Past events

The question of how to measure a smell has troubled scientists for over a century. It was none other than Alexander Graham Bell that raised the challenge: "we have very many different kinds of smells, all the way from the odor of violets and roses up to asafoetida. But until you can measure their likenesses and differences you can have no science of odor”. Such a measure of smell can be naturally derived from a model of olfactory perceptual quality space, and several such models have recently been put forth. These typically rely on finding mathematical rules that link odorant structure to aspects of odor perception. Here, I collected 49,788 perceptual odor estimates from 199 participants, and built such a model, finalizing a physicochemical measure of smell. This measure, expressed in radians, predicts real-world odorant pairwise perceptual similarity from odorant structure alone. Using this measure, I met Bell's challenge by accurately predicting the perceptual similarity of rose, violet and asafoetida, from their physicochemical structure. Next, based on thousands of comparisons, I identified a cutoff in this measure, below 0.05 radians, where discrimination between pairs of mixtures becomes highly challenging. To assess the usefulness of this measure, I investigated whether it can be used to create olfactory metamers, namely non-overlapping molecular compositions that share a common percept. Characterizing the link between physical structure and ensuing perception in vision and audition, and the creation of perceptual entities such as metamers, was important towards understanding their underlying dimensionality, brain mechanisms, and towards their ultimate digitization. I suggest that olfactory metamers can similarly aid these goals in olfaction. Zoom link to join: https://weizmann.zoom.us/j/93360836031?pwd=dDZEdTQ1QUkxUVVONVErVm9CcUJWQT09 Meeting ID: 933 6083 6031 Password: 591230

The hippocampus and related medial temporal lobe structures (entorhinal cortex, perirhinal cortex, etc.) play a vital role in transforming experience into long-term memories that are then stored in the cortex, however the cellular mechanisms which designate single neurons to be part of a memory trace remain unknown. Part of the difficulty in addressing the mechanisms of transformation of short-term to long-term memories is the distributed nature of the resulting “engram” at synapses throughout the cortex. We therefore used a behavioral paradigm dependent on both the hippocampus and neocortex that enabled us to generate memory traces rapidly and reliably in a specific cortical location, by training rodents to associate the direct electrical microstimulation of the primary sensory neocortex with a reward. We found that medial-temporal input to neocortical Layer 1 (L1) gated the emergence of specific firing responses in subpopulations of Layer 5 pyramidal neurons marked by increased burstiness related to apical dendritic activity. Following learning and during memory retrieval, these neocortical responses became independent of the medial-temporal influence but continued to evoke behaviour with single bursts sufficient to elicit a correct response. These findings suggest that L1 is the locus for hippocampal-dependent associative learning in the neocortex, where memory engrams are established in subsets of pyramidal neurons by enhancing the sensitivity of tuft dendrites to contextual inputs and driving burst firing. Zoom link to join- https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

The ability to use information to adaptively guide behavior is central to intelligence. A vital research challenge is to establish how people decide what they want to know. In this talk I will present our recent research characterizing three key motives of information seeking. We find that participants automatically assess (i) how useful information is in directing action, (ii) how it will make them feel, and (iii) how it will influence their ability to predict and understand the world around them. They then integrate these assessments into a calculation of the value of information that guides information-seeking or its avoidance. These diverse influences are captured by separate brain regions along the dopamine reward pathway and are differentially modulated by pharmacological manipulation of dopamine function. The findings yield predictions about how information-seeking behavior will alter in disorders in which the reward system malfunctions. We test these predictions using a linguistic analysis of participants’ web searches ‘in the wild’ to quantify their motives for seeking information and relate those to reported psychiatric symptoms. Finally, using controlled behavioral experiments we show that the three motives for seeking information follow different developmental trajectories that are consistent with what would be predicted from our neuroimaging data. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

