Biology subjects

Sussillo, D.

Publications and source records attributed to Sussillo, D..

7 recordsLinked to original sources

bioRxiv2025

Improved interpretability in LFADS models using a learned, context-dependent per-trial bias

The computation-through-dynamics perspective argues that biological neural circuits process information via the continuous evolution of their internal states. Inspired by this perspective, Latent Factor Activity using Dynamical systems (LFADS, [1]) identifies a generative model consistent with the neural activity recordings. LFADS models neural dynamics with a recurrent neural network (RNN) generator, which results in excellent fit to the data. However, it has been difficult to understand the dynamics of the LFADS generator. In this work, we show that this poor interpretability arises in part because the generator implements complex, multi-stable dynamics. We introduce a simple modification to LFADS that ameliorates issues with interpretability by providing an inferred per-trial bias (modeled as a constant input) to the RNN generator, enabling it to contextually adapt a simpler dynamical system to individual trials. In both simulated neural recordings from pendulum oscillations and real recordings during arm movements in nonhuman primates, we observed that the standard LFADS learned complex, multi-stable dynamics, whereas the modified LFADS learned easier-to-understand contextual dynamics. This enabled direct analysis of the generator, which reproduced at a single-trial level previous results shown only through more complex analyses at the trial average. Finally, we applied the per-trial inferred bias LFADS model to human intracortical brain computer interface recordings during attempted finger movements and speech. We show that modifying neural dynamics using linear operations of the per-trial bias addresses non-stationarity and identifies the extent of behavioral variability, problems known to plague BCI. We call our modification to LFADS as "contextual LFADS".

neuroscience↗

bioRxiv2025

Computation-through-Dynamics Benchmark: Simulated datasets and quality metrics for dynamical models of neural activity

A primary goal of systems neuroscience is to discover how ensembles of neurons transform inputs into goal-directed behavior, a process known as neural computation. A powerful framework for understanding neural computation uses neural dynamics - the rules that govern how neural activity evolves over time - to explain how goal-directed input-output transformations occur. As dynamical rules are not directly observable, we need computational models that can infer neural dynamics from recorded neural activity. We typically validate such models using synthetic datasets with known ground-truth dynamics, but unfortunately existing synthetic datasets dont reflect fundamental features of neural computation and may therefore be poor proxies for neural systems. Further, the field lacks validated metrics for quantifying the accuracy of the dynamics inferred by models. The Computation-through-Dynamics Toolkit (CtDToolkit) addresses these critical gaps by providing: 1) synthetic datasets that reflect computational properties of biological neural circuits, 2) interpretable metrics for quantifying model performance, and 3) a standardized pipeline for training and evaluating models with or without known external inputs. In this manuscript, we demonstrate how CtDToolkit can help guide the development, tuning, and troubleshooting of neural dynamics models. In summary, CtDToolkit provides a necessary framework for model developers to better understand and characterize neural computation through the lens of dynamics. Author SummaryUnderstanding how the brain works requires interpretable accounts of how populations of neurons process information to produce behavior. One powerful approach is to study "neural dynamics", the patterns of how neural activity evolves over time. Scientists develop computational models to infer these dynamics from neural recordings, but it has been challenging to know when the inferred dynamics are trustworthy. Existing datasets often lack key features of biological neural circuits, and current performance metrics can provide an incomplete picture of model quality. We developed the Computation-through-Dynamics Toolkit (CtDToolkit) to solve these problems. Our toolkit provides three key resources: biologically motivated synthetic datasets, improved metrics that provide more holistic accounts of model performance, and a standardized workflow for training and evaluating models. We hope that CtDToolkit enables researchers to rigorously test, improve, and troubleshoot their models before applying them to real brain data. This work establishes a crucial foundation for developing better methods to understand neural computation, ultimately advancing our ability to decode how the brain transforms sensory information into thought and action.

neuroscience↗

bioRxiv2024

A generic noninvasive neuromotor interface for human-computer interaction

Since the advent of computing, humans have sought computer input technologies that are expressive, intuitive, and universal. While diverse modalities have been developed, including keyboards, mice, and touchscreens, they require interaction with an intermediary device that can be limiting, especially in mobile scenarios. Gesture-based systems utilize cameras or inertial sensors to avoid an intermediary device, but they tend to perform well only for unobscured or overt movements. Brain computer interfaces (BCIs) have been imagined for decades to solve the interface problem by allowing for input to computers via thought alone. However high-bandwidth communication has only been demonstrated using invasive BCIs with decoders designed for single individuals, and so cannot scale to the general public. In contrast, neuromotor signals found at the muscle offer access to subtle gestures and force information. Here we describe the development of a noninvasive neuromotor interface that allows for computer input using surface electromyography (sEMG). We developed a highly-sensitive and robust hardware platform that is easily donned/doffed to sense myoelectric activity at the wrist and transform intentional neuromotor commands into computer input. We paired this device with an infrastructure optimized to collect training data from thousands of consenting participants, which allowed us to develop generic sEMG neural network decoding models that work across many people without the need for per-person calibration. Test users not included in the training set demonstrate closed-loop median performance of gesture decoding at 0.5 target acquisitions per second in a continuous navigation task, 0.9 gesture detections per second in a discrete gesture task, and handwriting at 17.0 adjusted words per minute. We demonstrate that input bandwidth can be further improved up to 30% by personalizing sEMG decoding models to the individual, anticipating a future in which humans and machines co-adapt to provide seamless translation of human intent. To our knowledge this is the first high-bandwidth neuromotor interface that directly leverages biosignals with performant out-of-the-box generalization across people.

neuroscience↗

bioRxiv2022

A new theoretical framework jointly explains behavioral and neural variability across subjects performing flexible decision-making

