Start

Welcome

The Neural Data Science group led by Alexander Ecker works at the interface of machine learning and computational neuroscience. We develop new methods and algorithms to make sense of large-scale neuroscience data. Moreover, we work on novel approaches to computer vision based on insights we gain from biological vision.

Currently, the Neural Data Science Group has 10 PhD students and 4 postdocs.

Data Science is a highly collaborative endeavour. We work closely with a number of labs in Göttingen, Germany and the United States:

Teaching

Summer term 2024

Practical course on applying deep learning for image generation.

Alexander Ecker and Timo Lüddecke

Winter term 2023/2024

Seminar where recent deep learning papers are presented and discussed.

Alexander Ecker, Michaela Vystrčilová, Richard Vogg

Introduction to Graph Machine Learning

Martin Ritzert and Alexander Ecker

Bachelor’s and Master’s theses

General requirements

If you’re interested in joining our lab for a thesis, have a look at this document describing how we work and what we expect from you.

We expect prospective students to have substantial knowledge in machine learning, its mathematical foundations and Python programming. We therefore expect students interested in doing their thesis in our lab to take our courses on Machine Learning and Deep Learning for Computer Vision unless they have acquired equivalent knowledge otherwise. For Bachelor’s students, we also recommend the Practical Course Data Science.

Further recommended courses are:

Please note, our thesis supervision capacity is limited and we receive more thesis inquiries than we are able supervise. Therefore, we have to select candidates. If you are interested, please write an email with the subject “Master’s thesis” or “Bachelor’s thesis” containing one to three sentences about what you would like to work on and your study record to the supervisor stated below.

We will get back to you within a few days. Otherwise, do not hesitate to remind us :).

Thesis offers

Biologically inspired CNN for ganglion cell response prediction
Implement a Convolutional Neural Network constraint by insights from the retinal circuitry to predict the responses of retinal ganglion cells.
Supervisor: Michaela Vystrčilová
Deep Embedding Clustering for the visual cortex cells embeddings
Improve functional neuronal clustering
Supervisor: Polina Turishcheva
Design Optimization for Acoustics
Using Neural Networks to predict the frequency responses on beading patterns.
Supervisor: Jan van Delden
Embeddings for neurons function and how they relate to morphology and cell types
Improve functional neuronal clustering
Supervisor: Polina Turishcheva
Implicit_learning_for_neuronal_prepresentation
Improve functional neuronal clustering
Supervisor: Polina Turishcheva
Model Neurons Interactions in time and between each other
Adjust readouts for neuroscience vision models to consider time and neurons interactions
Supervisor: Polina Turishcheva
Perspective module to module eye focus for mouse visual cortex
Adjust the module from the foundational model and try to make it sharable between animals
Supervisor: Polina Turishcheva
Self-Supervised Pretraining for Training Robust Monkey Detection and Tracking Models
Leveraging self-supervised pretraining to improve the robustness of monkey detection and tracking models in diverse environmental conditions.
Supervisor: Felix Müller
Solving Citation Networks with Large Language Models
By focussing on the abstract, LLMs should be able to effectively solve Cora and other citation datasets
Supervisor: Martin Ritzert
Tranformers as predictive models of the retina
Design and implement a transformer network to beat our state-of-the-art CNNs on retinal ganglion cell prediction tasks.
Supervisor: Michaela Vystrčilová
Treewidth-Based Positional Encodings
Generalizing a positional encoding for trees to general graphs using the notion of treewidth
Supervisor: Martin Ritzert