A fundamental feature of sleep is that a sensory stimulus does not reliably affect behavior or subjective experience. What mediates such “sensory disconnection”? Do similar processes occur during anesthesia, cognitive lapses, and some neuropsychiatric disorders? In a series of studies in humans and rodents, we compared neuronal responses to identical auditory stimuli across wakefulness and sleep. In A1, early single-neuron spiking responses are largely comparable across wakefulness, natural sleep, and light anesthesia. However, robust differences emerge in downstream high-level regions and late-responding neurons, and in top-down response signatures, suggesting that sleep impairs effective cortical connectivity. We reconcile the apparent discrepancy with the classic “thalamic gating” notion by showing that in contrast to natural sleep, deep anesthesia does lead to attenuation already in A1. Next, we show that reduced locus coeruleus-noradrenaline (LC-NE) activity during sleep mediates sensory disconnection. We find that in freely behaving rats, LC-NE activity is a key mechanism that determines the likelihood of sensory-evoked awakenings (SEA): the level of ongoing tonic LC activity during sleep anticipates SEAs, while minimal optogenetic LC activation or silencing increases and decreases SEA, respectively. In humans, pharmacological manipulation of NE levels modulates sensory perception and late sensory responses, suggesting that NE links sensory awareness to external world events. We are exploring novel methods such as transcutaneous vagal nerve stimulation to modulate LC-NE non-invasively in humans. In the last part of the talk I will present recent results on sleep and memory consolidation. Using unilateral olfactory stimulation during sleep we find that ‘local’ targeted memory reactivation (TMR) in human sleep selectively promotes specific memories associated with regional sleep oscillations. In epilepsy patients implanted with depth electrodes we investigate the effects of intracranial electrical closed loop stimulation during sleep on memory and hippocampal-neocortical dialogue at single-neuron resolution. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

We all capture the world around us through digital data such as images, videos and sound. However, in many cases, we are interested in certain properties of the data that are either not available or difficult to perceive directly from the input signal. My goal is to “Re-render Reality”, i.e., develop algorithms that analyze digital signals and then create a new version of it that allows us to see and hear better. In this talk, I’ll present a variety of methodologies aimed at enhancing the way we perceive our world through modified, re-rendered output. These works combine ideas from signal processing, optimization, computer graphics, and machine learning, and address a wide range of applications. More specifically, I’ll demonstrate how we can automatically reveal subtle geometric imperfection in images, visualize human motion in 3D, and use visual signals to help us separate and mute interference sound in a video. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

Alzheimer’s disease (AD) is one of the most pressing global medical issues to date with no effective therapeutic strategies. Despite extensive research much remains unknown regarding the crosstalk between brain cells and the role of non-neuronal cells in the progression of Alzheimer’s disease (AD). We use single nucleus RNA-sequencing and machine learning algorithms to build detailed cellular maps of mice and human brain and to follow molecular changes in each cell type along disease progression. Our maps revealed new disease associated states in glia cells as well as unique multi-cellular communities linked to AD. Specifically, we found a link between populations of disease-associated astrocytes (DAAs), microglia, oligodendrocytes and GABAergic neurons to AD related traits in mouse models and in post-mortem human brains. Expanding the data analysis across multiple cell types, we found co-occurrences of cellular populations across individuals, which we define as multi-cellular communities. Among these communities we discovered a unique cellular community linked to cognitive decline and Alzheimer’s disease pathology. These new insights are shaping our understanding of the unique cellular environment of the Alzheimer’s disease brains. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

Memristive technologies are attractive candidates to replace conventional memory technologies and can also be used to combine data storage and computing to enable novel non-von Neumann computer architecture. One such non-von Neumann computer architecture is neuromorphic computing, where brain-inspired circuits are built for massive parallelism and in-place computing. This talk focuses on neuromorphic computing with memristors. I will show how we can get inspiration from the brain to build electronic circuits that are energy efficient and perform both inference and training extremely fast and efficient. We will see that this approach can be used not only to accelerate machine learning applications, but also for novel mixed-signal circuits and for near-sensor processing. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070

While various forms of cells have been found in relation to the hippocampus cognitive map and navigation system, how these cells are formed and what is read from them is still a mystery. In the current lecture I will talk about several projects which tackle these issues. First, I will show how the formation of border cells in the cognitive map is related to a coordinate transformation, second I will discuss the interaction between the reward system (VTA) and the hippocampus. Finally, I will describe a project using place cells as a proxy for associative memory for assessing deficits in Alzheimer's disease. Zoom link to join: https://weizmann.zoom.us/j/96608033618?pwd=SEdJUkR2ZzRBZ3laUUdGbWR1VFJTdz09 Meeting ID: 966 0803 3618 Password: 564068 Host: Dr. Rita Schmidt rita.schmidt@weizmann.ac.il tel: 9070