Year in Review

From Wu Tsai Faculty Member Rajit Manohar’s research group, published in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems:

Software algorithms can be built into the hardware of computer chips to dramatically improve performance. This work creates a rapid prototyping framework to simplify this process, by translating message-passing algorithms into a description language that can be mapped to massively parallel field-programmable gate arrays (FPGAs). This framework can support a wide range of circuits and is now being used to accelerate the training of models that are inspired by computational principles of the visual cortex in mammals.

From a collaboration between the research groups of Wu Tsai Faculty Members Smita Krishnaswamy and Nick Turk-Browne, published in Nature Computational Science:

The human brain is incredibly complex but its activity follows predictable patterns. This work develops a new algorithm (T-PHATE) optimized for human brain imaging that captures these dynamic patterns on a non-linear manifold of potential brain states. When applied to functional magnetic resonance imaging (fMRI) data collected during cognitive tasks, T-PHATE can extract and visualize trajectories through these brain states, revealing how the mind is thinking, comprehending, and remembering over time.

From Wu Tsai Faculty Member Priya Panda’s research group, published in ACM Computing Surveys:

Deep learning underlies many modern technological innovations such as self-driving cars and voice assistants, yet training and running these AI models consumes a tremendous amount of energy. This work reviews the field of neuromorphic computing, which develops algorithms and systems that draw inspiration from the brain. Advances such as spiking neural networks and non-volatile memory-based computing could drastically reduce the energy use, compute latency, and learning complexity in deep learning of the future.

From a collaboration between the research groups of Wu Tsai Faculty Members Thierry Emonet and Damon Clark, published in Nature:

Animals can navigate through the world using different senses. This study discovered how one species, the fruit fly, navigates based on smell. This requires identifying the source of odors, which flies achieve by sensing the direction that the odor moves across their antennae corrected for the direction of the wind. This reliance on odor motion is more efficient than other olfactory cues and mimics strategies used to detect visual motion through the eye.

From Wu Tsai Faculty Member Charlie Greer’s research group, published in eNeuro:

Immune cells are ubiquitous in the brain but much less is known about their function than neurons. This study discovered that microglia, a major class of brain immune cells, maintain corridors in the developing brain that facilitate the movement of new neurons to their target destination. These findings have important implications for understanding normal and disordered development of neural circuits and connectivity.

From Wu Tsai Faculty Member Flora Vaccarino’s research group, published in Molecular Psychiatry:

Developmental disorders such as Tourette syndrome provide an important window into the link between genetics, biology, and behavior. Using stem cells to grow organoids — person-specific samples of tissue that mirror properties of the brain — this study discovered that Tourette’s is associated with decreased production of inhibitory interneurons in certain brain regions. This early mispatterning alters the inhibitory tone of networks connecting the cortex and basal ganglia, which may help explain the disabling behavioral tics that occur in this disorder.

From Wu Tsai Faculty Member Joy Hirsch’s research group, published in IEEE Photonics Conference:

Reading faces and making eye contact are essential for effective social interaction, and yet these processes are poorly understood because there are limited options for brain imaging during live, two-person interactions. This study uses functional near-infrared spectroscopy to record brain activity simultaneously in two individuals in a face-to-face task exploring emotions. Compared with typically developed individuals, those with autism spectrum disorder showed reduced activity in areas of the brain involved in social skills.

From Wu Tsai Faculty Member Steve Chang’s research group, published in Neuron:

The eyes of another individual provide a fundamental source of information for social behavior. This study investigated how this information about interactive social gaze is processed in the primate brain. Gazing at other individuals resulted in widespread activity, not limited or timed to certain regions, suggesting a model of social interaction that depends on distributed computations across large brain networks.

From Wu Tsai Faculty Member Julian Jara-Ettinger’s research group, published in Cognition:

Learning from others is an efficient way to glean information without having to experience everything ourselves, but this is only useful if our sources are reliable. This study reveals how young children make inferences about the knowledge and trustworthiness of others. The results show that children use sophisticated mental models to make these inferences and determine who is worth sharing knowledge with.

The brain is organized by a set of principles whose disruption can be a sign — or even a cause — of cognitive decline. To better understand the healthy brain and what happens when disease affects the brain, we seek to illuminate these organizing principles. We are working on a novel multimodal neuroimaging approach that integrates positron emission tomography, optical imaging, and functional magnetic resonance imaging. By applying this framework in a mouse model of Alzheimer’s Disease, we are working to uncover early changes in brain activity patterns that are indicative of disease and the neurobiological changes that drive them. In this degenerative model, we have observed early hyperactivity followed by a loss of synapses and neural connectivity between specific brain regions. In the coming year, we aim to determine the relative roles of excitatory/inhibitory synapse balance versus overall synapse loss to these changes in brain activity.

Detecting motion in the environment as we move or objects move toward us is a critical task of our perceptual systems. Compared with motion detection in the visual system, relatively little is known about the computation of motion signals in the auditory and motor systems. Our goal is to compare neural algorithms for motion detection across perceptual systems. This year we have been investigating how the human brain perceives rising or falling pitches. We have shown how humans combine loudness cues at different frequencies and times. Strikingly, we find commonalities in the algorithms used to detect auditory pitch changes and those used to detect visual motion, suggesting deep parallels in neural computations across senses. We are analyzing speech databases to reveal how these algorithms relate to human speech patterns and will soon begin work comparing visual and auditory to touch systems.

The biennial Real-Time Functional Imaging and Neurofeedback (rtFIN) conference was hosted at Yale with the Institute’s support. Co-organizers Michelle Hampson and Nick Turk-Browne welcomed 162 faculty and trainees from 14 countries. The conference included sessions focused on cognitive training, ethics and diversity, novel imaging methods, data sharing initiatives, and clinical and basic science applications.

The monthly Computational Neuroscience Chalk Talks, run by Damon Clark, welcomed faculty to give and receive feedback on theoretical and computational approaches to questions in neuroscience. Topics included deriving rules for growth of dendritic arbors, developing biophysical models to explain synaptic adaptation, and applying principles of information bottlenecks to explain visual processing.

Students and faculty interested in philosophy and neuroscience gathered at weekly journal clubs and regular external speaker events through Applied Philosophy in Neuroscience (APHINE), led by graduate students Alec Sheffield, Clayton Barnes, and Natalia Castelo Branco. Topics included consciousness, neurolinguistics, evolution of the mind, and neuroaesthetics.

We joined tens of thousands of neuroscientists at the annual conference of the Society for Neuroscience (SfN), hosted in San Diego, California. It was the Institute’s first conference exhibit, and the booth sparked interest, questions, and conversations from numerous researchers across career stages and around the world.

The Institute’s newly launched website was nominated for two Webby Awards in the categories of Best Website and Mobile Sites for a School/University and Best Mobile Visual Design Aesthetic (Mobile Features).

We supported Black in Neuro in their work to diversify the field of neuroscience by celebrating and empowering Black scholars and professionals. We especially recognize the efforts of Black in Neuro co-founder and Wu Tsai Postdoctoral Fellow De-Shaine Murray.