Research

2024

R. Vogg, T. Lüddecke, J. Henrich, S. Dey, M. Nuske, V. Hassler, D. Murphy, J. Fischer, J. Ostner, O. Schülke, P. Kappeler, C. Fichtel, A. Gail, S. Treue, H. Scherberger, F. Wörgötter, A. Ecker
Computer Vision for Primate Behavior Analysis in the Wild
arXiv preprint arXiv:2401.16424, 2024
show abstract
Advances in computer vision as well as increasingly widespread video-based behavioral monitoring have great potential for transforming how we study animal cognition and behavior. However, there is still a fairly large gap between the exciting prospects and what can actually be achieved in practice today, especially in videos from the wild. With this perspective paper, we want to contribute towards closing this gap, by guiding behavioral scientists in what can be expected from current methods and steering computer vision researchers towards problems that are relevant to advance research in animal behavior. We start with a survey of the state-of-the-art methods for computer vision problems that are directly relevant to the video-based study of animal behavior, including object detection, multi-individual tracking, (inter)action recognition and individual identification. We then review methods for effort-efficient learning, which is one of the biggest challenges from a practical perspective. Finally, we close with an outlook into the future of the emerging field of computer vision for animal behavior, where we argue that the field should move fast beyond the common frame-by-frame processing and treat video as a first-class citizen.
@article{vogg2024computer, title: {Computer Vision for Primate Behavior Analysis in the Wild}, author: {R. Vogg, T. Lüddecke, J. Henrich, S. Dey, M. Nuske, V. Hassler, D. Murphy, J. Fischer, J. Ostner, O. Schülke, P. Kappeler, C. Fichtel, A. Gail, S. Treue, H. Scherberger, F. Wörgötter, A. Ecker}, year: {2024}, journal: {arXiv preprint arXiv:2401.16424}, }
.bib
M. Vystrčilová, S. Sridhar, M. Burg, T. Gollisch, A. Ecker
Convolutional neural network models of the primate retina reveal adaptation to natural stimulus statistics
bioRxiv, 2024
show abstract
The diverse nature of visual environments demands that the retina, the first stage of the visual system, encodes a vast range of stimuli with various statistics. The retina adapts its computations to some specific features of the input, such as brightness, contrast or motion. However, it is less clear whether it also adapts to the statistics of natural scenes compared to white noise, the latter of which is often used to infer models of retinal computation. To address this question, we analyzed neural activity of retinal ganglion cells (RGCs) in response to both white noise and naturalistic movie stimuli. We performed a systematic comparative analysis of traditional linear-nonlinear (LN) and recent convolutional neural network (CNN) models and tested their generalization across stimulus domains. We found that no model type trained on one stimulus ensemble was able to accurately predict neural activity on the other, suggesting that retinal processing depends on the stimulus statistics. Under white noise stimulation, the receptive fields of the neurons were mostly lowpass, while under natural image statistics they exhibited a more pronounced surround resembling the whitening filters predicted by efficient coding. Together, these results suggest that retinal processing dynamically adapts to the stimulus statistics.
@article{vystrvcilova2024convolutional, title: {Convolutional neural network models of the primate retina reveal adaptation to natural stimulus statistics}, author: {M. Vystrčilová, S. Sridhar, M. Burg, T. Gollisch, A. Ecker}, year: {2024}, journal: {bioRxiv}, }
.bib
F. Müller, R. Görge, A. Bernzen, J. Pirk, M. Poretschkin
LLMs and Memorization: On Quality and Specificity of Copyright Compliance
2024 AAAI/ACM Conference on AI, Ethics, and Society (to appear), 2024
show abstract
Memorization in large language models (LLMs) is a growing concern. LLMs have been shown to easily reproduce parts of their training data, including copyrighted work. This is an important problem to solve, as it may violate existing copyright laws as well as the European AI Act. In this work, we propose a systematic analysis to quantify the extent of potential copyright infringements in LLMs using European law as an example. Unlike previous work, we evaluate instruction-finetuned models in a realistic end-user scenario. Our analysis builds on a proposed threshold of 160 characters, which we borrow from the German Copyright Service Provider Act and a fuzzy text matching algorithm to identify potentially copyright-infringing textual reproductions. The specificity of countermeasures against copyright infringement is analyzed by comparing model behavior on copyrighted and public domain data. We investigate what behaviors models show instead of producing protected text (such as refusal or hallucination) and provide a first legal assessment of these behaviors. We find that there are huge differences in copyright compliance, specificity, and appropriate refusal among popular LLMs. Alpaca, GPT 4, GPT 3.5, and Luminous perform best in our comparison, with OpenGPT-X, Alpaca, and Luminous producing a particularly low absolute number of potential copyright violations. Code will be published soon.
@article{mueller2024llmsmemorizationqualityspecificity, title: {LLMs and Memorization: On Quality and Specificity of Copyright Compliance}, author: {F. Müller, R. Görge, A. Bernzen, J. Pirk, M. Poretschkin}, year: {2024}, journal: {2024 AAAI/ACM Conference on AI, Ethics, and Society (to appear)}, }
.bib
S. Sridhar, M. Vystrčilová, M. Khani, D. Karamanlis, H. Schreyer, V. Ramakrishna, S. Krueppel, S. Zapp, M. Mietsch, A. Ecker, T. Gollisch
Modeling spatial contrast sensitivity in responses of primate retinal ganglion cells to natural movies
bioRxiv, 2024
show abstract
Retinal ganglion cells, the output neurons of the vertebrate retina, often display nonlinear summation of visual signals over their receptive fields. This creates sensitivity to spatial contrast, letting the cells respond to spatially structured visual stimuli, such as a contrast-reversing grating, even when no net change in overall illumination of the receptive field occurs. Yet, computational models of ganglion cell responses are often based on linear receptive fields. Nonlinear extensions, on the other hand, such as subunit models, which separate receptive fields into smaller, nonlinearly combined subfields, are often cumbersome to fit to experimental data, in particular when natural stimuli are considered. Previous work in the salamander retina has shown that sensitivity to spatial contrast in response to flashed images can be partly captured by a model that combines signals from the mean and variance of luminance signals inside the receptive field. Here, we extend this spatial contrast model for application to spatiotemporal stimulation and explore its performance on spiking responses that we recorded from retinas of marmosets under artificial and natural movies. We show how the model can be fitted to experimental data and that it outperforms common models with linear spatial integration, in particular for parasol ganglion cells. Finally, we use the model framework to infer the cells’ spatial scale of nonlinear spatial integration and contrast sensitivity. Our work shows that the spatial contrast model provides a simple approach to capturing aspects of nonlinear spatial integration with only few free parameters, which can be used to assess the cells’ functional properties under natural stimulation and which provides a simple-to-obtain benchmark for comparison with more detailed nonlinear encoding models.
@article{sridhar2024modeling, title: {Modeling spatial contrast sensitivity in responses of primate retinal ganglion cells to natural movies}, author: {S. Sridhar, M. Vystrčilová, M. Khani, D. Karamanlis, H. Schreyer, V. Ramakrishna, S. Krueppel, S. Zapp, M. Mietsch, A. Ecker, T. Gollisch}, year: {2024}, journal: {bioRxiv}, }
.bib

