the resulting unsupervised, local learning rules enable the learning of an efficient representation via "dendritic predictive coding".

Neural mechanisms of vocal learning and production in songbirds

Daniela Vallentin, Max Planck Institute for Biological Intelligence, Germany

Learning and execution of complex motor skills are often modulated by sensory feedback and contextual cues arriving across multiple sensory modalities. Vocal motor behaviors are primarily influenced by auditory inputs, both during learning and mature vocal production. The importance of auditory input in shaping vocal output has been investigated in several songbird species that acquire their adult song based on auditory exposure to a tutor during development. We explored song imitation in juvenile zebra finches raised either in the presence or absence of females providing vocal feedback. We found that male zebra finches raised with a female copied the spectral and temporal features of the tutor song more accurately than compared to birds, that were raised socially isolated. We found that females emitted more calls as young birds improved their song performance, indicating that females can provide practice-specific feedback. To decipher whether female vocal feedback has an impact on the neural activity within the song learning pathway, we performed intracellular recordings in HVC, a premotor area involved in song learning and production, in singing and listening zebra finches. In juvenile zebra finches, we found that female vocalizations can modulate neural activity in HVC during passively listening and singing. These results highlight the contribution of female vocal feedback to developmental song learning and how vocal input other than the tutor song can influence the neural circuit involved in song learning and production. Once the bird reaches adulthood this song remains stable. We discovered that inhibition within the premotor area HVC plays a major role in closing this critical period by suppressing the influence of the tutor once song proficiency has been achieved. We then developed a cell-type specific viral strategy to target inhibitory neurons in adult zebra finches and were able to re-open the critical period by teaching an adult zebra finch novel song elements. This finding might have important implications to understand and expand motor skill learning capabilities or improve sensory and motor recovery after injury.

Information theory for the age of neural networks

Michael Wibral, University of Göttingen, Germany

Since the late 1950's the concepts of redundant synergistic coding in neural systems have gathered intense and sustained interest. In redundant coding multiple neurons code for information such that this information can be obtained from any of them, while in synergistic coding the encoded information can only be recovered by considering all neurons jointly. As one coding scheme provides robustness while the other maximizes efficiency in terms of the neruons needed for an encoding task, both seem to have a place in neural systems.

While the presence and use of these coding schemes in various neural systems, thus, seems almost self-evident, the question of which coding type dominated in which neural system has been fiercely debated at times. This confusion was only resolved recently when Williams and Beer showed in their seminal work that the concepts of redundant and synergistic information in a system had been lacking a consistent mathematical definition free of internal contradictions.

Williams' and Beer's work further showed that there is one more important type of information besides redudant and synergistic information, i.e. unique information, that can only be obtained from a certain neuron or part of the system. Together redudant, synergistic and unique information form a full decomposition of the mutual information between a set of source variables (e.g. a bunch of neurons) and a target variable (e.g. a stimulus to be decoded). This decomposition is called a partial information decomposition (PID). Since this a firm mathematical foundation for PID has been etsablished, it has seen a rapid surge of interest and of applications in neuroscience resulting in several high rank publications, also in the field of EEG and MEG. Due to the mathematical difficulty and intricacy of the PID concept, however, there have been some doubts with respect to its usefulness as a tool to understand neural computations.

We here show that indeed the concepts of PID are extremely useful to describe neural computation in general, even beyond neural coding problems. We do so in a constructive way by defining an intuitively understandable local goal function for neurons based on PID. This goal function is then used to train artifical neurons that model biological layer 5 pyramidal neurons. We demonstrate that this local goal function in combination with the artificical layer 5 neurons is sufficient to build networks that solve clustering, assocviative memory and classification problems surprisingly well, i.e. surpassing hopfield networks and having a performance on par with equally-sized networks trained globally with backpropagation. At the same time the local neural goal functions are intuitively understandable, and can be tracked neuron-per-neuron and class-per-class over the course of learning.

In our view, this demonstrates that indeed PID offers a powerful, yet intuitively understandable, new way to analyze, understand and construct neural computation.

From the fovea to the periphery and back: mechanisms of trans-saccadic visual information transfer in the superior colliculus

Tong Zhang, University of Tübingen, Germany

The superior colliculus (SC) possesses visual machinery supporting both foveal analysis and peripheral object detection. This structure also emits movement-related discharge that is relayed to both the downstream oculomotor control network and upstream cortical areas. This places the SC in an ideal position to both control orienting responses as well as bridge periods of sensory uncertainty associated with rapid eyeball rotations. Yet, the mechanisms with which foveal state influences peripheral visual sensitivity, or with which peripheral visual information is trans-saccadically relayed to foveal SC visual representations, are not fully understood. Here we will first describe how foveal SC state can have a strong impact on peripheral SC visual sensitivity. Using real-time gaze-contingent image control of instantaneous foveal eye position error, we will demonstrate how a foveal error of only 1-2 min arc is sufficient to not only drive microsaccades, but also strongly influence peripheral SC visual responses and orienting response efficiency. We will then show how SC movement-related discharge is itself not a pure neuronal movement command, but instead represents the sensory state of the periphery at the time of saccades. Thus, SC “motor” bursts not only represent where gaze will shift towards, but they also provide a peripheral preview of the visual appearance of saccade targets. Once these targets are foveated, our final set of results will demonstrate how foveal SC visual representations are predictively sensitive to pre-saccadic peripheral target appearance. Thus, the SC encompasses a full active vision loop, from the fovea to the periphery and back.

