Convergent large-scale network and local vulnerabilities underlie brain atrophy across Parkinson’s disease stages

The SIMple microscope: Development of a fibre-based platform for accessible SIM imaging in unconventional environments

Advancements in imaging speed, depth and resolution have made structured illumination microscopy (SIM) an increasingly powerful optical sectioning (OS) and super-resolution (SR) technique, but these developments remain inaccessible to many life science researchers due to the cost, optical complexity and delicacy of these instruments. We address these limitations by redesigning the optical path using in-line fibre components that are compact, lightweight and easily assembled in a “Plug & Play” modality, without compromising imaging performance. They can be integrated into an existing widefield microscope with a minimum of optical components and alignment, making OS-SIM more accessible to researchers with less optics experience. We also demonstrate a complete SR-SIM imaging system with dimensions 300 mm × 300 mm × 450 mm. We propose to enable accessible SIM imaging by utilising its compact, lightweight and robust design to transport it where it is needed, and image in “unconventional” environments where factors such as temperature and biosafety considerations currently limit imaging experiments.

OpenNeuro FitLins GLM: An Accessible, Semi-Automated Pipeline for OpenNeuro Task fMRI Analysis

In this talk, I will discuss the OpenNeuro Fitlins GLM package and provide an illustration of the analytic workflow. OpenNeuro FitLins GLM is a semi-automated pipeline that reduces barriers to analyzing task-based fMRI data from OpenNeuro's 600+ task datasets. Created for psychology, psychiatry and cognitive neuroscience researchers without extensive computational expertise, this tool automates what is largely a manual process and compilation of in-house scripts for data retrieval, validation, quality control, statistical modeling and reporting that, in some cases, may require weeks of effort. The workflow abides by open-science practices, enhancing reproducibility and incorporates community feedback for model improvement. The pipeline integrates BIDS-compliant datasets and fMRIPrep preprocessed derivatives, and dynamically creates BIDS Statistical Model specifications (with Fitlins) to perform common mass univariate [GLM] analyses. To enhance and standardize reporting, it generates comprehensive reports which includes design matrices, statistical maps and COBIDAS-aligned reporting that is fully reproducible from the model specifications and derivatives. OpenNeuro Fitlins GLM has been tested on over 30 datasets spanning 50+ unique fMRI tasks (e.g., working memory, social processing, emotion regulation, decision-making, motor paradigms), reducing analysis times from weeks to hours when using high-performance computers, thereby enabling researchers to conduct robust single-study, meta- and mega-analyses of task fMRI data with significantly improved accessibility, standardized reporting and reproducibility.

A personal journey on understanding intelligence

The focus of this talk is not about my research in AI or Robotics but my own journey on trying to do research and understand intelligence in a rapidly evolving research landscape. I will trace my path from conducting early-stage research during graduate school, to working on practical solutions within a startup environment, and finally to my current role where I participate in more structured research at a major tech company. Through these varied experiences, I will provide different perspectives on research and talk about how my core beliefs on intelligence have changed and sometimes even been compromised. There are no lessons to be learned from my stories, but hopefully they will be entertaining.

FLUXSynID: High-Resolution Synthetic Face Generation for Document and Live Capture Images

Synthetic face datasets are increasingly used to overcome the limitations of real-world biometric data, including privacy concerns, demographic imbalance, and high collection costs. However, many existing methods lack fine-grained control over identity attributes and fail to produce paired, identity-consistent images under structured capture conditions. In this talk, I will present FLUXSynID, a framework for generating high-resolution synthetic face datasets with user-defined identity attribute distributions and paired document-style and trusted live capture images. The dataset generated using FLUXSynID shows improved alignment with real-world identity distributions and greater diversity compared to prior work. I will also discuss how FLUXSynID’s dataset and generation tools can support research in face recognition and morphing attack detection (MAD), enhancing model robustness in both academic and practical applications.

Open SPM: A Modular Framework for Scanning Probe Microscopy

OpenSPM aims to democratize innovation in the field of scanning probe microscopy (SPM), which is currently dominated by a few proprietary, closed systems that limit user-driven development. Our platform includes a high-speed OpenAFM head and base optimized for small cantilevers, an OpenAFM controller, a high-voltage amplifier, and interfaces compatible with several commercial AFM systems such as the Bruker Multimode, Nanosurf DriveAFM, Witec Alpha SNOM, Zeiss FIB-SEM XB550, and Nenovision Litescope. We have created a fully documented and community-driven OpenSPM platform, with training resources and sourcing information, which has already enabled the construction of more than 15 systems outside our lab. The controller is integrated with open-source tools like Gwyddion, HDF5, and Pycroscopy. We have also engaged external companies, two of which are integrating our controller into their products or interfaces. We see growing interest in applying parts of the OpenSPM platform to related techniques such as correlated microscopy, nanoindentation, and scanning electron/confocal microscopy. To support this, we are developing more generic and modular software, alongside a structured development workflow. A key feature of the OpenSPM system is its Python-based API, which makes the platform fully scriptable and ideal for AI and machine learning applications. This enables, for instance, automatic control and optimization of PID parameters, setpoints, and experiment workflows. With a growing contributor base and industry involvement, OpenSPM is well positioned to become a global, open platform for next-generation SPM innovation.

From Spiking Predictive Coding to Learning Abstract Object Representation

In a first part of the talk, I will present Predictive Coding Light (PCL), a novel unsupervised learning architecture for spiking neural networks. In contrast to conventional predictive coding approaches, which only transmit prediction errors to higher processing stages, PCL learns inhibitory lateral and top-down connectivity to suppress the most predictable spikes and passes a compressed representation of the input to higher processing stages. We show that PCL reproduces a range of biological findings and exhibits a favorable tradeoff between energy consumption and downstream classification performance on challenging benchmarks. A second part of the talk will feature our lab’s efforts to explain how infants and toddlers might learn abstract object representations without supervision. I will present deep learning models that exploit the temporal and multimodal structure of their sensory inputs to learn representations of individual objects, object categories, or abstract super-categories such as „kitchen object“ in a fully unsupervised fashion. These models offer a parsimonious account of how abstract semantic knowledge may be rooted in children's embodied first-person experiences.