2023

N. Wu, I. Valera, A. Ecker, T. Euler, Y. Qiu
Bayesian Neural System Identification with Response Variability
arXiv preprint arXiv:2308.05990, 2023
@article{wu2023bayesian, title: {Bayesian Neural System Identification with Response Variability}, author: {N. Wu, I. Valera, A. Ecker, T. Euler, Y. Qiu}, year: {2023}, journal: {arXiv preprint arXiv:2308.05990}, }
.bib
Z. Ding, D. Tran, K. Ponder, E. Cobos, Z. Ding, P. Fahey, E. Wang, T. Muhammad, J. Fu, S. Cadena, others
Bipartite invariance in mouse primary visual cortex
bioRxiv, 2023
@article{ding2023bipartite, title: {Bipartite invariance in mouse primary visual cortex}, author: {Z. Ding, D. Tran, K. Ponder, E. Cobos, Z. Ding, P. Fahey, E. Wang, T. Muhammad, J. Fu, S. Cadena, others}, year: {2023}, journal: {bioRxiv}, }
.bib
A. Barreto, L. Reifenrath, R. Vogg, F. Sinz, A. Mahlein
Data Augmentation for Mask-Based Leaf Segmentation of UAV-Images as a Basis to Extract Leaf-Based Phenotyping Parameters
KI-Künstliche Intelligenz, 2023
show abstract
In crop protection, disease quantification parameters such as disease incidence (DI) and disease severity (DS) are the principal indicators for decision making, aimed at ensuring the safety and productivity of crop yield. The quantification is standardized with leaf organs, defined as individual scoring units. This study focuses on identifying and segmenting individual leaves in agricultural fields using unmanned aerial vehicle (UAV), multispectral imagery of sugar beet fields, and deep instance segmentation networks (Mask R-CNN). Five strategies for achieving network robustness with limited labeled images are tested and compared, employing simple and copy-paste image augmentation techniques. The study also evaluates the impact of environmental conditions on network performance. Metrics of performance show that multispectral UAV images recorded under sunny conditions lead to a performance drop. Focusing on the practical application, we employ Mask R-CNN models in an image-processing pipeline to calculate leaf-based parameters including DS and DI. The pipeline was applied in time-series in an experimental trial with five varieties and two fungicide strategies to illustrate epidemiological development. Disease severity calculated with the model with highest Average Precision (AP) shows the strongest correlation with the same parameter assessed by experts. The time-series development of disease severity and disease incidence demonstrates the advantages of multispectral UAV-imagery in contrasting varieties for resistance, as well as the limits for disease control measurements. This study identifies key components for automatic leaf segmentation of diseased plants using UAV imagery, such as illumination and disease condition. It also provides a tool for delivering leaf-based parameters relevant to optimize crop production through automated disease quantification by imaging tools.
@article{barreto2023data, title: {Data Augmentation for Mask-Based Leaf Segmentation of UAV-Images as a Basis to Extract Leaf-Based Phenotyping Parameters}, author: {A. Barreto, L. Reifenrath, R. Vogg, F. Sinz, A. Mahlein}, year: {2023}, journal: {KI-Künstliche Intelligenz}, }
.bib
J. Schultz, J. van Delden, C. Blech, S. Langer, T. Lüddecke
Deep learning for frequency response prediction of a multimass oscillator
PAMM, 2023
@article{schultz2023deep, title: {Deep learning for frequency response prediction of a multimass oscillator}, author: {J. Schultz, J. van Delden, C. Blech, S. Langer, T. Lüddecke}, year: {2023}, journal: {PAMM}, }
.bib
K. Willeke, K. Restivo, K. Franke, A. Nix, S. Cadena, T. Shinn, C. Nealley, G. Rodriguez, S. Patel, A. Ecker, others
Deep learning-driven characterization of single cell tuning in primate visual area V4 unveils topological organization
bioRxiv, 2023
@article{willeke2023deep, title: {Deep learning-driven characterization of single cell tuning in primate visual area V4 unveils topological organization}, author: {K. Willeke, K. Restivo, K. Franke, A. Nix, S. Cadena, T. Shinn, C. Nealley, G. Rodriguez, S. Patel, A. Ecker, others}, year: {2023}, journal: {bioRxiv}, }
.bib
J. van Delden, J. Schultz, C. Blech, S. Langer, T. Lüddecke
Learning to Predict Structural Vibrations
arXiv preprint arXiv:2310.05469, 2023
@article{delden2023vibroacoustic, title: {Learning to Predict Structural Vibrations}, author: {J. van Delden, J. Schultz, C. Blech, S. Langer, T. Lüddecke}, year: {2023}, journal: {arXiv preprint arXiv:2310.05469}, }
.bib
J. Fu, S. Shrinivasan, K. Ponder, T. Muhammad, Z. Ding, E. Wang, Z. Ding, D. Tran, P. Fahey, S. Papadopoulos, others
Pattern completion and disruption characterize contextual modulation in mouse visual cortex
bioRxiv, 2023
@article{fu2023pattern, title: {Pattern completion and disruption characterize contextual modulation in mouse visual cortex}, author: {J. Fu, S. Shrinivasan, K. Ponder, T. Muhammad, Z. Ding, E. Wang, Z. Ding, D. Tran, P. Fahey, S. Papadopoulos, others}, year: {2023}, journal: {bioRxiv}, }
.bib
M. A. Weis, L. Hansel, T. Lüddecke, A. S. Ecker
Self-Supervised Graph Representation Learning for Neuronal Morphologies
Transactions on Machine Learning Research, 2023
show abstract
