2024 Research Grant Awards

Awarded to Research Professor Carlos Lois (BBE)
Are the observed genomic changes in individual neurons just random mistakes, or is there a pattern underlying physiological functions?
Over the last 15 years several works have reported different types of changes in the genomes of individual neurons. However, to date, there is no evidence that these genomic changes are relevant for any physiological function, and the general consensus is that they may be simply the result of random mistakes in DNA transactions. We hypothesize that we have not been able to unveil any meaningful pattern in these genomic changes because the techniques used until now lacked the necessary resolution. We propose to clone mature neurons from the adult mouse brain by nuclear transfer, and use new high-fidelity and long-read high-resolution sequencing technologies capable of detecting all types of genomic changes to explore whether there are discernible patterns behind these changes.

Awarded to Professor Scott Cushing (CCE) and Professor Elizabeth Hong (BBE)
Development of Entangled Photon Sources for Portable Single-Molecule Fluorescence Lifetime Imaging
Entangled photons enable compact, battery-powered, wavelength tunable excitation sources in a small physical footprint for low-cost, portable fluorescence lifetime microscopy and sensing. We propose to develop entangled photon sources optimized for fluorescence lifetime imaging (in terms of wavelength, form factor, and beam size) using a new microscope built jointly between the Cushing and Hong labs. Performance will be compared to classical laser sources in a realistic biological application — lifetime measurement of FRET-dependent fluorescence sensors of cyclic AMP and PKA activity in olfactory presynaptic terminals — to guide optimization of the entangled photon source for routine neuroscience applications.

Awarded to Research Professor Tsui-Fen Chou (BBE) and Professor Yuki Oka (BBE)
Discovering new circulating factors underlying homeostatic behavior
Detection of circulating hormones and blood osmolality changes by the brain triggers robust homeostatic behaviors such as feeding and drinking. A handful of these circulating factors have been characterized in the past decades. Our results suggest the existence of uncharacterized blood circulating factors that controls thirst and drinking responses in mice. This project seeks to combine cell-type-specific activity monitoring and mass spectrometry to identify a novel factor underlying homeostatic behavior.

Awarded to Professor Thanos Siapas (BBE)
Spike sorting algorithms for high-density neural probes
Chronic electrophysiological recordings with high density probes are a powerful tool for systems neuroscience and brain-machine interfaces, but spike-sorting remains a significant challenge. We previously developed a generative model for spike sorting tetrode data using a mixture of drifting t-distributions. This model captures two key features of chronic extracellular recordings: cluster drift over time and heavy tails in the distribution of spike features. We propose to extend this spike sorting approach to high-density probes, such as Neuropixels, to enable automated spike sorting over long periods of time and reliable quantitative measures of single-unit isolation. This approach will also enable real-time spike sorting for closed-loop experiments with brain-machine interfaces.

Awarded to Professor Paul Sternberg (BBE)
Transcriptional changes in male mating behavior plasticity
We will exploit a newly discovered experience-dependent behavior in C. elegans to address how a sensory-motor processing system adapts to different environments and alters an ethologically-relevant behavior, male copulatory behavior.The proposed project takes my long-term interest in neural circuits underlying innate behavior into the realm of plasticity. We will use a newly developed protocol for multiplexed single molecule FISH of C. elegans to identify cells and genes potentially involved in the plasticity. We can then test the function of cells (and ultimately genes) by now-standard methods of cell inactivation and activation to obtain a mechanistic model of the plasticity.

Awarded to Professor Ralph Adolphs (HSS), Professor Frederick Eberhardt (HSS) and Assistant Professor Matt Thomson (BBE)
Leveraging large language models to explore emotion
Transfomer-based models, such as GPT-4, exhibit interesting emergent properties and can be used flexibly across many application domains, yet also exhibit notable gaps when compared to biological systems. Here we focus on one such gap: emotion. Our goal is to articulate design principles and criteria for evaluating emotion in such models, so that the knowledge gleaned can help us better understand emotion in biological systems.

Awarded to Professor Azita Emami (EAS)
Software-Hardware Co-Design for Robust Feature Engineering in Brain Machine
Convolutional neural networks offer a significantly more robust and information rich approach for extracting neural features from broad band neural data. This project aims to optimize algorithmic parameters and hardware architecture to implement a low power feature extractor capable of processing the large bandwidth of neural channels required in today's brain-machine interfaces. We will also develop strategies for dynamic selection of channels with highest level of information content.

Awarded to Professor Mikhail Shapiro (EAS)
Developing Volumetric Ultrasound for Human Brain-Machine Interfaces
Brain machine interfaces (BMIs) help patients with disabilities achieve greater independence by using their thoughts to control assistive devices but commonly require highly invasive open brain surgery for electrode implantation since conventional non-invasive imaging techniques like fMRI and EEG lack sufficient spatiotemporal resolution for real-time BMI use. We recently demonstrated that functional ultrasound (fUS), a novel neuroimaging technique, can image from outside the dura and be used to decode behavior in non-human primates and through a polymeric acoustic skull window in a human patient, enabling a minimally invasive BMI. However, the human brain's size and complexity as well as difficulties with registering 2D fUS planes across sessions and to MRI make decoding 2D fUS signal challenging. We propose developing 3D fUS in humans to improve decoding capabilities through increased signal density and, in turn, using 3D fUS to decode real-time motor and speech information, offering a new minimally invasive alternative to standard BMIs.