Faculty Bibliography

Though the temporal precision of neural computation has been studied intensively, a data-driven determination of this precision remains a fundamental challenge. Reproducible spike patterns may be obscured on single trials by uncontrolled temporal variability in behavior and cognition and may not be time locked to measurable signatures in behavior or local field potentials (LFP). To overcome these challenges, we describe a general-purpose time warping framework that reveals precise spike-time patterns in an unsupervised manner, even when these patterns are decoupled from behavior or are temporally stretched across single trials. We demonstrate this method across diverse systems: cued reaching in nonhuman primates, motor sequence production in rats, and olfaction in mice. This approach flexibly uncovers diverse dynamical firing patterns, including pulsatile responses to behavioral events, LFP-aligned oscillatory spiking, and even unanticipated patterns, such as 7 Hz oscillations in rat motor cortex that are not time locked to measured behaviors or LFP.

How does neural activity generate perception? The spatial identities and temporal latencies of activated units correlate with external sensory features, but finding the subspace of activity that is consequential for perception, remains challenging. We trained mice to recognize synthetic odors: optogenetically-driven spatiotemporal patterns of glomerular activity in the olfactory bulb. We then performed precise spatial or temporal perturbations on trained patterns and measured how recognition changes. Changes in recognition reflect the perceptual relevance of the modified feature. We modeled recognition as the matching of glomerular activity to learned templates, and uncovered what forms a perceptually-meaningful pattern template: activation sequences ordered by latencies relative to each other, with surprisingly minimal effect of sniff. Within templates, spatially-identified glomeruli contribute additively, with larger contributions from earlier-activated glomeruli. Template matching with these perceptually-meaningful features can account for animals' responses, with the degree of mismatch predicting changes in recognition. The model accurately generalizes to novel spatio-temporal manipulations of patterns, and produces non-linear responses that resemble the non-linear responses in the data. This is the first report to our knowledge, that not only establishes a causal role for neural activity sequences in perception, but also uncovers the perceptually-relevant coding schemes governing these sequences. Our synthetic approach reveals the fundamental logic of the olfactory code, and provides a general framework for testing links between sensory activity and perception

The piriform cortex is the first area of integration for all peripheral odor information and it is believed to generate a unique and wholistic representation of behavioral relevance, sensory object. However, what properties of the cortical neural population activity define odor objects remains unknown. To address this question, we recorded cortical spiking responses to synthetic odors made of fully parameterized optogenetic activity patterns in the olfactory bulb, enabling independent and precise control of the incoming neural responses unattainable with natural odorants. Then, we measured changes in the neural response to a range of controlled spatial and temporal perturbations of the pattern for which we previously established their behavioral relevance. We developed an experimental approach to systematically probe cortical neural activity and found features of the population code which represent behaviorally relevant information

Animals are capable of detecting odorants in a single sniff, at extremely low concentrations. This ability is crucial for survival, yet it is unknown how the olfactory system supports detection at the perceptual limit. In the mouse olfactory bulb, inhalation of different odors leads to changes in the set of neurons activated, as well as when neurons are activated relative to each other (synchrony), and the onset of inhalation (latency). A key question is which features of stimulus evoked activity (e.g. rate, synchrony, or latency) are used to guide detection behavior? Here, we probed the sensitivity of mice to perturbations across each stimulus dimension using holographic two-photon (2P) optogenetic stimulation of olfactory bulb neurons, with cellular and single action potential resolution and millisecond precision. We found that mice can detect single action potentials evoked synchronously across <20 olfactory bulb neurons. Mice exhibited this sensitivity for artificial ensembles of mitral cells, as well as mixed ensembles of mitral and granule cells. Further, we discovered that detection depends strongly on the synchrony of activation across neurons, with detectability falling to near-chance levels with an imposed stimulus spread 3 30 ms, while detection performance was minimally perturbed by changes in the latency of activation relative to inhalation. These results reveal that mice are acutely attuned to single neurons and action potentials in olfactory circuits, and that synchrony across neurons may be a critical feature supporting the perceptibility of sparse ensemble activity signals