Unsupervised graph representation learning has recently gained interest in several application domains such as neuroscience, where modeling the diverse morphology of cell types in the brain is one of the key challenges. It is currently unknown how many excitatory cortical cell types exist and what their defining morphological features are. Here we present GraphDINO, a purely data-driven approach to learn low-dimensional representations of 3D neuronal morphologies from unlabeled large-scale datasets. GraphDINO is a novel transformer-based representation learning method for spatially-embedded graphs. To enable self-supervised learning on transformers, we (1) developed data augmentation strategies for spatially-embedded graphs, (2) adapted the positional encoding and (3) introduced a novel attention mechanism, AC-Attention, which combines attention-based global interaction between nodes and classic graph convolutional processing. We show, in two different species and across multiple brain areas, that this method yields morphological cell type clusterings that are on par with manual feature-based classification by experts, but without using prior knowledge about the structural features of neurons. Moreover, it outperforms previous approaches on quantitative benchmarks predicting expert labels. Our method could potentially enable data-driven discovery of novel morphological features and cell types in large-scale datasets. It is applicable beyond neuroscience in settings where samples in a dataset are graphs and graph-level embeddings are desired.
@article{weis2023selfsupervised, title: {Self-Supervised Graph Representation Learning for Neuronal Morphologies}, author: {M. A. Weis, L. Hansel, T. Lüddecke, A. S. Ecker}, year: {2023}, journal: {Transactions on Machine Learning Research}, }
.bib
P. Turishcheva, P. Fahey, L. Hansel, R. Froebe, K. Ponder, M. Vystrčilová, K. Willeke, M. Bashiri, E. Wang, Z. Ding, others
The Dynamic Sensorium competition for predicting large-scale mouse visual cortex activity from videos
arXiv preprint arXiv:2305.19654, 2023
show abstract
Understanding how biological visual systems process information is challenging due to the complex nonlinear relationship between neuronal responses and high-dimensional visual input. Artificial neural networks have already improved our understanding of this system by allowing computational neuroscientists to create predictive models and bridge biological and machine vision. During the Sensorium 2022 competition, we introduced benchmarks for vision models with static input. However, animals operate and excel in dynamic environments, making it crucial to study and understand how the brain functions under these conditions. Moreover, many biological theories, such as predictive coding, suggest that previous input is crucial for current input processing. Currently, there is no standardized benchmark to identify state-of-the-art dynamic models of the mouse visual system. To address this gap, we propose the Sensorium 2023 Competition with dynamic input. This includes the collection of a new large-scale dataset from the primary visual cortex of five mice, containing responses from over 38,000 neurons to over 2 hours of dynamic stimuli per neuron. Participants in the main benchmark track will compete to identify the best predictive models of neuronal responses for dynamic input. We will also host a bonus track in which submission performance will be evaluated on out-of-domain input, using withheld neuronal responses to dynamic input stimuli whose statistics differ from the training set. Both tracks will offer behavioral data along with video stimuli. As before, we will provide code, tutorials, and strong pre-trained baseline models to encourage participation. We hope this competition will continue to strengthen the accompanying Sensorium benchmarks collection as a standard tool to measure progress in large-scale neural system identification models of the entire mouse visual hierarchy and beyond.
@article{turishcheva2023dynamic, title: {The Dynamic Sensorium competition for predicting large-scale mouse visual cortex activity from videos}, author: {P. Turishcheva, P. Fahey, L. Hansel, R. Froebe, K. Ponder, M. Vystrčilová, K. Willeke, M. Bashiri, E. Wang, Z. Ding, others}, year: {2023}, journal: {arXiv preprint arXiv:2305.19654}, }
.bib
T. Jiang, M. Freudenberg, C. Kleinn, A. Ecker, N. Nölke
The Impacts of Quality-Oriented Dataset Labeling on Tree Cover Segmentation Using U-Net: A Case Study in WorldView-3 Imagery
Remote Sensing, 2023
@article{jiang2023impacts, title: {The Impacts of Quality-Oriented Dataset Labeling on Tree Cover Segmentation Using U-Net: A Case Study in WorldView-3 Imagery}, author: {T. Jiang, M. Freudenberg, C. Kleinn, A. Ecker, N. Nölke}, year: {2023}, journal: {Remote Sensing}, }
.bib
E. Wang, P. Fahey, K. Ponder, Z. Ding, A. Change, T. Muhammad, S. Patel, Z. Ding, D. Tran, J. Fu, others
Towards a foundation model of the mouse visual cortex
bioRxiv, 2023
@article{wang2023towards, title: {Towards a foundation model of the mouse visual cortex}, author: {E. Wang, P. Fahey, K. Ponder, Z. Ding, A. Change, T. Muhammad, S. Patel, Z. Ding, D. Tran, J. Fu, others}, year: {2023}, journal: {bioRxiv}, }
.bib

