Neuromatch 5 (Neuromatch Conference 2022) was a fully virtual conference focused on computational neuroscience broadly construed, including machine learning work with explicit biological links:contentReference[oaicite:11]{index=11}. After four successful Neuromatch conferences, the fifth edition consolidated proven innovations from past events, featuring a series of talks hosted on Crowdcast and flash talk sessions (pre-recorded videos) with dedicated discussion times on Reddit:contentReference[oaicite:12]{index=12}.

Neural representations of time are central to our understanding of the world around us. I review cognitive, neurophysiological and theoretical work that converges on three simple ideas. First, the time of past events is remembered via populations of neurons with a continuum of functional time constants. Second, these time constants evenly tile the log time axis. This results in a neural Weber-Fechner scale for time which can support behavioral Weber-Fechner laws and characteristic behavioral effects in memory experiments. Third, these populations appear as dual pairs---one type of population contains cells that change firing rate monotonically over time and a second type of population that has circumscribed temporal receptive fields. These ideas can be used to build artificial neural networks that have novel properties. Of particular interest, a convolutional neural network built using these principles can generalize to arbitrary rescaling of its inputs. That is, after learning to perform a classification task on a time series presented at one speed, it successfully classifies stimuli presented slowed down or sped up. This result illustrates the point that this confluence of ideas originating in cognitive psychology and measured in the mammalian brain could have wide-reaching impacts on AI research.

Brains must integrate complex sensory information and compare to past events to generate appropriate behavioral responses. The neural circuit basis of these computations is unclear and the underlying structure unknown. Here, we mapped the comprehensive synaptic wiring diagram of the fruit fly larva brain, which contains 3,013 neurons and 544K synaptic sites. It is the most complete insect connectome to date: 1) Both brain hemispheres are reconstructed, allowing investigation of neural pathways that include contralateral axons, which we found in 37% of brain neurons. 2) All sensory neurons and descending neurons are reconstructed, allowing one to follow signals in an uninterrupted chain—from the sensory periphery, through the brain, to motor neurons in the nerve cord. We developed novel computational tools, allowing us to cluster the brain and investigate how information flows through it. We discovered that feedforward pathways from sensory to descending neurons are multilayered and highly multimodal. Robust feedback was observed at almost all levels of the brain, including descending neurons. We investigated how the brain hemispheres communicate with each other and the nerve cord, leading to identification of novel circuit motifs. This work provides the complete blueprint of a brain and a strong foundation to study the structure-function relationship of neural circuits.

Many everyday decisions are based on both external cues and internal hunches. How does the brain put these together? We addressed this question in mice trained to make decisions based on sensory stimuli and on past events. While mice made these decisions, we causally probed the roles of cortical areas and recorded from thousands of neurons throughout the brain, with an emphasis on frontal cortex. The results are not what we thought based on textbook notions of how the brain works. This talk is based on work led by Nick Steinmetz, Peter Zatka-Haas, Armin Lak, and Pip Coen, in the laboratory I share with Kenneth Harris

