Browsing by Author "Gong, Yiyang"

The brain encodes diverse cognitive functions through the coordinated activity of interacting neural circuits. Neural ensembles are groups of coactive neurons in these circuits that respond to similar stimuli. Neural ensembles are found throughout the brain and have been associated with many cognitive processes including memory, motor control, and perception. However, a key goal of systems neuroscience is to establish a functional link between neural activity and behavior and these previous studies established only a correlation between ensembles and behavior. Demonstrating a functional link between ensembles and behavior requires precise manipulation of ensemble activity. Manipulating ensemble activity allows neuroscientists to determine the patterns of neural activity that are necessary and sufficient to drive behavior. Additionally, recording and analyzing the activity of hundreds to thousands of neurons simultaneously allows neuroscientists to elucidate the patterns of neural activity underlying behavior. In this dissertation, we developed novel computational tools to help scientists selectively activate ensembles and analyze large-scale neural activity with single-cell resolution.One method to precisely activate cortical ensembles while limiting off-target effects is to stimulate pattern completion neurons. Pattern completion neurons are subsets of neurons in an ensemble that, when activated, can trigger the activation of the rest of the ensemble. However, scientists currently lack methods to reliably identify pattern completion neurons. The first project in this dissertation used computational modeling to identify characteristics of pattern completion neurons in cortical ensembles. We developed a realistic spiking model of layer 2/3 of the mouse visual cortex. We then identified ensembles in the network and quantified the pattern completion capability of different neuron pairs in an ensemble. We analyzed the relationship between structural and dynamic parameters and pattern completion capability. We found that multiple graph theory parameters, and degree in particular, could predict the pattern completion capability of a neuron pair. Additionally, we found that neurons that fired earlier in an ensemble recall event were more likely to have pattern completion properties than neurons that fired later. Lastly, we showed that we can measure this temporal latency in vivo with modern calcium indicators. The later projects in this dissertation used deep learning to improve calcium imaging analysis. First, we developed a semi-supervised pipeline for neuron segmentation to reduce the burden of manual labeling. We compensated for the low number of ground truth labels in two ways. First, we augmented the training data with pseudolabels generated with ensemble learning. Next, we used domain-specific knowledge to predict optimal hyperparameters from the limited ground truth labels. Our pipeline achieved state-of-the-art accuracy when trained on only 25% the number of manual labels as supervised methods. Lastly, we developed a spatiotemporal deep learning pipeline to predict the underlying electrical activity from calcium imaging videos. Calcium imaging provides only an indirect measurement of spiking neural activity, and various spike inference pipelines have attempted to accurately recover spiking timing and rate. Our pipeline improved the detection of single-spike events and improved spike rate prediction throughout the video. This improved performance will help scientists reconstruct neural circuits and study single-cell responses to stimuli. Overall, the tools developed in this dissertation will help systems neuroscience researchers establish a causal link between neural activity and behavior and will help determine the precise patterns of neural activity underlying these behaviors.

The brain supports numerous complex processes ranging from signal processing and motor control to learning and memory. These processes rely on signal transduction between interconnected networks of neurons that form neural circuits. Understanding how neural circuits function requires non-invasive, genetically specific technologies to both record and manipulate neural activity. Recording neural activity establishes a correlative relationship between the activity and cognitive function, while manipulating neural activity establishes a causal relationship between the activity and behavioral or physiological processes. Genetically encoded protein tools facilitate neuroscience research in both experimental paradigms. Genetically encoded sensors enable optical recording of neural activity across a wide spatiotemporal range. These indicators detect diverse forms of neural activity, including calcium ion flux, membrane voltage potential, and neurotransmitter concentration. Conversely, optogenetic actuators enable targeted, optical excitation or inhibition of neurons upon activation with a specific wavelength of light.

Advancement of genetically encoded tools will allow researchers to access new experimental regimes of neuroscience. Enhancing the fluorescence response and temporal fidelity of genetically encoded sensors improves signal detection fidelity, enabling neuroscientists to access more neurons at once and more precisely analyze neural circuits. Expanding the spectral diversity of genetically encoded tools makes it possible to record from multiple neural populations simultaneously or to optogenetically excite one population with a specific wavelength of light while recording the activity of another in a distinct optical channel. Such multi-channel experiments enable neuroscientists to investigate the influence of the activity of an ensemble of neurons on the activity of another ensemble downstream in a neural circuit or feedback between neural circuits. However, expanding the palette of protein sensors and actuators for such multi-channel experiments has been challenging. Most state-of-the-art genetically encoded sensors fuse cyan-light-sensitive green fluorescent protein to a sensing domain, so the dual channel experiments described above require a complementary sensor or actuator that is sensitive to a spectrally distinct wavelength of light. However, the performance of red fluorescent genetically encoded tools typically lags relative to their green counterparts, and using cyan-light-activated sensors in conjunction with green-light-activated actuators introduces high optical crosstalk. Additionally, the dynamic properties and context-dependent performance of genetically encoded sensors make high-throughput screens of this class of tools labor intensive and time consuming. This constraint on the throughput of screens has limited development efforts to a miniscule fraction of the possible variants of each sensor.

In this dissertation, I expanded the spectral diversity of the tools described above and developed a novel strategy for high throughput evolution of genetically encoded sensors. First, I developed a red fluorescent genetically encoded voltage sensor by engineering the fluorescence resonance energy transfer (FRET) efficiency between a voltage sensitive domain and a red fluorescent protein. This red fluorescent sensor enabled high fidelity recordings of neural activity with sub-millisecond temporal resolution, dual-channel recordings in parallel with green fluorescent sensors, and simultaneous optogenetic excitation and voltage imaging with minimal optical crosstalk. Second, I developed an optogenetic actuator with a blue-shifted activation spectrum by employing this same FRET mechanism. I demonstrated that the activation spectrum of optogenetic tools could be tuned by engineering FRET efficiency between the actuator domain and a compatible fluorescent protein. This straightforward strategy represents a technical step forward for engineering the spectra of optogenetic actuators, which has been difficult to achieve without compromising functionality. Third, I developed a screening method that enabled pooled, high-throughput screens of diverse libraries containing genetically encoded sensor mutants. This method employed both experimental and computational advancements. I used in situ optical mRNA sequencing to determine the sequence of each screened protein variant and machine learning to predict the function of unscreened variants. I expanded the coverage of the possible sequence space by over an order of magnitude compared to traditional directed evolution.