2022

L. Hoefling, K. Szatko, C. Behrens, Y. Qiu, D. Klindt, Z. Jessen, G. Schwartz, M. Bethge, P. Berens, K. Franke, A. Ecker, T. Euler
A chromatic feature detector in the retina signals visual context changes
bioRxiv, 2022
@article{hoefling2022chromatic, title: {A chromatic feature detector in the retina signals visual context changes}, author: {L. Hoefling, K. Szatko, C. Behrens, Y. Qiu, D. Klindt, Z. Jessen, G. Schwartz, M. Bethge, P. Berens, K. Franke, A. Ecker, T. Euler}, year: {2022}, journal: {bioRxiv}, }
.bib
M. Goldin, B. Lefebvre, S. Virgili, M. Pham Van Cang, A. Ecker, T. Mora, U. Ferrari, O. Marre
Context-dependent selectivity to natural images in the retina
Nature Communications, 2022
@article{goldin2022context, title: {Context-dependent selectivity to natural images in the retina}, author: {M. Goldin, B. Lefebvre, S. Virgili, M. Pham Van Cang, A. Ecker, T. Mora, U. Ferrari, O. Marre}, year: {2022}, journal: {Nature Communications}, }
.bib
S. Cadena, K. Willeke, K. Restivo, G. Denfield, F. Sinz, M. Bethge, A. Tolias, A. Ecker
Diverse task-driven modeling of macaque V4 reveals functional specialization towards semantic tasks
bioRxiv, 2022
@article{cadena2022diverse, title: {Diverse task-driven modeling of macaque V4 reveals functional specialization towards semantic tasks}, author: {S. Cadena, K. Willeke, K. Restivo, G. Denfield, F. Sinz, M. Bethge, A. Tolias, A. Ecker}, year: {2022}, journal: {bioRxiv}, }
.bib
T. Lüddecke, A. S. Ecker
Image Segmentation Using Text and Image Prompts
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022
@article{lueddecke2021clipseg, title: {Image Segmentation Using Text and Image Prompts}, author: {T. Lüddecke, A. S. Ecker}, year: {2022}, journal: {2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, }
.bib
M. Weis, S. Papadopoulos, L. Hansel, T. Lüddecke, B. Celii, P. Fahey, J. Bae, A. Bodor, D. Brittain, J. Buchanan, D. Bumbarger, M. Castro, E. Cobos, F. Collman, N. da Costa, S. Dorkenwald, L. Elabbady, E. Froudarakis, A. Halageri, Z. Jia, C. Jordan, D. Kapner, N. Kemnitz, S. Kinn, K. Lee, K. Li, R. Lu, T. Macrina, G. Mahalingam, E. Mitchell, S. Mondal, S. Mu, B. Nehoran, S. Patel, X. Pitkow, S. Popovych, R. Reid, C. Schneider-Mizell, H. Seung, W. Silversmith, F. Sinz, M. Takeno, R. Torres, N. Turner, W. Wong, J. Wu, W. Yin, S. Yu, J. Reimer, A. Tolias, A. Ecker
Large-scale unsupervised discovery of excitatory morphological cell types in mouse visual cortex
bioRxiv, 2022
show abstract
Neurons in the neocortex exhibit astonishing morphological diversity which is critical for properly wiring neural circuits and giving neurons their functional properties. The extent to which the morphological diversity of excitatory neurons forms a continuum or is built from distinct clusters of cell types remains an open question. Here we took a data-driven approach using graph-based machine learning methods to obtain a low-dimensional morphological bar code describing more than 30,000 excitatory neurons in mouse visual areas V1, AL and RL that were reconstructed from a millimeter scale serial-section electron microscopy volume. We found a set of principles that captured the morphological diversity of the dendrites of excitatory neurons. First, their morphologies varied with respect to three major axes: soma depth, total apical and basal skeletal length. Second, neurons in layer 2/3 showed a strong trend of a decreasing width of their dendritic arbor and a smaller tuft with increasing cortical depth. Third, in layer 4, atufted neurons were primarily located in the primary visual cortex, while tufted neurons were more abundant in higher visual areas. Fourth, we discovered layer 4 neurons in V1 on the border to layer 5 which showed a tendency towards avoiding deeper layers with their dendrites. In summary, excitatory neurons exhibited a substantial degree of dendritic morphological variation, both within and across cortical layers, but this variation mostly formed a continuum, with only a few notable exceptions in deeper layers.Competing Interest StatementThe authors have declared no competing interest.
@article{weis2022largescale, title: {Large-scale unsupervised discovery of excitatory morphological cell types in mouse visual cortex}, author: {M. Weis, S. Papadopoulos, L. Hansel, T. Lüddecke, B. Celii, P. Fahey, J. Bae, A. Bodor, D. Brittain, J. Buchanan, D. Bumbarger, M. Castro, E. Cobos, F. Collman, N. da Costa, S. Dorkenwald, L. Elabbady, E. Froudarakis, A. Halageri, Z. Jia, C. Jordan, D. Kapner, N. Kemnitz, S. Kinn, K. Lee, K. Li, R. Lu, T. Macrina, G. Mahalingam, E. Mitchell, S. Mondal, S. Mu, B. Nehoran, S. Patel, X. Pitkow, S. Popovych, R. Reid, C. Schneider-Mizell, H. Seung, W. Silversmith, F. Sinz, M. Takeno, R. Torres, N. Turner, W. Wong, J. Wu, W. Yin, S. Yu, J. Reimer, A. Tolias, A. Ecker}, year: {2022}, journal: {bioRxiv}, }
.bib
