The internship aims to develop a controller for a social mobile robot to have a conversation with people using large language models (LLMs) such as ChatGPT. The internship is part of the European SPRING project, which aims to develop mobile robots for healthcare environments. The intern will develop a controller (Python, ROS) for ARI, the social robot. The controller will navigate towards a human (or group), have a conversation with them, and leave the conversation. The intern will use existing components from the SPRING project such as mapping and localization of the robot and humans, human-aware navigation, speech recognition, and a simple dialogue system based on ChatGPT. The intern will also investigate how to optimally use LLMs such as ChatGPT for natural and comfortable conversation with the robot, for example, by using prompt engineering. The intern will have the chance to develop and implement their own ideas to improve the conversation with the robot, for example, by investigating gaze, gestures, or emotions.

In the second of this year’s Brain Prize webinars, Rob Singer (Einstein Medical College, USA), Florence Besse (Institut de Biologie Valrose, France) and Jennifer Lippincott-Schwartz (Janelia Farm Research Campus, USA) will present their work on mRNA transport, trafficking, and localization. Each speaker will present for 25 minutes, and the webinar will conclude with an open discussion. The webinar will be moderated by the winners of the 2023 Brain Prize, Michael Greenberg, Erin Schuman and Christine Holt.

The advent of next generation sequencing ushered in a ten-year period of exuberant technology development, enabling the quantification of gene expression and epigenetic features within individual cells, and within intact tissue sections.  In this seminar, I will outline our technological contributions, beginning with the development of Drop-seq, a method for high-throughput single cell analysis, followed by the development of Slide-seq, a technique for measuring genome-wide expression at 10 micron spatial resolution.  Using a combination of these techniques, we recently constructed a comprehensive cell type atlas of the adult mouse brain, positioning cell types within individual brain structures.  I will discuss the major findings from this dataset, including emerging principles of neurotransmission, and the localization of disease gene signatures to specific cell types.  Finally, I will introduce a new spatial technology, Slide-tags, that unifies single cell and spatial genomics into a single, highly scalable assay.

There is growing recognition that host-associated microbiotas modulate intrinsic neurodevelopmental programs including those underlying human social behavior. Despite this awareness, the fundamental processes are generally not understood. We discovered that the zebrafish microbiota is necessary for normal social behavior. By examining neuronal correlates of behavior, we found that the microbiota restrains neurite complexity and targeting of key forebrain neurons within the social behavior circuitry. The microbiota is also necessary for both localization and molecular functions of forebrain microglia, brain-resident phagocytes that remodel neuronal arbors. In particular, the microbiota promotes expression of complement signaling pathway components important for synapse remodeling. Our work provides evidence that the microbiota modulates zebrafish social behavior by stimulating microglial remodeling of forebrain circuits during early neurodevelopment and suggests molecular pathways for therapeutic interventions during atypical neurodevelopment.

The development of circuit-based therapeutics to treat neurological and neuropsychiatric diseases require detailed localization and understanding of electrophysiological signals in the human brain. Electrodes can record and stimulate circuits in many ways, and we often rely on non-invasive imaging methods to predict the location to implant electrodes. However, electrophysiological and imaging signals measure the underlying tissue in a fundamentally different manner. To integrate multimodal data and benefit from these complementary measurements, I will describe an approach that considers how different measurements integrate signals across the underlying tissue. I will show how this approach helps relate fMRI and intracranial EEG measurements and provides new insights into how electrical stimulation influences human brain networks.