“Development and application of gaze control models for active perception”

Gaze shifts in humans serve to direct high-resolution vision provided by the fovea towards areas in the environment. Gaze can be considered a proxy for attention or indicator of the relative importance of different parts of the environment. In this talk, we discuss the development of generative models of human gaze in response to visual input. We discuss how such models can be learned, both using supervised learning and using implicit feedback as an agent interacts with the environment, the latter being more plausible in biological agents. We also discuss two ways such models can be used. First, they can be used to improve the performance of artificial autonomous systems, in applications such as autonomous navigation. Second, because these models are contingent on the human’s task, goals, and/or state in the context of the environment, observations of gaze can be used to infer information about user intent. This information can be used to improve human-machine and human robot interaction, by making interfaces more anticipative. We discuss example applications in gaze-typing, robotic tele-operation and human-robot interaction.

Developmental and evolutionary perspectives on thalamic function

Brain organization and function is a complex topic. We are good at establishing correlates of perception and behavior across forebrain circuits, as well as manipulating activity in these circuits to affect behavior. However, we still lack good models for the large-scale organization and function of the forebrain. What are the contributions of the cortex, basal ganglia, and thalamus to behavior? In addressing these questions, we often ascribe function to each area as if it were an independent processing unit. However, we know from the anatomy that the cortex, basal ganglia, and thalamus, are massively interconnected in a large network. One way to generate insight into these questions is to consider the evolution and development of forebrain systems. In this talk, I will discuss the developmental and evolutionary (comparative anatomy) data on the thalamus, and how it fits within forebrain networks. I will address questions including, when did the thalamus appear in evolution, how is the thalamus organized across the vertebrate lineage, and how can the change in the organization of forebrain networks affect behavioral repertoires.

An Ecological and Objective Neural Marker of Implicit Unfamiliar Identity Recognition

We developed a novel paradigm measuring implicit identity recognition using Fast Periodic Visual Stimulation (FPVS) with EEG among 16 students and 12 police officers with normal face processing abilities. Participants' neural responses to a 1-Hz tagged oddball identity embedded within a 6-Hz image stream revealed implicit recognition with high-quality mugshots but not CCTV-like images, suggesting optimal resolution requirements. Our findings extend previous research by demonstrating that even unfamiliar identities can elicit robust neural recognition signatures through brief, repeated passive exposure. This approach offers potential for objective validation of face processing abilities in forensic applications, including assessment of facial examiners, Super-Recognisers, and eyewitnesses, potentially overcoming limitations of traditional behavioral assessment methods.

Open Hardware Microfluidics

What’s the point of having scientific and technological innovations when only a few can benefit from them? How can we make science more inclusive? Those questions are always in the back of my mind when we perform research in our laboratory, and we have a strong focus on the scientific accessibility of our developed methods from microfabrication to sensor development.

Restoring Sight to the Blind: Effects of Structural and Functional Plasticity

Visual restoration after decades of blindness is now becoming possible by means of retinal and cortical prostheses, as well as emerging stem cell and gene therapeutic approaches. After restoring visual perception, however, a key question remains. Are there optimal means and methods for retraining the visual cortex to process visual inputs, and for learning or relearning to “see”? Up to this point, it has been largely assumed that if the sensory loss is visual, then the rehabilitation focus should also be primarily visual. However, the other senses play a key role in visual rehabilitation due to the plastic repurposing of visual cortex during blindness by audition and somatosensation, and also to the reintegration of restored vision with the other senses. I will present multisensory neuroimaging results, cortical thickness changes, as well as behavioral outcomes for patients with Retinitis Pigmentosa (RP), which causes blindness by destroying photoreceptors in the retina. These patients have had their vision partially restored by the implantation of a retinal prosthesis, which electrically stimulates still viable retinal ganglion cells in the eye. Our multisensory and structural neuroimaging and behavioral results suggest a new, holistic concept of visual rehabilitation that leverages rather than neglects audition, somatosensation, and other sensory modalities.

“A Focus on 3D Printed Lenses: Rapid prototyping, low-cost microscopy and enhanced imaging for the life sciences”

High-quality glass lenses are commonplace in the design of optical instrumentation used across the biosciences. However, research-grade glass lenses are often costly, delicate and, depending on the prescription, can involve intricate and lengthy manufacturing - even more so in bioimaging applications. This seminar will outline 3D printing as a viable low-cost alternative for the manufacture of high-performance optical elements, where I will also discuss the creation of the world’s first fully 3D printed microscope and other implementations of 3D printed lenses. Our 3D printed lenses were generated using consumer-grade 3D printers and pose a 225x materials cost-saving compared to glass optics. Moreover, they can be produced in any lab or home environment and offer great potential for education and outreach. Following performance validation, our 3D printed optics were implemented in the production of a fully 3D printed microscope and demonstrated in histological imaging applications. We also applied low-cost fabrication methods to exotic lens geometries to enhance resolution and contrast across spatial scales and reveal new biological structures. Across these applications, our findings showed that 3D printed lenses are a viable substitute for commercial glass lenses, with the advantage of being relatively low-cost, accessible, and suitable for use in optical instruments. Combining 3D printed lenses with open-source 3D printed microscope chassis designs opens the doors for low-cost applications for rapid prototyping, low-resource field diagnostics, and the creation of cheap educational tools.

Functional Plasticity in the Language Network – evidence from Neuroimaging and Neurostimulation

Efficient cognition requires flexible interactions between distributed neural networks in the human brain. These networks adapt to challenges by flexibly recruiting different regions and connections. In this talk, I will discuss how we study functional network plasticity and reorganization with combined neurostimulation and neuroimaging across the adult life span. I will argue that short-term plasticity enables flexible adaptation to challenges, via functional reorganization. My key hypothesis is that disruption of higher-level cognitive functions such as language can be compensated for by the recruitment of domain-general networks in our brain. Examples from healthy young brains illustrate how neurostimulation can be used to temporarily interfere with efficient processing, probing short-term network plasticity at the systems level. Examples from people with dyslexia help to better understand network disorders in the language domain and outline the potential of facilitatory neurostimulation for treatment. I will also discuss examples from aging brains where plasticity helps to compensate for loss of function. Finally, examples from lesioned brains after stroke provide insight into the brain’s potential for long-term reorganization and recovery of function. Collectively, these results challenge the view of a modular organization of the human brain and argue for a flexible redistribution of function via systems plasticity.