K. F. Willeke, P. G. Fahey, M. Bashiri, L. Hansel, M. F. Burg, C. Blessing, S. A. Cadena, Z. Ding, K. Lurz, K. Ponder, T. Muhammad, S. S. Patel, A. S. Ecker, A. S. Tolias, F. H. Sinz
Retrospective on the SENSORIUM 2022 competition
PMLR, 2022
show abstract
The neural underpinning of the biological visual system is challenging to study experi- mentally, in particular as neuronal activity becomes increasingly nonlinear with respect to visual input. Artificial neural networks (ANNs) can serve a variety of goals for improving our understanding of this complex system, not only serving as predictive digital twins of sensory cortex for novel hypothesis generation in silico, but also incorporating bio-inspired architectural motifs to progressively bridge the gap between biological and machine vision. The mouse has recently emerged as a popular model system to study visual information processing, but no standardized large-scale benchmark to identify state-of-the-art mod- els of the mouse visual system has been established. To fill this gap, we proposed the SENSORIUM benchmark competition. We collected a large-scale dataset from mouse primary visual cortex containing the responses of more than 28,000 neurons across seven mice stimu- lated with thousands of natural images, together with simultaneous behavioral measurements that include running speed, pupil dilation, and eye movements. The benchmark challenge ranked models based on predictive performance for neuronal responses on a held-out test set, and included two tracks for model input limited to either stimulus only (SENSORIUM) or stimulus plus behavior (SENSORIUM+). As a part of the NeurIPS 2022 competition track, we received 172 model submissions from 26 teams, with the winning teams improving our previous state-of-the-art model by more than 15%. Dataset access and infrastructure for evaluation of model predictions will remain online as an ongoing benchmark. We would like to see this as a starting point for regular challenges and data releases, and as a standard tool for measuring progress in large-scale neural system identification models of the mouse visual system and beyond.
@article{willeke2022bsensorium, title: {Retrospective on the SENSORIUM 2022 competition}, author: {K. F. Willeke, P. G. Fahey, M. Bashiri, L. Hansel, M. F. Burg, C. Blessing, S. A. Cadena, Z. Ding, K. Lurz, K. Ponder, T. Muhammad, S. S. Patel, A. S. Ecker, A. S. Tolias, F. H. Sinz}, year: {2022}, journal: {PMLR}, }
.bib
K. F. Willeke, P. G. Fahey, M. Bashiri, L. Pede, M. F. Burg, C. Blessing, S. A. Cadena, Z. Ding, K. Lurz, K. Ponder, T. Muhammad, S. S. Patel, A. S. Ecker, A. S. Tolias, F. H. Sinz
The Sensorium competition on predicting large-scale mouse primary visual cortex activity
arXiv preprint arXiv:2206.08666, 2022
show abstract
The neural underpinning of the biological visual system is challenging to study experimentally, in particular as the neuronal activity becomes increasingly nonlinear with respect to visual input. Artificial neural networks (ANNs) can serve a variety of goals for improving our understanding of this complex system, not only serving as predictive digital twins of sensory cortex for novel hypothesis generation in silico, but also incorporating bio-inspired architectural motifs to progressively bridge the gap between biological and machine vision. The mouse has recently emerged as a popular model system to study visual information processing, but no standardized large-scale benchmark to identify state-of-the-art models of the mouse visual system has been established. To fill this gap, we propose the Sensorium benchmark competition. We collected a large-scale dataset from mouse primary visual cortex containing the responses of more than 28,000 neurons across seven mice stimulated with thousands of natural images, together with simultaneous behavioral measurements that include running speed, pupil dilation, and eye movements. The benchmark challenge will rank models based on predictive performance for neuronal responses on a held-out test set, and includes two tracks for model input limited to either stimulus only (Sensorium) or stimulus plus behavior (Sensorium+). We provide a starting kit to lower the barrier for entry, including tutorials, pre-trained baseline models, and APIs with one line commands for data loading and submission. We would like to see this as a starting point for regular challenges and data releases, and as a standard tool for measuring progress in large-scale neural system identification models of the mouse visual system and beyond.
@article{willeke2022sensorium, title: {The Sensorium competition on predicting large-scale mouse primary visual cortex activity}, author: {K. F. Willeke, P. G. Fahey, M. Bashiri, L. Pede, M. F. Burg, C. Blessing, S. A. Cadena, Z. Ding, K. Lurz, K. Ponder, T. Muhammad, S. S. Patel, A. S. Ecker, A. S. Tolias, F. H. Sinz}, year: {2022}, journal: {arXiv preprint arXiv:2206.08666}, }
.bib

