Topic: hidden Markov models

This webinar addressed computational perspectives on how animals and humans make decisions, spanning normative, descriptive, and mechanistic models. Sam Gershman (Harvard) presented a capacity-limited reinforcement learning framework in which policies are compressed under an information bottleneck constraint. This approach predicts pervasive perseveration, stimulus‐independent “default” actions, and trade-offs between complexity and reward. Such policy compression reconciles observed action stochasticity and response time patterns with an optimal balance between learning capacity and performan

Calcium imaging (CI) has emerged as a prominent method for recording neural population activity. Therefore, CI recordings are prime candidates for the application of latent variable models designed to identify latent neural states and dynamics. However, in CI experiments spiking events are measured only indirectly via calcium ions that rapidly enter a cell during action potentials and then slowly seep back out. To analyze CI data, per-neuron fluorescence time-traces are typically either de-convolved to approximate spiking events or analyzed directly. The former approach, although enabling many

Understanding how internally driven cognitive processes (COPs) give rise to behavior in different species is one of the key challenges in systems neuroscience. Traditionally, neuroscience has aimed to isolate individual COPs by simplifying sensory stimuli and behavioral readouts. Yet in natural behavior, several COPs can co-occur, often with overlapping neural representations. The dynamics of such superimposed COPs are yet to be understood, as well as the extent to which they generalize across species. To tackle this question, we developed a naturalistic Virtual Reality setup in which two spec

The hippocampus and the medial prefrontal cortex (mPFC) participate in episodic and working memory, respectively. In trace conditioning, animals learn to associate a sensory stimulus (CS) and a reward (US) separated by a delay (trace period). Although the neural activity in both areas can be linked to memory representation formation, the neural mechanisms involved in learning the temporal association between CS and US are not known. In this work, we use auditory trace conditioning in head-fixed mice with simultaneous 128-channel silicon probe recordings from the hippocampal CA1 subfield and th

Uncovering how a stream of sensory inputs maps onto a sequence of actions is crucial for understanding the nervous system. However, behavior is shaped not only by recent sensory stimuli but also by longer-term experiences, internal neuronal dynamics, and physiological states. As a result, the same experimental conditions can result in variable behaviors. We study the role of internal states in zebrafish larvae behavior, leveraging a rich dataset of three million movements recorded during spontaneous behavior, photomotor responses, optomotor responses, and visually driven escape in the zebraf

Population recording offer a window into the underlying dynamics of brain regions. Latent Variable Models (LVMs) infer these dynamics treating the observations as projections from the latents. Evaluating the quality of these models remains challenging, as we don't have access to the true underlying dynamics. A common approach, "co-smoothing", involves jointly estimating latent variables from neural data and predicting the firing rates of held-out neurons. However, we find that optimizing this single objective often yields models that include extraneous dynamics not relevant to the underlying n

The patch-clamp technique allows us to eavesdrop the gating behavior of individual ion channels with unprecedented temporal resolution. The recorded signals arise from conformational changes of the protein or protein complex forming the ion channels as they make rapid transitions between conducting and non-conducting states. However, unambiguous analysis of single-channel datasets is challenging given the inadvertently low signal-to-noise ratio as well as signal distortions caused by low-pass filtering and limited recording bandwidth. Ion channel kinetics are typically described using hidden-M

This study investigates brain functional network dynamics of freely behaving but head-fixed mice using functional ultrasound (fUS). Unlike traditional fMRI, fUS allows recording brain activity in awake, behaving mice without anesthesia. We established a pipeline for processing fUS data and applied it to behavioral mouse data from a preclinical model of autism carrying a Shank2 gene knockout mutation. The pipeline includes several pre-processing stages, a deconvolution-based method for investigating neuronal activity underlying the fUS signal, and an analysis using hidden Markov models (HMM) to

When completing a task for a prolonged period, our ability to sustain attention fluctuates over time. In mice, errors and biased responses cluster in time rather than occurring randomly, indicating that animals transition into and out of disengaged states over the course of an experimental session. What neural and physiological processes trigger these state transitions? Here, we investigated the role of pupil-linked arousal, as relatively high or low levels of arousal are associated with poor task performance, and there is evidence that the activity of ascending arousal systems triggers larges