Using Fast Periodic Visual Stimulation to measure cognitive function in dementia

Fast periodic visual stimulation (FPVS) has emerged as a promising tool for assessing cognitive function in individuals with dementia. This technique leverages electroencephalography (EEG) to measure brain responses to rapidly presented visual stimuli, offering a non-invasive and objective method for evaluating a range of cognitive functions. Unlike traditional cognitive assessments, FPVS does not rely on behavioural responses, making it particularly suitable for individuals with cognitive impairment. In this talk I will highlight a series of studies that have demonstrated its ability to detect subtle deficits in recognition memory, visual processing and attention in dementia patients using EEG in the lab, at home and in clinic. The method is quick, cost-effective, and scalable, utilizing widely available EEG technology. FPVS holds significant potential as a functional biomarker for early diagnosis and monitoring of dementia, paving the way for timely interventions and improved patient outcomes.

Simulating Thought Disorder: Fine-Tuning Llama-2 for Synthetic Speech in Schizophrenia

Computational modelling of ocular pharmacokinetics

Pharmacokinetics in the eye is an important factor for the success of ocular drug delivery and treatment. Pharmacokinetic features determine the feasible routes of drug administration, dosing levels and intervals, and it has impact on eventual drug responses. Several physical, biochemical, and flow-related barriers limit drug exposure of anterior and posterior ocular target tissues during treatment during local (topical, subconjunctival, intravitreal) and systemic administration (intravenous, per oral). Mathematical models integrate joint impact of various barriers on ocular pharmacokinetics (PKs) thereby helping drug development. The models are useful in describing (top-down) and predicting (bottom-up) pharmacokinetics of ocular drugs. This is useful also in the design and development of new drug molecules and drug delivery systems. Furthermore, the models can be used for interspecies translation and probing of disease effects on pharmacokinetics. In this lecture, ocular pharmacokinetics and current modelling methods (noncompartmental analyses, compartmental, physiologically based, and finite element models) are introduced. Future challenges are also highlighted (e.g. intra-tissue distribution, prediction of drug responses, active transport).

Neurosurgery & Consciousness: Bridging Science and Philosophy in the Age of AI

Overview of neurosurgery specialty interplay between neurology, psychiatry and neurosurgery. Discussion on benefits and disadvantages of classifications. Presentation of sub-specialties: trauma, oncology, functional, pediatric, vascular and spine. How does an ordinary day of a neurosurgeon look like; outpatient clinic, emergencies, pre/intra/post operative patient care. An ordinary operation. Myth-busting and practical insights of every day practice. An ordinary operation. Hint for research on clinical problems to be solved. The coming ethical frontiers of neuroprosthetics. In part two we will explore the explanatory gap and its significance. We will review the more than 200 theories of the hard problem of consciousness, from the prevailing to the unconventional. Finally, we are going to reflect on the AI advancements and the claims of LLMs becoming conscious

Resonancia Magnética y Detección Remota: No se Necesita Estar tan Cerca”

La resonancia magnética nuclear está basada en el fenómeno del magnetismo nuclear que más aplicaciones ha encontrado para el estudio de enfermedades humanas. Usualmente la señal de RM es recibida y transmitida a distancias cercanas al objeto del que se quiere obtener una imagen. Otra alternativa es emitir y recibir la misma señal de manera remota haciendo uso de guías de onda. Este enfoque tiene la ventaja que se puede aplicar a altos campos magnéticos, la absorción de energía es menor, además es posible cubrir mayores regiones de interés y comodidad para el paciente. Por otro lado, sufre de baja calidad de imagen en algunos casos. En esta ocasión hablaremos de nuestra experiencia haciendo uso de este enfoque empleando una guía de ondas abierta y metamateriales tanto para sistemas clínicos como preclínicos de IRM.

Cognitive maps as expectations learned across episodes – a model of the two dentate gyrus blades

How can the hippocampal system transition from episodic one-shot learning to a multi-shot learning regime and what is the utility of the resultant neural representations? This talk will explore the role of the dentate gyrus (DG) anatomy in this context. The canonical DG model suggests it performs pattern separation. More recent experimental results challenge this standard model, suggesting DG function is more complex and also supports the precise binding of objects and events to space and the integration of information across episodes. Very recent studies attribute pattern separation and pattern integration to anatomically distinct parts of the DG (the suprapyramidal blade vs the infrapyramidal blade). We propose a computational model that investigates this distinction. In the model the two processing streams (potentially localized in separate blades) contribute to the storage of distinct episodic memories, and the integration of information across episodes, respectively. The latter forms generalized expectations across episodes, eventually forming a cognitive map. We train the model with two data sets, MNIST and plausible entorhinal cortex inputs. The comparison between the two streams allows for the calculation of a prediction error, which can drive the storage of poorly predicted memories and the forgetting of well-predicted memories. We suggest that differential processing across the DG aids in the iterative construction of spatial cognitive maps to serve the generation of location-dependent expectations, while at the same time preserving episodic memory traces of idiosyncratic events.