2021

M. A. Weis, K. Chitta, Y. Sharma, W. Brendel, M. Bethge, A. Geiger, A. S. Ecker
Benchmarking Unsupervised Object Representations for Video Sequences
Journal of Machine Learning Research, 2021
@article{Weis2021, title: {Benchmarking Unsupervised Object Representations for Video Sequences}, author: {M. A. Weis, K. Chitta, Y. Sharma, W. Brendel, M. Bethge, A. Geiger, A. S. Ecker}, year: {2021}, journal: {Journal of Machine Learning Research}, }
.bib
K. Lurz, M. Bashiri, K. Willeke, A. Jagadish, E. Wang, E. Y. Walker, S. A. Cadena, T. Muhammad, E. Cobos, A. S. Tolias, A. S. Ecker, F. H. Sinz
Generalization in data-driven models of primary visual cortex
International Conference on Learning Representations, 2021
@inproceedings{lurz2021generalization, title: {Generalization in data-driven models of primary visual cortex}, author: {K. Lurz, M. Bashiri, K. Willeke, A. Jagadish, E. Wang, E. Y. Walker, S. A. Cadena, T. Muhammad, E. Cobos, A. S. Tolias, A. S. Ecker, F. H. Sinz}, year: {2021}, booktitle: {International Conference on Learning Representations}, }
.bib
M. Burg, S. Cadena, G. Denfield, E. Walker, A. Tolias, M. Bethge, A. Ecker
Learning divisive normalization in primary visual cortex
PLOS Computational Biology, 2021
show abstract
Divisive normalization (DN) is a prominent computational building block in the brain that has been proposed as a canonical cortical operation. Numerous experimental studies have verified its importance for capturing nonlinear neural response properties to simple, artificial stimuli, and computational studies suggest that DN is also an important component for processing natural stimuli. However, we lack quantitative models of DN that are directly informed by measurements of spiking responses in the brain and applicable to arbitrary stimuli. Here, we propose a DN model that is applicable to arbitrary input images. We test its ability to predict how neurons in macaque primary visual cortex (V1) respond to natural images, with a focus on nonlinear response properties within the classical receptive field. Our model consists of one layer of subunits followed by learned orientation-specific DN. It outperforms linear-nonlinear and wavelet-based feature representations and makes a significant step towards the performance of state-of-the-art convolutional neural network (CNN) models. Unlike deep CNNs, our compact DN model offers a direct interpretation of the nature of normalization. By inspecting the learned normalization pool of our model, we gained insights into a long-standing question about the tuning properties of DN that update the current textbook description: we found that within the receptive field oriented features were normalized preferentially by features with similar orientation rather than non-specifically as currently assumed.
@article{burg_2021_learning_divisive_normalization, title: {Learning divisive normalization in primary visual cortex}, author: {M. Burg, S. Cadena, G. Denfield, E. Walker, A. Tolias, M. Bethge, A. Ecker}, year: {2021}, journal: {PLOS Computational Biology}, }
.bib
M. A. Weis, L. Pede, T. Lüddecke, A. S. Ecker
Self-supervised Representation Learning of Neuronal Morphologies
arXiv, 2021
show abstract
Understanding the diversity of cell types and their function in the brain is one of the key challenges in neuroscience. The advent of large-scale datasets has given rise to the need of unbiased and quantitative approaches to cell type classification. We present GraphDINO, a purely data-driven approach to learning a low dimensional representation of the 3D morphology of neurons. GraphDINO is a novel graph representation learning method for spatial graphs utilizing self-supervised learning on transformer models. It smoothly interpolates between attention-based global interaction between nodes and classic graph convolutional processing. We show that this method is able to yield morphological cell type clustering that is comparable to manual feature-based classification and shows a good correspondence to expert-labeled cell types in two different species and cortical areas. Our method is applicable beyond neuroscience in settings where samples in a dataset are graphs and graph-level embeddings are desired.
@article{weis2021selfsupervised, title: {Self-supervised Representation Learning of Neuronal Morphologies}, author: {M. A. Weis, L. Pede, T. Lüddecke, A. S. Ecker}, year: {2021}, journal: {arXiv}, }
.bib
D. Kobak, Y. Bernaerts, M. Weis, F. Scala, A. Tolias, P. Berens
Sparse reduced-rank regression for exploratory visualisation of paired multivariate data
Journal of the Royal Statistical Society: Series C (Applied Statistics), 2021
@article{weis2021sparse, title: {Sparse reduced-rank regression for exploratory visualisation of paired multivariate data}, author: {D. Kobak, Y. Bernaerts, M. Weis, F. Scala, A. Tolias, P. Berens}, year: {2021}, journal: {Journal of the Royal Statistical Society: Series C (Applied Statistics)}, }
.bib

2020

V. Benson, A. Ecker
Assessing out-of-domain generalization for robust building damage detection
NeurIPS 2020 Workshop on Artificial Intelligence for Humanitarian Assistance and Disaster Response (AI+HADR 2020), 2020
@inproceedings{benson2020assessing, title: {Assessing out-of-domain generalization for robust building damage detection}, author: {V. Benson, A. Ecker}, year: {2020}, booktitle: {NeurIPS 2020 Workshop on Artificial Intelligence for Humanitarian Assistance and Disaster Response (AI+HADR 2020)}, }
.bib
C. Michaelis, M. Bethge, A. Ecker
Closing the generalization gap in one-shot object detection
arXiv preprint arXiv:2011.04267, 2020
@article{michaelis2020closing, title: {Closing the generalization gap in one-shot object detection}, author: {C. Michaelis, M. Bethge, A. Ecker}, year: {2020}, journal: {arXiv preprint arXiv:2011.04267}, }
.bib
T. Lüddecke, A. Ecker
CNNs efficiently learn long-range dependencies
NeurIPS 2020 Workshop on Shared Visual Representations in Human & Machine Intelligence, 2020
@inproceedings{luddeckecnns, title: {CNNs efficiently learn long-range dependencies}, author: {T. Lüddecke, A. Ecker}, year: {2020}, booktitle: {NeurIPS 2020 Workshop on Shared Visual Representations in Human & Machine Intelligence}, }
.bib
T. Lüddecke, F. Wörgötter
Fine-grained action plausibility rating
Robotics and Autonomous Systems (RAS), 2020
@article{lueddecke20, title: {Fine-grained action plausibility rating}, author: {T. Lüddecke, F. Wörgötter}, year: {2020}, journal: {Robotics and Autonomous Systems (RAS)}, }
.bib
M. Rolínek, V. Musil, A. Paulus, M. Vlastelica, C. Michaelis, G. Martius
Optimizing Rank-based Metrics with Blackbox Differentiation
Computer Vision and Pattern Recognition (CVPR), 2020
@inproceedings{Rolínek2019a, title: {Optimizing Rank-based Metrics with Blackbox Differentiation}, author: {M. Rolínek, V. Musil, A. Paulus, M. Vlastelica, C. Michaelis, G. Martius}, year: {2020}, booktitle: {Computer Vision and Pattern Recognition (CVPR)}, }
.bib
I. Ustyuzhaninov, S. A. Cadena, E. Froudarakis, P. G. Fahey, E. Y. Walker, E. Cobos, J. Reimer, F. H. Sinz, A. S. Tolias, M. Bethge, A. S. Ecker
Rotation-invariant clustering of functional cell types in primary visual cortex
International Conference on Learning Representations (ICLR), 2020
@inproceedings{Ustyuzhaninov2020a, title: {Rotation-invariant clustering of functional cell types in primary visual cortex}, author: {I. Ustyuzhaninov, S. A. Cadena, E. Froudarakis, P. G. Fahey, E. Y. Walker, E. Cobos, J. Reimer, F. H. Sinz, A. S. Tolias, M. Bethge, A. S. Ecker}, year: {2020}, booktitle: {International Conference on Learning Representations (ICLR)}, }
.bib
Z. Zhao, D. Klindt, A. M. Chagas, K. P. Szatko, L. Rogerson, D. Protti, C. Behrens, D. Dalkara, T. Schubert, M. Bethge, K. Franke, P. Berens, A. S. Ecker, T. Euler
The temporal structure of the inner retina at a single glance
Scientific Reports, 2020
@article{Z*2020a, title: {The temporal structure of the inner retina at a single glance}, author: {Z. Zhao, D. Klindt, A. M. Chagas, K. P. Szatko, L. Rogerson, D. Protti, C. Behrens, D. Dalkara, T. Schubert, M. Bethge, K. Franke, P. Berens, A. S. Ecker, T. Euler}, year: {2020}, journal: {Scientific Reports}, }
.bib