Registration

Deadline extended until September 22, 2024

Please use the contact form to contact the organizers

Registration fees:

  • Regular: €200
  • Student: €150
  • Members of the “Neurowissenschaftliche Gesellschaft” receive a discount of € 25
  • For students of the Graduate Training Center Tübingen (PhD and MSc), the registration fee is fully covered by the Graduate School (no transfer required).
  • Members/employees of the Max Planck Institutes in Tübingen will have their registration paid directly by the MPI. In this case, please do not transfer the registration fee yourself.

The registration fee includes lunch on both conference days and the conference dinner.

Bank Account

Please transfer the registration fee to:

Recipient: Universitätsklinikum Tübingen
IBAN: DE41 6005 0101 7477 5037 93
BIC: SOLADEST600
Bank: Baden-Wuerttembergische Bank Stuttgart

Payment reference (Verwendungszweck): D.33.07956 - Neuroscience Symposium 24 YOUR NAME

Tübingen

Tübingen is a city at the heart of Baden-Württemberg in the South of Germany. It lies at the banks of the Neckar river, 30 km South of Stuttgart, and pleases the eyes of its visitors with its beautiful old town, the castle Hohentübingen and the famous Neckarfront with the tower of poet Hölderin. Tübingen is home to one of Germany’s best universities, the Eberhard Karls University which was founded in 1477. With around 30,000 students at only 85,000 inhabitants, life in Tübingen is dominated by students and academics.

Neuroscience in Tübingen

Neuroscience is a major research focus in Tübingen. Neuroscience research is conducted not only at the university and the university hospital (the ‘UKT’), which is home of the Tübingen MEG center, but also in several other research institutions such as the Werner Reichardt Center for Integrative Neuroscience (CIN), the Hertie Institute for Clinical Brain Research (HIH), the Bernstein Center for Computational Neuroscience Tübingen (BCCN), the German Center for Neurodegenertive Diseaeses (DZNE), the German Center for Diabetes Research (DZD) and the Max Planck Institutes for Intelligent Systems and for Biological Cybernetics.

Young scientists are an important part of the academic workforce in the Tübingen neuroscience community. Their support, advancement and education is fostered by the Graduate Training Center of Neuroscience, with over 300 enrolled master and PhD students.

Restaurants

Fun facts

Did you know that Tübingen …

  • … is the youngest city in Germany? The mean age was 39.0 years on December 31, 2011.
  • … contains the geographic center of Baden-Württemberg? It is marked by a stone, not far from the campus of the university hospital.

Travel & Accomodation

How to get to Tübingen?

The university town of Tübingen, geographically the center of Baden-Württemberg at 48°31’17” northern latitude and 9°3’16” eastern longitude, is situated within the beautiful area between the Alb foothills and Schönbuch in the Neckar valley.

The „town of short distances“ offers numerous possibilities for parking at the border of the Old Town. Shops and sights, municipal offices and institutions are mostly located in centre; further destinations are easy to access with public transportation. Shared taxis and night buses make a nightlife without car possible. A well-developed network of bike paths makes it easy to get around in town. At specific bus stops it is possible to take bikes along in the bus.

Railway

Tübingen serves as a railway junction with connections to Stuttgart, Black Forest, the Swabian Alb and the region of Lake Constance. From Stuttgart central station it takes just 55 minutes to the centre of Tübingen.

Plane

From Stuttgart airport by car via B27 in 25 minutes to the centre of Tübingen, or by airport shuttle, „Airport Sprinter“ No. 828, in 50 minutes to Tübingen’s bus terminal. The bus runs hourly from the central bus stop in front of Terminal 1 of the arrival section.

Car

From the autobahn A8 (München-Stuttgart) via B27 in 30 minutes; from autobahn A81 (Singen-Heilbronn) via B28 in 25 minutes. Tübingen contains a Green Zone, where environmental badges are obligatory for every vehicle. This measure was taken to cut down the load of particulate matter.

Local transport association (German Website)
www.naldo.de

Busses and cabs in Tübingen  (German Website)
www.svtue.de

Stuttgart Airport
www.flughafen-stuttgart.de

Intercity busses

Germany has a lot of interesting cities that are connected with the university town of Tübingen via daily intercity busses. Just follow the links to see schedules or buy tickets online.

Destination Provider
Augsburg www.muenchenlinie.de
Bayreuth www.flixbus.de
Berlin www.flixbus.de
www.meinfernbus.de
Europa-Park Rust www.meinfernbus.de
Frankfurt am Main www.deinbus.de
Freiburg www.deinbus.de
www.meinfernbus.de
München www.deinbus.de
www.muenchenlinie.de
Nürnberg www.flixbus.de
www.meinfernbus.de
Heidelberg www.deinbus.de
Heilbronn www.flixbus.de
www.deinbus.de
Karlsruhe www.meinfernbus.de
Konstanz www.deinbus.de
Lörrach www.meinfernbus.de
Singen www.deinbus.de
Stuttgart www.meinfernbus.de
Villingen-Schwenningen www.deinbus.de
Where is Tübingen?

Where is Tübingen?

Source: http://www.tuebingen.de

Where to stay?

Hotels and private rooms in Tübingen (click here)

Contact us