Vision for perception versus vision for action: dissociable contributions of visual sensory drives from primary visual cortex and superior colliculus neurons to orienting behaviors

The primary visual cortex (V1) directly projects to the superior colliculus (SC) and is believed to provide sensory drive for eye movements. Consistent with this, a majority of saccade-related SC neurons also exhibit short-latency, stimulus-driven visual responses, which are additionally feature-tuned. However, direct neurophysiological comparisons of the visual response properties of the two anatomically-connected brain areas are surprisingly lacking, especially with respect to active looking behaviors. I will describe a series of experiments characterizing visual response properties in primate V1 and SC neurons, exploring feature dimensions like visual field location, spatial frequency, orientation, contrast, and luminance polarity. The results suggest a substantial, qualitative reformatting of SC visual responses when compared to V1. For example, SC visual response latencies are actively delayed, independent of individual neuron tuning preferences, as a function of increasing spatial frequency, and this phenomenon is directly correlated with saccadic reaction times. Such “coarse-to-fine” rank ordering of SC visual response latencies as a function of spatial frequency is much weaker in V1, suggesting a dissociation of V1 responses from saccade timing. Consistent with this, when we next explored trial-by-trial correlations of individual neurons’ visual response strengths and visual response latencies with saccadic reaction times, we found that most SC neurons exhibited, on a trial-by-trial basis, stronger and earlier visual responses for faster saccadic reaction times. Moreover, these correlations were substantially higher for visual-motor neurons in the intermediate and deep layers than for more superficial visual-only neurons. No such correlations existed systematically in V1. Thus, visual responses in SC and V1 serve fundamentally different roles in active vision: V1 jumpstarts sensing and image analysis, but SC jumpstarts moving. I will finish by demonstrating, using V1 reversible inactivation, that, despite reformatting of signals from V1 to the brainstem, V1 is still a necessary gateway for visually-driven oculomotor responses to occur, even for the most reflexive of eye movement phenomena. This is a fundamental difference from rodent studies demonstrating clear V1-independent processing in afferent visual pathways bypassing the geniculostriate one, and it demonstrates the importance of multi-species comparisons in the study of oculomotor control.

Brain macrophage transplantation for research and therapy development

Towards open meta-research in neuroimaging

When meta-research (research on research) makes an observation or points out a problem (such as a flaw in methodology), the project should be repeated later to determine whether the problem remains. For this we need meta-research that is reproducible and updatable, or living meta-research. In this talk, we introduce the concept of living meta-research, examine prequels to this idea, and point towards standards and technologies that could assist researchers in doing living meta-research. We introduce technologies like natural language processing, which can help with automation of meta-research, which in turn will make the research easier to reproduce/update. Further, we showcase our open-source litmining ecosystem, which includes pubget (for downloading full-text journal articles), labelbuddy (for manually extracting information), and pubextract (for automatically extracting information). With these tools, you can simplify the tedious data collection and information extraction steps in meta-research, and then focus on analyzing the text. We will then describe some living meta-research projects to illustrate the use of these tools. For example, we’ll show how we used GPT along with our tools to extract information about study participants. Essentially, this talk will introduce you to the concept of meta-research, some tools for doing meta-research, and some examples. Particularly, we want you to take away the fact that there are many interesting open questions in meta-research, and you can easily learn the tools to answer them. Check out our tools at https://litmining.github.io/

The Brain Prize winners' webinar

This webinar brings together three leaders in theoretical and computational neuroscience—Larry Abbott, Haim Sompolinsky, and Terry Sejnowski—to discuss how neural circuits generate fundamental aspects of the mind. Abbott illustrates mechanisms in electric fish that differentiate self-generated electric signals from external sensory cues, showing how predictive plasticity and two-stage signal cancellation mediate a sense of self. Sompolinsky explores attractor networks, revealing how discrete and continuous attractors can stabilize activity patterns, enable working memory, and incorporate chaotic dynamics underlying spontaneous behaviors. He further highlights the concept of object manifolds in high-level sensory representations and raises open questions on integrating connectomics with theoretical frameworks. Sejnowski bridges these motifs with modern artificial intelligence, demonstrating how large-scale neural networks capture language structures through distributed representations that parallel biological coding. Together, their presentations emphasize the synergy between empirical data, computational modeling, and connectomics in explaining the neural basis of cognition—offering insights into perception, memory, language, and the emergence of mind-like processes.

Decision and Behavior

This webinar addressed computational perspectives on how animals and humans make decisions, spanning normative, descriptive, and mechanistic models. Sam Gershman (Harvard) presented a capacity-limited reinforcement learning framework in which policies are compressed under an information bottleneck constraint. This approach predicts pervasive perseveration, stimulus‐independent “default” actions, and trade-offs between complexity and reward. Such policy compression reconciles observed action stochasticity and response time patterns with an optimal balance between learning capacity and performance. Jonathan Pillow (Princeton) discussed flexible descriptive models for tracking time-varying policies in animals. He introduced dynamic Generalized Linear Models (Sidetrack) and hidden Markov models (GLM-HMMs) that capture day-to-day and trial-to-trial fluctuations in choice behavior, including abrupt switches between “engaged” and “disengaged” states. These models provide new insights into how animals’ strategies evolve under learning. Finally, Kenji Doya (OIST) highlighted the importance of unifying reinforcement learning with Bayesian inference, exploring how cortical-basal ganglia networks might implement model-based and model-free strategies. He also described Japan’s Brain/MINDS 2.0 and Digital Brain initiatives, aiming to integrate multimodal data and computational principles into cohesive “digital brains.”

LLMs and Human Language Processing

This webinar convened researchers at the intersection of Artificial Intelligence and Neuroscience to investigate how large language models (LLMs) can serve as valuable “model organisms” for understanding human language processing. Presenters showcased evidence that brain recordings (fMRI, MEG, ECoG) acquired while participants read or listened to unconstrained speech can be predicted by representations extracted from state-of-the-art text- and speech-based LLMs. In particular, text-based LLMs tend to align better with higher-level language regions, capturing more semantic aspects, while speech-based LLMs excel at explaining early auditory cortical responses. However, purely low-level features can drive part of these alignments, complicating interpretations. New methods, including perturbation analyses, highlight which linguistic variables matter for each cortical area and time scale. Further, “brain tuning” of LLMs—fine-tuning on measured neural signals—can improve semantic representations and downstream language tasks. Despite open questions about interpretability and exact neural mechanisms, these results demonstrate that LLMs provide a promising framework for probing the computations underlying human language comprehension and production at multiple spatiotemporal scales.