2019

A. S. Ecker, F. H. Sinz, E. Froudarakis, P. G. Fahey, S. A. Cadena, E. Y. Walker, E. Cobos, J. Reimer, A. S. Tolias, M. Bethge
A rotation-equivariant convolutional neural network model of primary visual cortex
International Conference on Learning Representations (ICLR), 2019
@inproceedings{ecker_2019, title: {A rotation-equivariant convolutional neural network model of primary visual cortex}, author: {A. S. Ecker, F. H. Sinz, E. Froudarakis, P. G. Fahey, S. A. Cadena, E. Y. Walker, E. Cobos, J. Reimer, A. S. Tolias, M. Bethge}, year: {2019}, journal: {International Conference on Learning Representations (ICLR)}, }
.bib
C. Michaelis, B. Mitzkus, R. Geirhos, E. Rusak, O. Bringmann, A. S. Ecker, M. Bethge, W. Brendel
Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming
Machine Learning for Autonomous Driving Workshop, NeurIPS 2019, 2019
@inproceedings{michaelis2019dragon, title: {Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming}, author: {C. Michaelis, B. Mitzkus, R. Geirhos, E. Rusak, O. Bringmann, A. S. Ecker, M. Bethge, W. Brendel}, year: {2019}, booktitle: {Machine Learning for Autonomous Driving Workshop, NeurIPS 2019}, }
.bib
T. Lüddecke, T. Kulvicius, F. Wörgötter
Context-based Affordance Segmentation from 2D Images for Robot Action
Robotics and Autonomous Systems (RAS), 2019
@article{lueddecke19a, title: {Context-based Affordance Segmentation from 2D Images for Robot Action}, author: {T. Lüddecke, T. Kulvicius, F. Wörgötter}, year: {2019}, journal: {Robotics and Autonomous Systems (RAS)}, }
.bib
S. A. Cadena, G. H. Denfield, E. Y. Walker, L. A. Gatys, A. S. Tolias, M. Bethge, A. S. Ecker
Deep convolutional models improve predictions of macaque V1 responses to natural images
PLoS Computational Biology, 2019
@article{Cadena2019, title: {Deep convolutional models improve predictions of macaque V1 responses to natural images}, author: {S. A. Cadena, G. H. Denfield, E. Y. Walker, L. A. Gatys, A. S. Tolias, M. Bethge, A. S. Ecker}, year: {2019}, journal: {PLoS Computational Biology}, }
.bib
T. Lüddecke, A. Agostini, M. Fauth, M. Tamosiunaite, F. Wörgötter
Distributional Semantics of Objects in Visual Scenes in Comparison to Text
Artificial Intelligence, 2019
@article{lueddecke19, title: {Distributional Semantics of Objects in Visual Scenes in Comparison to Text}, author: {T. Lüddecke, A. Agostini, M. Fauth, M. Tamosiunaite, F. Wörgötter}, year: {2019}, journal: {Artificial Intelligence}, }
.bib
S. A. Cadena, F. H. Sinz, T. Muhammad, E. Froudarakis, E. Cobos, E. Y. Walker, J. Reimer, M. Bethge, A. Tolias, A. S. Ecker
How well do deep neural networks trained on object recognition characterize the mouse visual system?
NeurIPS Neuro AI Workshop, 2019
@inproceedings{Cadena2019b, title: {How well do deep neural networks trained on object recognition characterize the mouse visual system?}, author: {S. A. Cadena, F. H. Sinz, T. Muhammad, E. Froudarakis, E. Cobos, E. Y. Walker, J. Reimer, M. Bethge, A. Tolias, A. S. Ecker}, year: {2019}, journal: {NeurIPS Neuro AI Workshop}, }
.bib
R. Geirhos, P. Rubisch, C. Michaelis, M. Bethge, F. A. Wichmann, W. Brendel
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness
International Conference on Learning Representations (ICLR), 2019
@inproceedings{Geirhos2019a, title: {ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness}, author: {R. Geirhos, P. Rubisch, C. Michaelis, M. Bethge, F. A. Wichmann, W. Brendel}, year: {2019}, journal: {International Conference on Learning Representations (ICLR)}, }
.bib
E. Y. Walker, F. H. Sinz, E. Froudarakis, P. G. Fahey, T. Muhammad, A. S. Ecker, E. Cobos, J. Reimer, X. Pitkow, A. S. Tolias
Inception loops discover what excites neurons most using deep predictive models
Nature Neuroscience, 2019
@article{Walker2019, title: {Inception loops discover what excites neurons most using deep predictive models}, author: {E. Y. Walker, F. H. Sinz, E. Froudarakis, P. G. Fahey, T. Muhammad, A. S. Ecker, E. Cobos, J. Reimer, X. Pitkow, A. S. Tolias}, year: {2019}, journal: {Nature Neuroscience}, }
.bib
Neural Data Science Group
Institute of Computer Science
University of Goettingen