This seminar is part of our “Emergent Scientists” series, an initiative that provides a platform for scientists at the critical PhD/postdoc transition period to share their work with a broad audience and network. Summary: These talks cover Alzheimer’s disease (AD) research in both mice and humans. Christiana will discuss in particular the translational aspects of applying mouse work to humans and the importance of timing in disease pathology and intervention (e.g. timing between AD biomarkers vs. symptom onset, timing of therapy, etc.). Siddharth will discuss a rare variant of Alzheimer’s disease called “Logopenic Progressive Aphasia”, which presents with temporo-parietal atrophy yet relative sparing of hippocampal circuitry. Siddharth will discuss how, despite the unusual anatomical basis underlying this AD variant, degeneration of the angular gyrus in the left inferior parietal lobule contributes to memory deficits similar to those of typical amnesic Alzheimer’s disease. Christiana’s abstract: Alzheimer’s disease (AD) is a debilitating neurodegenerative disorder that causes severe deterioration of memory, cognition, behavior, and the ability to perform daily activities. The disease is characterized by the accumulation of two proteins in fibrillar form; Amyloid-β forms fibrils that accumulate as extracellular plaques while tau fibrils form intracellular tangles. Here we aim to translate findings from a commonly used AD mouse model to AD patients. Here we initiate and chronically inhibit neuropathology in lateral entorhinal cortex (LEC) layer two neurons in an AD mouse model. This is achieved by over-expressing P301L tau virally and chronically activating hM4Di DREADDs intracranially using the ligand dechloroclozapine. Biomarkers in cerebrospinal fluid (CSF) is measured longitudinally in the model using microdialysis, and we use this same system to intracranially administer drugs aimed at halting AD-related neuropathology. The models are additionally tested in a novel contextual memory task. Preliminary findings indicate that viral injections of P301L tau into LEC layer two reveal direct projections between this region and the outer molecular layer of dentate gyrus and the rest of hippocampus. Additionally, phosphorylated tau co-localize with ‘starter cells’ and appear to spread from the injection site. Preliminary microdialysis results suggest that the concentrations of CSF amyloid-β and tau proteins mirror changes observed along the disease cascade in patients. The disease-modifying drugs appear to halt neuropathological development in this preclincial model. These findings will lead to a novel platform for translational AD research, linking the extensive research done in rodents to clinical applications. Siddharth’s abstract: A distributed brain network supports our ability to remember past events. The parietal cortex is a critical member of this network, yet, its exact contributions to episodic remembering remain unclear. Neurodegenerative syndromes affecting the posterior neocortex offer a unique opportunity to understand the importance and role of parietal regions to episodic memory. In this talk, I introduce and explore the rare neurodegenerative syndrome of Logopenic Progressive Aphasia (LPA), an aphasic variant of Alzheimer’s disease presenting with early, left-lateralized temporo-parietal atrophy, amidst relatively spared hippocampal integrity. I then discuss two key studies from my recent Ph.D. work showcasing pervasive episodic and autobiographical memory dysfunction in LPA, to a level comparable to typical, amnesic Alzheimer’s disease. Using multimodal neuroimaging, I demonstrate how degeneration of the angular gyrus in the left inferior parietal lobule, and its structural connections to the hippocampus, contribute to amnesic profiles in this syndrome. I finally evaluate these findings in the context of memory profiles in other posterior cortical neurodegenerative syndromes as well as recent theoretical models underscoring the importance of the parietal cortex in the integration and representation of episodic contextual information.

I will present work on a novel fMRI analysis method that allows us to investigate sequential reactivation in the hippocampus. Our method focuses on analysing the time courses of probabilistic multivariate classifiers and allows us to infer the presence and frequency of fast sequential reactivation events. Using a paradigm in which we controlled the speed of sequential visually elicited activations, we validated the method in visual cortex for event sequences with only 32 ms between items. We show that detectability remains possible if low signal-to-noise ratio and when sequence events occur at unknown times. In a preliminary analysis, we show that even the exposure to our visual paradigm elicits reactivations in visual cortex at rest following the task. I then present work in which we tested how representations influence replay by asking whether transitions between task-state representations are reactivated at rest during hippocampal replay events. Participants learned to make decisions about ambiguous stimuli that depended on past events and attentionally filtered stimulus processing. FMRI signals during rest periods following this task indicated sequential reactivation of task states. These results indicate that adaptive task state representations are computed and replayed at different cortical sites. In combination with other methods, fMRI may allow us to unravel this coordinated nature of replay.

Plasticity shapes the brain during development, and mechanisms of plasticity continue into adulthood to enable learning and memory. Nearly all brain functions are influenced by past events, reinforcing the view that the confluence of plasticity and computation in the same circuit elements is a core component of biological intelligence. My laboratory studies plasticity in the cerebral cortex during development, and plasticity during behaviour that is manifest as cortical dynamics. I will describe how cortical plasticity is implemented by learning rules that involve not only Hebbian changes and synaptic scaling but also dendritic renormalization. By using advanced techniques such as optical measurements of single-synapse function and structure in identified neurons in awake behaving mice, we have recently demonstrated locally coordinated plasticity in dendrites whereby specific synapses are strengthened and adjacent synapses with complementary features are weakened. Together, these changes cooperatively implement functional plasticity in neurons. Such plasticity relies on the dynamics of activity-dependent molecules within and between synapses. Alongside, it is increasingly clear that risk genes associated with neurodevelopmental disorders disproportionately target molecules of plasticity. Deficits in renormalization contribute fundamentally to dysfunctional neuronal circuits and computations, and may be a unifying mechanistic feature of these disorders.