Brain-Wide Compositionality and Learning Dynamics in Biological Agents

Biological agents continually reconcile the internal states of their brain circuits with incoming sensory and environmental evidence to evaluate when and how to act. The brains of biological agents, including animals and humans, exploit many evolutionary innovations, chiefly modularity—observable at the level of anatomically-defined brain regions, cortical layers, and cell types among others—that can be repurposed in a compositional manner to endow the animal with a highly flexible behavioral repertoire. Accordingly, their behaviors show their own modularity, yet such behavioral modules seldom correspond directly to traditional notions of modularity in brains. It remains unclear how to link neural and behavioral modularity in a compositional manner. We propose a comprehensive framework—compositional modes—to identify overarching compositionality spanning specialized submodules, such as brain regions. Our framework directly links the behavioral repertoire with distributed patterns of population activity, brain-wide, at multiple concurrent spatial and temporal scales. Using whole-brain recordings of zebrafish brains, we introduce an unsupervised pipeline based on neural network models, constrained by biological data, to reveal highly conserved compositional modes across individuals despite the naturalistic (spontaneous or task-independent) nature of their behaviors. These modes provided a scaffolding for other modes that account for the idiosyncratic behavior of each fish. We then demonstrate experimentally that compositional modes can be manipulated in a consistent manner by behavioral and pharmacological perturbations. Our results demonstrate that even natural behavior in different individuals can be decomposed and understood using a relatively small number of neurobehavioral modules—the compositional modes—and elucidate a compositional neural basis of behavior. This approach aligns with recent progress in understanding how reasoning capabilities and internal representational structures develop over the course of learning or training, offering insights into the modularity and flexibility in artificial and biological agents.

Unmotivated bias

In this talk, I will explore how social affective biases arise even in the absence of motivational factors as an emergent outcome of the basic structure of social learning. In several studies, we found that initial negative interactions with some members of a group can cause subsequent avoidance of the entire group, and that this avoidance perpetuates stereotypes. Additional cognitive modeling discovered that approach and avoidance behavior based on biased beliefs not only influences the evaluative (positive or negative) impressions of group members, but also shapes the depth of the cognitive representations available to learn about individuals. In other words, people have richer cognitive representations of members of groups that are not avoided, akin to individualized vs group level categories. I will end presenting a series of multi-agent reinforcement learning simulations that demonstrate the emergence of these social-structural feedback loops in the development and maintenance of affective biases.

Imagining and seeing: two faces of prosopagnosia

Feedback-induced dispositional changes in risk preferences

Contrary to the original normative decision-making standpoint, empirical studies have repeatedly reported that risk preferences are affected by the disclosure of choice outcomes (feedback). Although no consensus has yet emerged regarding the properties and mechanisms of this effect, a widespread and intuitive hypothesis is that repeated feedback affects risk preferences by means of a learning effect, which alters the representation of subjective probabilities. Here, we ran a series of seven experiments (N= 538), tailored to decipher the effects of feedback on risk preferences. Our results indicate that the presence of feedback consistently increases risk-taking, even when the risky option is economically less advantageous. Crucially, risk-taking increases just after the instructions, before participants experience any feedback. These results challenge the learning account, and advocate for a dispositional effect, induced by the mere anticipation of feedback information. Epistemic curiosity and regret avoidance may drive this effect in partial and complete feedback conditions, respectively.

How Generative AI is Revolutionizing the Software Developer Industry

Generative AI is fundamentally transforming the software development industry by improving processes such as software testing, bug detection, bug fixes, and developer productivity. This talk explores how AI-driven techniques, particularly large language models (LLMs), are being utilized to generate realistic test scenarios, automate bug detection and repair, and streamline development workflows. As these technologies evolve, they promise to improve software quality and efficiency significantly. The discussion will cover key methodologies, challenges, and the future impact of generative AI on the software development lifecycle, offering a comprehensive overview of its revolutionary potential in the industry.

Comparing supervised learning dynamics: Deep neural networks match human data efficiency but show a generalisation lag

Recent research has seen many behavioral comparisons between humans and deep neural networks (DNNs) in the domain of image classification. Often, comparison studies focus on the end-result of the learning process by measuring and comparing the similarities in the representations of object categories once they have been formed. However, the process of how these representations emerge—that is, the behavioral changes and intermediate stages observed during the acquisition—is less often directly and empirically compared. In this talk, I'm going to report a detailed investigation of the learning dynamics in human observers and various classic and state-of-the-art DNNs. We develop a constrained supervised learning environment to align learning-relevant conditions such as starting point, input modality, available input data and the feedback provided. Across the whole learning process we evaluate and compare how well learned representations can be generalized to previously unseen test data. Comparisons across the entire learning process indicate that DNNs demonstrate a level of data efficiency comparable to human learners, challenging some prevailing assumptions in the field. However, our results also reveal representational differences: while DNNs' learning is characterized by a pronounced generalisation lag, humans appear to immediately acquire generalizable representations without a preliminary phase of learning training set-specific information that is only later transferred to novel data.

Optogenetic control of Nodal signaling patterns

Embryos issue instructions to their cells in the form of patterns of signaling activity. Within these patterns, the distribution of signaling in time and space directs the fate of embryonic cells. Tools to perturb developmental signaling with high resolution in space and time can help reveal how these patterns are decoded to make appropriate fate decisions. In this talk, I will present new optogenetic reagents and an experimental pipeline for creating designer Nodal signaling patterns in live zebrafish embryos. Our improved optoNodal reagents eliminate dark activity and improve response kinetics, without sacrificing dynamic range. We adapted an ultra-widefield microscopy platform for parallel light patterning in up to 36 embryos and demonstrated precise spatial control over Nodal signaling activity and downstream gene expression. Using this system, we demonstrate that patterned Nodal activation can initiate specification and internalization movements of endodermal precursors. Further, we used patterned illumination to generate synthetic signaling patterns in Nodal signaling mutants, rescuing several characteristic developmental defects. This study establishes an experimental toolkit for systematic exploration of Nodal signaling patterns in live embryos.

Physical Activity, Sedentary Behaviour and Brain Health

Why age-related macular degeneration is a mathematically tractable disease

Among all prevalent diseases with a central neurodegeneration, AMD can be considered the most promising in terms of prevention and early intervention, due to several factors surrounding the neural geometry of the foveal singularity. • Steep gradients of cell density, deployed in a radially symmetric fashion, can be modeled with a difference of Gaussian curves. • These steep gradients give rise to huge, spatially aligned biologic effects, summarized as the Center of Cone Resilience, Surround of Rod Vulnerability. • Widely used clinical imaging technology provides cellular and subcellular level information. • Data are now available at all timelines: clinical, lifespan, evolutionary • Snapshots are available from tissues (histology, analytic chemistry, gene expression) • A viable biogenesis model exists for drusen, the largest population-level intraocular risk factor for progression. • The biogenesis model shares molecular commonality with atherosclerotic cardiovascular disease, for which there has been decades of public health success. • Animal and cell model systems are emerging to test these ideas.