With recent advances in learning algorithms, recurrent networks of spiking neurons are achieving performance competitive with standard recurrent neural networks. Still, these learning algorithms are limited to small networks of simple spiking neurons and modest-length temporal sequences, as they impose high memory requirements, have difficulty training complex neuron models, and are incompatible with online learning.Taking inspiration from the concept of Liquid Time-Constant (LTCs), we introduce a novel class of spiking neurons, the Liquid Time-Constant Spiking Neuron (LTC-SN), resulting in functionality similar to the gating operation in LSTMs. We integrate these neurons in SNNs that are trained with FPTT and demonstrate that thus trained LTC-SNNs outperform various SNNs trained with BPTT on long sequences while enabling online learning and drastically reducing memory complexity. We show this for several classical benchmarks that can easily be varied in sequence length, like the Add Task and the DVS-gesture benchmark. We also show how FPTT-trained LTC-SNNs can be applied to large convolutional SNNs, where we demonstrate novel state-of-the-art for online learning in SNNs on a number of standard benchmarks (S-MNIST, R-MNIST, DVS-GESTURE) and also show that large feedforward SNNs can be trained successfully in an online manner to near (Fashion-MNIST, DVS-CIFAR10) or exceeding (PS-MNIST, R-MNIST) state-of-the-art performance as obtained with offline BPTT. Finally, the training and memory efficiency of FPTT enables us to directly train SNNs in an end-to-end manner at network sizes and complexity that was previously infeasible: we demonstrate this by training in an end-to-end fashion the first deep and performant spiking neural network for object localization and recognition. Taken together, we out contribution enable for the first time training large-scale complex spiking neural network architectures online and on long temporal sequences.

Animals modify their behavioral outputs in response to changes in external and internal environments. We use the nematode, C. elegans to probe the pathways linking changes in internal states like hunger with behavior. We find that acute food deprivation alters the localization of two transcription factors, likely releasing an insulin-like peptide from the intestine, which in turn modifies chemosensory neurons and alters behavior. These results present a model for how inter-tissue signals to generate flexible behaviors via gut-brain signaling.

The complex morphology of neurons, with synapses located hundreds of microns from the cell body, necessitates the localization of important cell biological machines, including ribosomes, within dendrites and axons. Local translation of mRNAs is important for the function and plasticity of synapses. Using advanced sequencing and imaging techniques we have updated our understanding of the local transcriptome and identified the local translatome- identifying over 800 transcripts for which local translation is the dominant source of protein. In addition, we have explored the unique mechanisms neurons use to meet protein demands at synapses, identifying surprising features of neuronal and synaptic protein synthesis.

Glymphatic perivascular exchange is supported by the astroglial water channel aquaporin-4 (AQP4), which localizes to perivascular astrocytic endfeet surrounding the cerebral vasculature. In aging mice, impairment of glymphatic function is associated with reduced perivascular AQP4 localization, yet whether these changes contribute to the development of neurodegenerative disease, such as Alzheimer’s disease (AD), remains unknown. Using post mortem human tissue, we evaluated perivascular AQP4 localization in the frontal cortical gray matter, white matter, and hippocampus of cognitively normal subjects and those with AD. Loss of perivascular and increasing cellular localization of AQP4 in the frontal gray matter was specifically associated with AD status, amyloid β (Aβ) and tau pathology, and cognitive decline in the early stages of disease. Using AAV-PHP.B to drive expression on non-perivascular AQP4 in wild type and Tg2576 (APPSwe, mouse model of Aβ deposition) mice, increased cellular AQP4 localization did not slow glymphatic function or change Aβ deposition. Using the Snta1 knockout line (which lacks perivascular AQP4 localization), we observed that loss AQP4 from perivascular endfeet slowed glymphatic function in wild type mice and accelerated Aβ plaque deposition in Tg2576 mice. These findings demonstrate that loss of perivascular AQP4 localization, and not increased cellular AQP4 localization, slows glymphatic function and promotes the development of AD pathology. To evaluate whether naturally occurring variation in the human AQP4 gene, or the alpha syntrophin (SNTA1), dystrobrevin (DTNA) or dystroglycan (DAG1) genes (whose products maintain perivascular AQP4 localization) confer risk for or protection from AD pathology or clinical progression, we evaluated 56 tag single nucleotide polymorphisms (SNPs) across these genes for association with CSF AD biomarkers, MRI measures of cortical and hippocampal atrophy, and longitudinal cognitive decline in the Alzheimer’s Disease Neuroimaging Initiative I (ADNI I) cohort. We identify 25 different significant associations between AQP4, SNTA1, DTNA, and DAG1 tag SNPs and phenotypic measures of AD pathology and progression. These findings provide complimentary human genetic evidence for the contribution of perivascular glymphatic dysfunction to the development of AD in human populations.

