Biology subjects

Basrai, D.

Publications and source records attributed to Basrai, D..

4 recordsLinked to original sources

bioRxiv2025

Constraining inference of across-region interactions using neural activity perturbations

Functional interactions between brain regions are often inferred from multi-region models fit to neural activity recorded in behaving animals. Here, we show that inference of acrossregion interactions is hindered by the wide breadth of model fits consistent with naturally occurring neural activity. In contrast, models fit to activity that includes region-wide activity perturbations provide well-constrained estimates of across-region interactions; simulations suggest these estimates can be accurate.

neuroscience↗

bioRxiv2025

Interactions between motor cortical forelimb regions and their influence on muscles reorganize across behaviors

It remains unclear how classical models of motor cortical hierarchy align with emerging evidence of behavioral organization in motor cortex. To address this, we combined optogenetic inactivation, Neuropixels recording, and electromyography to quantify the pattern and influence of activity in the mouse analogs of forelimb premotor and primary motor cortex (RFA and CFA) during reaching and climbing. Results revealed that RFAs dominant influence on forelimb muscles and on CFA during reaching is replaced by a dominant influence of CFA on muscles and on RFA during climbing, even when forelimb muscle activity during climbing resembles that during reaching. Short-latency influence between regions on excitatory and inhibitory populations in different cortical laminae also showed behavioral specificity. Simultaneous recordings in both areas during climbing revealed a loss of activity timing differences seen during reaching previously interpreted as reflective of hierarchy. These findings demonstrate that hierarchical interactions between forelimb motor cortical regions are behavior-specific.

neuroscience↗

bioRxiv2023

Hierarchy in influence but not firing patterns among forelimb motor cortices

Though hierarchy is commonly invoked in descriptions of motor cortical function, its presence and manifestation in firing patterns remain poorly resolved. Here we use optogenetic inactivation to demonstrate that short-latency influence between forelimb premotor and primary motor cortices is asymmetric during reaching in mice, demonstrating a partial hierarchy between the endogenous activity in each region. Multi-region recordings revealed that some activity is captured by similar but delayed patterns where either regions activity leads, with premotor activity leading more. Yet firing in each region is dominated by patterns shared between regions and is equally predictive of firing in the other region at the single-neuron level. In dual-region network models fit to data, regions differed in their dependence on across-region input, rather than the amount of such input they received. Our results indicate that motor cortical hierarchy, while present, may not be exposed when inferring interactions between populations from firing patterns alone.

neuroscience↗

bioRxiv2021

A large-scale neural network training framework for generalized estimation of single-trial population dynamics

Recent technical advances have enabled recording of increasingly large populations of neural activity, even during natural, unstructured behavior. Deep sequential autoencoders are the current state-of-the-art for uncovering dynamics from these datasets. However, these highly complex models include many non-trainable hyperparameters (HPs) that are typically hand tuned with reference to supervisory information (e.g., behavioral data). This process is cumbersome and time consuming and biases model selection toward models with good representations of individual supervisory variables. Additionally, it cannot be applied to cognitive areas or unstructured tasks for which supervisory information is unavailable. Here we demonstrate AutoLFADS, an automated model-tuning framework that can characterize dynamics using only neural data, without the need for supervisory information. This enables inference of dynamics out-of-the-box in diverse brain areas and behaviors, which we demonstrate on several datasets: motor cortex during free-paced reaching, somatosensory cortex during reaching with perturbations, and dorsomedial frontal cortex during cognitive timing tasks. We also provide a cloud software package and comprehensive tutorials that enable new users to apply the method without dedicated computing resources.

neuroscience↗