Llama 3.1 Paper: The Llama Family of Models

Modern artificial intelligence (AI) systems are powered by foundation models. This paper presents a new set of foundation models, called Llama 3. It is a herd of language models that natively support multilinguality, coding, reasoning, and tool usage. Our largest model is a dense Transformer with 405B parameters and a context window of up to 128K tokens. This paper presents an extensive empirical evaluation of Llama 3. We find that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks. We publicly release Llama 3, including pre-trained and post-trained versions of the 405B parameter language model and our Llama Guard 3 model for input and output safety. The paper also presents the results of experiments in which we integrate image, video, and speech capabilities into Llama 3 via a compositional approach. We observe this approach performs competitively with the state-of-the-art on image, video, and speech recognition tasks. The resulting models are not yet being broadly released as they are still under development.

Error Consistency between Humans and Machines as a function of presentation duration

Within the last decade, Deep Artificial Neural Networks (DNNs) have emerged as powerful computer vision systems that match or exceed human performance on many benchmark tasks such as image classification. But whether current DNNs are suitable computational models of the human visual system remains an open question: While DNNs have proven to be capable of predicting neural activations in primate visual cortex, psychophysical experiments have shown behavioral differences between DNNs and human subjects, as quantified by error consistency. Error consistency is typically measured by briefly presenting natural or corrupted images to human subjects and asking them to perform an n-way classification task under time pressure. But for how long should stimuli ideally be presented to guarantee a fair comparison with DNNs? Here we investigate the influence of presentation time on error consistency, to test the hypothesis that higher-level processing drives behavioral differences. We systematically vary presentation times of backward-masked stimuli from 8.3ms to 266ms and measure human performance and reaction times on natural, lowpass-filtered and noisy images. Our experiment constitutes a fine-grained analysis of human image classification under both image corruptions and time pressure, showing that even drastically time-constrained humans who are exposed to the stimuli for only two frames, i.e. 16.6ms, can still solve our 8-way classification task with success rates way above chance. We also find that human-to-human error consistency is already stable at 16.6ms.

Open source FPGA tools for building research devices

Edmund will present why to use FPGAs when building scientific instruments, when and why to use open source FPGA tools, the history of their development, their development status, currently supported FPGA families and functions, current developments in design languages and tools, the community, freely available design blocks, and possible future developments.

Experimental research in patients with migraine

Trends in NeuroAI - Brain-like topography in transformers (Topoformer)

Dr. Nicholas Blauch will present on his work "Topoformer: Brain-like topographic organization in transformer language models through spatial querying and reweighting". Dr. Blauch is a postdoctoral fellow in the Harvard Vision Lab advised by Talia Konkle and George Alvarez. Paper link: https://openreview.net/pdf?id=3pLMzgoZSA Trends in NeuroAI is a reading group hosted by the MedARC Neuroimaging & AI lab (https://medarc.ai/fmri | https://groups.google.com/g/medarc-fmri).

The role of mitopohagy in neuronal physiology

Generative models for video games (rescheduled)

Developing agents capable of modeling complex environments and human behaviors within them is a key goal of artificial intelligence research. Progress towards this goal has exciting potential for applications in video games, from new tools that empower game developers to realize new creative visions, to enabling new kinds of immersive player experiences. This talk focuses on recent advances of my team at Microsoft Research towards scalable machine learning architectures that effectively capture human gameplay data. In the first part of my talk, I will focus on diffusion models as generative models of human behavior. Previously shown to have impressive image generation capabilities, I present insights that unlock applications to imitation learning for sequential decision making. In the second part of my talk, I discuss a recent project taking ideas from language modeling to build a generative sequence model of an Xbox game.

Generative models for video games

Developing agents capable of modeling complex environments and human behaviors within them is a key goal of artificial intelligence research. Progress towards this goal has exciting potential for applications in video games, from new tools that empower game developers to realize new creative visions, to enabling new kinds of immersive player experiences. This talk focuses on recent advances of my team at Microsoft Research towards scalable machine learning architectures that effectively capture human gameplay data. In the first part of my talk, I will focus on diffusion models as generative models of human behavior. Previously shown to have impressive image generation capabilities, I present insights that unlock applications to imitation learning for sequential decision making. In the second part of my talk, I discuss a recent project taking ideas from language modeling to build a generative sequence model of an Xbox game.

Modeling human brain development and disease: the role of primary cilia

Neurodevelopmental disorders (NDDs) impose a global burden, affecting an increasing number of individuals. While some causative genes have been identified, understanding the human-specific mechanisms involved in these disorders remains limited. Traditional gene-driven approaches for modeling brain diseases have failed to capture the diverse and convergent mechanisms at play. Centrosomes and cilia act as intermediaries between environmental and intrinsic signals, regulating cellular behavior. Mutations or dosage variations disrupting their function have been linked to brain formation deficits, highlighting their importance, yet their precise contributions remain largely unknown. Hence, we aim to investigate whether the centrosome/cilia axis is crucial for brain development and serves as a hub for human-specific mechanisms disrupted in NDDs. Towards this direction, we first demonstrated species-specific and cell-type-specific differences in the cilia-genes expression during mouse and human corticogenesis. Then, to dissect their role, we provoked their ectopic overexpression or silencing in the developing mouse cortex or in human brain organoids. Our findings suggest that cilia genes manipulation alters both the numbers and the position of NPCs and neurons in the developing cortex. Interestingly, primary cilium morphology is disrupted, as we find changes in their length, orientation and number that lead to disruption of the apical belt and altered delamination profiles during development. Our results give insight into the role of primary cilia in human cortical development and address fundamental questions regarding the diversity and convergence of gene function in development and disease manifestation. It has the potential to uncover novel pharmacological targets, facilitate personalized medicine, and improve the lives of individuals affected by NDDs through targeted cilia-based therapies.