Human vision is full of puzzles. Observers can grasp the essence of a scene in an instant, yet when probed for details they are at a loss. People have trouble finding their keys, yet they may be quite visible once found. How does one explain this combination of marvelous successes with quirky failures? I will describe our attempts to develop a unifying theory that brings a satisfying order to multiple phenomena. One key is to understand peripheral vision. A visual system cannot process everything with full fidelity, and therefore must lose some information. Peripheral vision must condense a mass of information into a succinct representation that nonetheless carries the information needed for vision at a glance. We have proposed that the visual system deals with limited capacity in part by representing its input in terms of a rich set of local image statistics, where the local regions grow — and the representation becomes less precise — with distance from fixation. This scheme trades off computation of sophisticated image features at the expense of spatial localization of those features. What are the implications of such an encoding scheme? Critical to our understanding has been the use of methodologies for visualizing the equivalence classes of the model. These visualizations allow one to quickly see that many of the puzzles of human vision may arise from a single encoding mechanism. They have suggested new experiments and predicted unexpected phenomena. Furthermore, visualization of the equivalence classes has facilitated the generation of testable model predictions, allowing us to study the effects of this relatively low-level encoding on a wide range of higher-level tasks. Peripheral vision helps explain many of the puzzles of vision, but some remain. By examining the phenomena that cannot be explained by peripheral vision, we gain insight into the nature of additional capacity limits in vision. In particular, I will suggest that decision processes face general-purpose limits on the complexity of the tasks they can perform at a given time.

We will highlight recent technological and methodological advances in deploying miniaturized technologies that can monitor the spatial electrophysiologic patterns of the visceral nervous system. As an example, we will discuss recent developments of thin, stretchable, wireless biosensor patches that can be embedded within routinely used medical adhesives for recording electrophysiologic patterns of the GI tract. We will also showcase recent developments in array signal processing that enable non-invasive tracking, and source localization, of the slow wave patterns associated with the GI tract. We will illustrate how such systems can also be used in tandem with novel miniaturized pacing devices to can enable closed-loop neuromodulation of the enteric nervous system. We will conclude with a summary of the knowns and unknowns in how multi-organ physiology research, technology miniaturization, and data science may create unique opportunities for the intersection of electrical engineering and neuroscience.

Perception can be strongly serially-dependent (i.e. biased toward previously seen stimuli). Recently, serial dependencies in perception were proposed as a mechanism for perceptual stability, increasing the apparent continuity of the complex environments we experience in everyday life. For example, stable scene perception can be actively achieved by the visual system through global serial dependencies, a special kind of serial dependence between summary statistical representations. Serial dependence occurs also between emotional expressions, but it is highly selective for the same identity. Overall, these results further support the notion of serial dependence as a global, highly specialized, and purposeful mechanism. However, serial dependence could also be a deleterious phenomenon in unnatural or unpredictable situations, such as visual search in radiological scans, biasing current judgments toward previous ones even when accurate and unbiased perception is needed. For example, observers make consistent perceptual errors when classifying a tumor- like shape on the current trial, seeing it as more similar to the shape presented on the previous trial. In a separate localization test, observers make consistent errors when reporting the perceived position of an objects on the current trial, mislocalizing it toward the position in the preceding trial. Taken together, these results show two opposite sides of serial dependence; it can be a beneficial mechanism which promotes perceptual stability, but at the same time a deleterious mechanism which impairs our percept when fine recognition is needed.