Two-photon (2P) optogenetic stimulation is currently the only method for precise, fast, and non-invasive cellular excitation deep inside brain tissue; it is typically combined with holographic wavefront-shaping techniques to generate distributed light patterns and target them to multiple specific cells in the brain. During propagation in the brain, these light patterns undergo severe distortion, mainly due to scattering, which leads to a discrepancy between the desired and actual light distribution. However, despite its importance, measurement of these tissue-induced distortions and their effects on the light patterns has yet to be demonstrated in situ. To this end, holographic optogenetics confocally unraveled sculpting (HOCUS), a system for real-time in situ evaluation of holographic light patterns, based on confocally descanning the stimulation light's reflection from the brain, is developed. HOCUS measures both tissue and wave propagation properties and enables the real-time measurement and correction of the dimensions and positions of holographic spots relative to neurons targeted for stimulation. It can also be used to measure tissue attenuation length, and thus should facilitate future attempts to optimize the generated hologram to pre-compensate for tissue-induced distortions, thereby improving the reliability of 2P holographic stimulation experiments.

In the olfactory system, odor percepts retain their identity despite substantial variations in concentration, timing, and background. We study a novel strategy for encoding intensity-invariant stimulus identity that is based on representing relative rather than absolute values of stimulus features. For example, in what is known as the primacy coding model, odorant identities are represented by the conditions that some odorant receptors are activated more strongly than others. Because, in this scheme, odorant identity depends only on the relative amplitudes of olfactory receptor responses, identity is invariant to changes in both intensity and monotonic nonlinear transformations of its neuronal responses. Here we show that sparse vectors representing odorant mixtures can be recovered in a compressed sensing framework via elastic net loss minimization. In the primacy model, this minimization is performed under the constraint that some receptors respond to a given odorant more strongly than others. Using duality transformation, we show that this constrained optimization problem can be solved by a neural network whose Lyapunov function represents the dual Lagrangian and whose neural responses represent the Lagrange coefficients of primacy and other constraints. The connectivity in such a dual network resembles known features of connectivity in olfactory circuits. We thus propose that networks in the piriform cortex implement dual computations to compute odorant identity with the sparse activities of individual neurons representing Lagrange coefficients. More generally, we propose that sparse neuronal firing rates may represent Lagrange multipliers, which we call the dual brain hypothesis. We show such a formulation is well suited to solve problems with multiple interacting relative constraints.

Neural recordings performed at an increasing scale and resolution have revealed complex, spatio-temporally precise patterns of activity in the olfactory system. Multiple models may explain the functional consequences of the spatio-temporal olfactory code, but the link to behavior remains unclear. Recent evidence in the field suggests a behavioral sensitivity to both fine spatial and temporal features in the code. How these features and combinations of features give rise to olfactory behavior is the subject of active research in the field. Modern genetic and optogenetic methods show great promise in testing the link between olfactory codes and behavior.

In many species, survival depends on olfaction, yet the mechanisms that underlie olfactory sensitivity are not well understood. Here we examine how a conserved subset of olfactory receptors, the trace amine-associated receptors (TAARs), determine odor detection thresholds of mice to amines. We find that deleting all TAARs, or even single TAARs, results in significant odor detection deficits. This finding is not limited to TAARs, as the deletion of a canonical odorant receptor reduced behavioral sensitivity to its preferred ligand. Remarkably, behavioral threshold is set solely by the most sensitive receptor, with no contribution from other highly sensitive receptors. In addition, increasing the number of sensory neurons (and glomeruli) expressing a threshold-determining TAAR does not improve detection, indicating that sensitivity is not limited by the typical complement of sensory neurons. Our findings demonstrate that olfactory thresholds are set by the single highest affinity receptor and suggest that TAARs are evolutionarily conserved because they determine the sensitivity to a class of biologically relevant chemicals.

Olfactory inputs are organized in an array of functional units (glomeruli), each relaying information from sensory neurons expressing a given odorant receptor to a small population of output neurons, mitral/tufted (MT) cells. MT cells respond heterogeneously to odorants, and how the responses encode stimulus features is unknown. We recorded in awake mice responses from "sister" MT cells that receive input from a functionally characterized, genetically identified glomerulus, corresponding to a specific receptor (M72). Despite receiving similar inputs, sister MT cells exhibit temporally diverse, concentration-dependent, excitatory and inhibitory responses to most M72 ligands. In contrast, the strongest known ligand for M72 elicits temporally stereotyped, early excitatory responses in sister MT cells, consistent across a range of concentrations. Our data suggest that information about ligand affinity is encoded in the collective stereotypy or diversity of activity among sister MT cells within a glomerular functional unit in a concentration-tolerant manner.