Improving Language Understanding by Generative Pre Training

Natural language understanding comprises a wide range of diverse tasks such as textual entailment, question answering, semantic similarity assessment, and document classification. Although large unlabeled text corpora are abundant, labeled data for learning these specific tasks is scarce, making it challenging for discriminatively trained models to perform adequately. We demonstrate that large gains on these tasks can be realized by generative pre-training of a language model on a diverse corpus of unlabeled text, followed by discriminative fine-tuning on each specific task. In contrast to previous approaches, we make use of task-aware input transformations during fine-tuning to achieve effective transfer while requiring minimal changes to the model architecture. We demonstrate the effectiveness of our approach on a wide range of benchmarks for natural language understanding. Our general task-agnostic model outperforms discriminatively trained models that use architectures specifically crafted for each task, significantly improving upon the state of the art in 9 out of the 12 tasks studied. For instance, we achieve absolute improvements of 8.9% on commonsense reasoning (Stories Cloze Test), 5.7% on question answering (RACE), and 1.5% on textual entailment (MultiNLI).

Distinctive features of experiential time: Duration, speed and event density

William James’s use of “time in passing” and “stream of thoughts” may be two sides of the same coin that emerge from the brain segmenting the continuous flow of information into discrete events. Departing from that idea, we investigated how the content of a realistic scene impacts two distinct temporal experiences: the felt duration and the speed of the passage of time. I will present you the results from an online study in which we used a well-established experimental paradigm, the temporal bisection task, which we extended to passage of time judgments. 164 participants classified seconds-long videos of naturalistic scenes as short or long (duration), or slow or fast (passage of time). Videos contained a varying number and type of events. We found that a large number of events lengthened subjective duration and accelerated the felt passage of time. Surprisingly, participants were also faster at estimating their felt passage of time compared to duration. The perception of duration heavily depended on objective duration, whereas the felt passage of time scaled with the rate of change. Altogether, our results support a possible dissociation of the mechanisms underlying the two temporal experiences.

Executive functions in the brain of deaf individuals – sensory and language effects

Executive functions are cognitive processes that allow us to plan, monitor and execute our goals. Using fMRI, we investigated how early deafness influences crossmodal plasticity and the organisation of executive functions in the adult human brain. Results from a range of visual executive function tasks (working memory, task switching, planning, inhibition) show that deaf individuals specifically recruit superior temporal “auditory” regions during task switching. Neural activity in auditory regions predicts behavioural performance during task switching in deaf individuals, highlighting the functional relevance of the observed cortical reorganisation. Furthermore, language grammatical skills were correlated with the level of activation and functional connectivity of fronto-parietal networks. Together, these findings show the interplay between sensory and language experience in the organisation of executive processing in the brain.

A Comprehensive Overview of Large Language Models

Large Language Models (LLMs) have recently demonstrated remarkable capabilities in natural language processing tasks and beyond. This success of LLMs has led to a large influx of research contributions in this direction. These works encompass diverse topics such as architectural innovations, better training strategies, context length improvements, fine-tuning, multi-modal LLMs, robotics, datasets, benchmarking, efficiency, and more. With the rapid development of techniques and regular breakthroughs in LLM research, it has become considerably challenging to perceive the bigger picture of the advances in this direction. Considering the rapidly emerging plethora of literature on LLMs, it is imperative that the research community is able to benefit from a concise yet comprehensive overview of the recent developments in this field. This article provides an overview of the existing literature on a broad range of LLM-related concepts. Our self-contained comprehensive overview of LLMs discusses relevant background concepts along with covering the advanced topics at the frontier of research in LLMs. This review article is intended to not only provide a systematic survey but also a quick comprehensive reference for the researchers and practitioners to draw insights from extensive informative summaries of the existing works to advance the LLM research.

Learning produces a hippocampal cognitive map in the form of an orthogonalized state machine

Cognitive maps confer animals with flexible intelligence by representing spatial, temporal, and abstract relationships that can be used to shape thought, planning, and behavior. Cognitive maps have been observed in the hippocampus, but their algorithmic form and the processes by which they are learned remain obscure. Here, we employed large-scale, longitudinal two-photon calcium imaging to record activity from thousands of neurons in the CA1 region of the hippocampus while mice learned to efficiently collect rewards from two subtly different versions of linear tracks in virtual reality. The results provide a detailed view of the formation of a cognitive map in the hippocampus. Throughout learning, both the animal behavior and hippocampal neural activity progressed through multiple intermediate stages, gradually revealing improved task representation that mirrored improved behavioral efficiency. The learning process led to progressive decorrelations in initially similar hippocampal neural activity within and across tracks, ultimately resulting in orthogonalized representations resembling a state machine capturing the inherent struture of the task. We show that a Hidden Markov Model (HMM) and a biologically plausible recurrent neural network trained using Hebbian learning can both capture core aspects of the learning dynamics and the orthogonalized representational structure in neural activity. In contrast, we show that gradient-based learning of sequence models such as Long Short-Term Memory networks (LSTMs) and Transformers do not naturally produce such orthogonalized representations. We further demonstrate that mice exhibited adaptive behavior in novel task settings, with neural activity reflecting flexible deployment of the state machine. These findings shed light on the mathematical form of cognitive maps, the learning rules that sculpt them, and the algorithms that promote adaptive behavior in animals. The work thus charts a course toward a deeper understanding of biological intelligence and offers insights toward developing more robust learning algorithms in artificial intelligence.

Dyslexia, Rhythm, Language and the Developing Brain

Are integrative, multidisciplinary, and pragmatic models possible? The #PsychMapping experience

This presentation delves into the necessity for simplified models in the field of psychological sciences to cater to a diverse audience of practitioners. We introduce the #PsychMapping model, evaluate its merits and limitations, and discuss its place in contemporary scientific culture. The #PsychMapping model is the product of an extensive literature review, initially within the realm of sport and exercise psychology and subsequently encompassing a broader spectrum of psychological sciences. This model synthesizes the progress made in psychological sciences by categorizing variables into a framework that distinguishes between traits (e.g., body structure and personality) and states (e.g., heart rate and emotions). Furthermore, it delineates internal traits and states from the externalized self, which encompasses behaviour and performance. All three components—traits, states, and the externalized self—are in a continuous interplay with external physical, social, and circumstantial factors. Two core processes elucidate the interactions among these four primary clusters: external perception, encompassing the mechanism through which external stimuli transition into internal events, and self-regulation, which empowers individuals to become autonomous agents capable of exerting control over themselves and their actions. While the model inherently oversimplifies intricate processes, the central question remains: does its pragmatic utility outweigh its limitations, and can it serve as a valuable tool for comprehending human behaviour?