Outside the foveola, visual acuity and other visual functions gradually deteriorate with increasing eccentricity. Humans compensate for these limitations by relying on a tight link between perception and action; rapid gaze shifts (saccades) occur 2-3 times every second, separating brief “fixation” intervals in which visual information is acquired and processed. During fixation, however, the eye is not immobile. Small eye movements incessantly shift the image on the retina even when the attended stimulus is already foveated, suggesting a much deeper coupling between visual functions and oculomotor activity. Thanks to a combination of techniques allowing for high-resolution recordings of eye position, retinal stabilization, and accurate gaze localization, we examined how attention and eye movements are controlled at this scale. We have shown that during fixation, visual exploration of fine spatial detail unfolds following visuomotor strategies similar to those occurring at a larger scale. This behavior compensates for non-homogenous visual capabilities within the foveola and is finely controlled by attention, which facilitates processing at selected foveal locations. Ultimately, the limits of high acuity vision are greatly influenced by the spatiotemporal modulations introduced by fixational eye movements. These findings reveal that, contrary to common intuition, placing a stimulus within the foveola is necessary but not sufficient for high visual acuity; fine spatial vision is the outcome of an orchestrated synergy of motor, cognitive, and attentional factors.

The precise spatial localization of molecular signals within tissues richly informs the mechanisms of tissue formation and function. Here, we’ll introduce Slide-seq, a technology which enables transcriptome-wide measurements with near-single cell spatial resolution. We’ll describe recent experimental and computational advances to enable Slide-seq in biological contexts in biological contexts where high detection sensitivity is important. More broadly, we’ll discuss the promise and challenges of spatial transcriptomics for tissue genomics. Lastly, we’ll touch upon novel molecular recording technologies, which allows recording of the absolute time dynamics of gene expression in live systems into DNA sequences.

The complex morphology of neurons, with synapses located 100’s of microns from the cell body, necessitates the localization of important cell biological machines and processes within dendrites and axons. Using expansion microscopy together with metabolic labeling we have discovered that both postsynaptic spines and presynaptic terminals exhibit rapid translation, which exhibits differential sensitivity to different neurotransmitters and neuromodulators. In addition, we have explored the unique mechanisms neurons use to meet protein demands at synapses, identifying the transcriptome and translatome in the neuropil.

Over the last several decades, the tractable response properties of parahippocampal neurons have provided a new access key to understanding the cognitive process of self-localization: the ability to know where you are currently located in space. Defined by functionally discrete response properties, neurons in the medial entorhinal cortex and hippocampus are proposed to provide the basis for an internal neural map of space, which enables animals to perform path-integration based spatial navigation and supports the formation of spatial memories. My lab focuses on understanding the mechanisms that generate this neural map of space and how this map is used to support behavior. In this talk, I’ll discuss how learning and experience shapes our internal neural maps of space to guide behavior.

My lab studies the design principles of cytoskeletal materials the drive cellular morphogenesis, with a focus on contractile machinery in adherent cells. In addition to force generation, a key feature of these materials are distributed force sensors which allow for rapid assembly, adaptation, repair and disintegration. Here I will discuss our recent identification of 18 proteins from the zyxin, paxillin, Tes and Enigma families with mechanosensitive LIM (Lin11, Isl- 1 & Mec-3) domains. We developed a screen to assess the force-dependent localization of LIM domain-containing region (LCR) from ~30 genes to the actin cytoskeleton and identified features common to their force-sensitive localization. Through in vitro reconstitution, we found that the LCR binds directly to mechanically stressed actin filaments. Moreover, the LCR from the fission yeast protein paxillin-like 1 is also mechanosensitive, suggesting force-sensitivity is highly conserved. We speculate that the evolutionary emergence of contractile F-actin machinery coincided with, or required, proteins that could report on the stresses present there to maintain homeostasis of actively stressed networks.

In many ways, cognitive neuroscience is the attempt to use physiological observation to clarify the mechanisms that shape behavior. Over the past 25 years, fMRI has provided a system-wide and yet somewhat spatially precise view of the response in human cortex evoked by a wide variety of stimuli and task contexts. The current talk focuses on the other direction of inference; the implications of this observed functional organization for behavior. To begin, we must interrogate the methodological and empirical frameworks underlying our derivation of this organization, partially by exploring its relationship to and predictability from gross neuroanatomy. Next, across a series of studies, the implications of two properties of functional organization for behavior will be explored: 1) the co-localization of visual working memory and perceptual processing and 2) implicit learning in the context of distributed responses. In sum, these results highlight the limitations of our current approach and hint at a new general mechanism for explaining observed behavior in context with the neural substrate.