Conversations with Caves? Understanding the role of visual psychological phenomena in Upper Palaeolithic cave art making

How central were psychological features deriving from our visual systems to the early evolution of human visual culture? Art making emerged deep in our evolutionary history, with the earliest art appearing over 100,000 years ago as geometric patterns etched on fragments of ochre and shell, and figurative representations of prey animals flourishing in the Upper Palaeolithic (c. 40,000 – 15,000 years ago). The latter reflects a complex visual process; the ability to represent something that exists in the real world as a flat, two-dimensional image. In this presentation, I argue that pareidolia – the psychological phenomenon of seeing meaningful forms in random patterns, such as perceiving faces in clouds – was a fundamental process that facilitated the emergence of figurative representation. The influence of pareidolia has often been anecdotally observed in Upper Palaeolithic art examples, particularly cave art where the topographic features of cave wall were incorporated into animal depictions. Using novel virtual reality (VR) light simulations, I tested three hypotheses relating to pareidolia in the caves of Upper Palaeolithic cave art in the caves of Las Monedas and La Pasiega (Cantabria, Spain). To evaluate this further, I also developed an interdisciplinary VR eye-tracking experiment, where participants were immersed in virtual caves based on the cave of El Castillo (Cantabria, Spain). Together, these case studies suggest that pareidolia was an intrinsic part of artist-cave interactions (‘conversations’) that influenced the form and placement of figurative depictions in the cave. This has broader implications for conceiving of the role of visual psychological phenomena in the emergence and development of figurative art in the Palaeolithic.

Deep non-linear mixed effects modelling of voltage-gated potassium channels

Bernstein Conference 2024

Age Effects on Eye Blink-Related Neural Activity and Functional Connectivity in Driving

Bernstein Conference 2024

Efficient nonlinear receptive field estimation across processing stages of sensory systems

Bernstein Conference 2024

Neuronal bursting from an interplay of fast voltage and slow concentration dynamics mediated by the Na+/K+-ATPase

Bernstein Conference 2024

Open-source solutions for research data management in neuroscience collaborations

Bernstein Conference 2024

Probing right-hemispheric neuronal representations in the language network of an individual with aphasia

Bernstein Conference 2024

tDCS montage optimization for the treatment of epilepsy using Neurotwins

Bernstein Conference 2024

Accurate Engagement of the Drosophila Central-Complex Compass During Head-Fixed Path-Constrained Navigation

COSYNE 2022

Emergence of an orientation map in the mouse superior colliculus from stage III retinal waves

COSYNE 2022

Emergent behavior and neural dynamics in artificial agents tracking turbulent plumes

COSYNE 2022

Engagement of the respiratory CPG for songbird vocalizations

COSYNE 2022

Integrating deep reinforcement learning agents with the C. elegans nervous system

COSYNE 2022

Integrating deep reinforcement learning agents with the C. elegans nervous system

COSYNE 2022

Mind the gradient: context-dependent selectivity to natural images in the retina revealed with a novel perturbative approach

COSYNE 2022

Mind the gradient: context-dependent selectivity to natural images in the retina revealed with a novel perturbative approach

COSYNE 2022

Predictive processing of natural images by V1 firing rates revealed by self-supervised deep neural networks

COSYNE 2022

Predictive processing of natural images by V1 firing rates revealed by self-supervised deep neural networks

COSYNE 2022

The smart image compression algorithm in the retina: recoding inputs in neural circuits

COSYNE 2022

The smart image compression algorithm in the retina: recoding inputs in neural circuits

COSYNE 2022

Statistics of sub-threshold voltage dynamics in cortical networks

COSYNE 2022

Statistics of sub-threshold voltage dynamics in cortical networks

COSYNE 2022

Time cell encoding in deep reinforcement learning agents depends on mnemonic demands

COSYNE 2022

Time cell encoding in deep reinforcement learning agents depends on mnemonic demands

COSYNE 2022

Using Markov Decision Processes to benchmark the performance of artificial and biological agents

COSYNE 2022

Using Markov Decision Processes to benchmark the performance of artificial and biological agents

COSYNE 2022

What do meta-reinforcement learning networks learn in two-stage decision-making?

COSYNE 2022

What do meta-reinforcement learning networks learn in two-stage decision-making?

COSYNE 2022

Alignment of ANN Language Models with Humans After a Developmentally Realistic Amount of Training

COSYNE 2023

“Attentional fingerprints” in conceptual space: Reliable, individuating patterns of visual attention revealed using natural language modeling

COSYNE 2023

Cortical dopamine enables deep reinforcement learning and leverages dopaminergic heterogeneity

COSYNE 2023

Exploring the role of image domains in self-supervised DNN models of the rodent brain

COSYNE 2023

Infinite storage in quasi-memory: a cryptographic principle underlining caching behavior in animals

COSYNE 2023

Input-dominated Hebbian learning enables image-computable E-I networks

COSYNE 2023

Language emergence in reinforcement learning agents performing navigational tasks

COSYNE 2023

A Large Dataset of Macaque V1 Responses to Natural Images Revealed Complexity in V1 Neural Codes

COSYNE 2023

Pre-training artificial neural networks with spontaneous retinal activity improves image prediction

COSYNE 2023

Comparing image representations in terms of sensitivities to local distortions

COSYNE 2025

A Data-Driven Approach to Estimating Animal Vocal Repertoires and their Usage

COSYNE 2025

Deep reinforcement learning trains agents to track odor plumes with active sensing

COSYNE 2025

Competition and integration of sensory signals in a deep reinforcement learning agent

Bernstein Conference 2024