The ability to flexibly switch our response to external stimuli according to contextual information is critical for successful interactions with a complex world. Context-dependent computations are necessary across many domains1-3, yet their neural implementations remain poorly understood. Here we developed a novel behavioral task in rats to study context-dependent selection and accumulation of evidence for decision-making4-6. Under assumptions supported by both monkey and rat data, we first show mathematically that a network can solve this problem through a combination of three defined components. These components can be identified and tested directly with experimental data. We further show that existing electrophysiological and modeling data are compatible with the full variety of possible combinations of these components, suggesting that different individuals could use different component combinations. To study variability across individual subjects, we developed automated, high-throughput methods to train rats on our task, and we trained many subjects on it. Consistent with theoretical predictions, neural and behavioral analyses revealed substantial heterogeneity across rats, despite uniformly good task performance. Our theory further predicts a specific link between behavioral and neural signatures, which was robustly supported in the data. In summary, our results provide a new experimentally-supported theoretical framework to analyze individual variability in biological and artificial systems performing flexible decision-making tasks, they open the door to cellular-resolution studies of individual variability in higher cognition, and they provide insights into neural mechanisms of context-dependent computation more generally.

neuroscience↗

bioRxiv2022

Flexible multitask computation in recurrent networks utilizes shared dynamical motifs

Flexible computation is a hallmark of intelligent behavior. Yet, little is known about how neural networks contextually reconfigure for different computations. Humans are able to perform a new task without extensive training, presumably through the composition of elementary processes that were previously learned. Cognitive scientists have long hypothesized the possibility of a compositional neural code, where complex neural computations are made up of constituent components; however, the neural substrate underlying this structure remains elusive in biological and artificial neural networks. Here we identified an algorithmic neural substrate for compositional computation through the study of multitasking artificial recurrent neural networks. Dynamical systems analyses of networks revealed learned computational strategies that mirrored the modular subtask structure of the task-set used for training. Dynamical motifs such as attractors, decision boundaries and rotations were reused across different task computations. For example, tasks that required memory of a continuous circular variable repurposed the same ring attractor. We show that dynamical motifs are implemented by clusters of units and are reused across different contexts, allowing for flexibility and generalization of previously learned computation. Lesioning these clusters resulted in modular effects on network performance: a lesion that destroyed one dynamical motif only minimally perturbed the structure of other dynamical motifs. Finally, modular dynamical motifs could be reconfigured for fast transfer learning. After slow initial learning of dynamical motifs, a subsequent faster stage of learning reconfigured motifs to perform novel tasks. This work contributes to a more fundamental understanding of compositional computation underlying flexible general intelligence in neural systems. We present a conceptual framework that establishes dynamical motifs as a fundamental unit of computation, intermediate between the neuron and the network. As more whole brain imaging studies record neural activity from multiple specialized systems simultaneously, the framework of dynamical motifs will guide questions about specialization and generalization across brain regions.

neuroscience↗

bioRxiv2021

Goal-Driven Recurrent Neural Network Models of the Ventral Visual Stream

The computational role of the abundant feedback connections in the ventral visual stream (VVS) is unclear, enabling humans and non-human primates to effortlessly recognize objects across a multitude of viewing conditions. Prior studies have augmented feedforward convolutional neural networks (CNNs) with recurrent connections to study their role in visual processing; however, often these recurrent networks are optimized directly on neural data or the comparative metrics used are undefined for standard feedforward networks that lack these connections. In this work, we develop task-optimized convolutional recurrent (ConvRNN) network models that more correctly mimic the timing and gross neuroanatomy of the ventral pathway. Properly chosen intermediate-depth ConvRNN circuit architectures, which incorporate mechanisms of feedforward bypassing and recurrent gating, can achieve high performance on a core recognition task, comparable to that of much deeper feedforward networks. We then develop methods that allow us to compare both CNNs and ConvRNNs to fine-grained measurements of primate categorization behavior and neural response trajectories across thousands of stimuli. We find that high performing ConvRNNs provide a better match to this data than feedforward networks of any depth, predicting the precise timings at which each stimulus is behaviorally decoded from neural activation patterns. Moreover, these ConvRNN circuits consistently produce quantitatively accurate predictions of neural dynamics from V4 and IT across the entire stimulus presentation. In fact, we find that the highest performing ConvRNNs, which best match neural and behavioral data, also achieve a strong Pareto-tradeoff between task performance and overall network size. Taken together, our results suggest the functional purpose of recurrence in the ventral pathway is to fit a high performing network in cortex, attaining computational power through temporal rather than spatial complexity.

neuroscience↗

bioRxiv2017

Inferring single-trial neural population dynamics using sequential auto-encoders

Neuroscience is experiencing a data revolution in which simultaneous recording of many hundreds or thousands of neurons is revealing structure in population activity that is not apparent from single-neuron responses. This structure is typically extracted from trial-averaged data. Single-trial analyses are challenging due to incomplete sampling of the neural population, trial-to-trial variability, and fluctuations in action potential timing. Here we introduce Latent Factor Analysis via Dynamical Systems (LFADS), a deep learning method to infer latent dynamics from single-trial neural spiking data. LFADS uses a nonlinear dynamical system (a recurrent neural network) to infer the dynamics underlying observed population activity and to extract de-noised single-trial firing rates from neural spiking data. We apply LFADS to a variety of monkey and human motor cortical datasets, demonstrating its ability to predict observed behavioral variables with unprecedented accuracy, extract precise estimates of neural dynamics on single trials, infer perturbations to those dynamics that correlate with behavioral choices, and combine data from non-overlapping recording sessions (spanning months) to improve inference of underlying dynamics. In summary, LFADS leverages all observations of a neural populations activity to accurately model its dynamics on single trials, opening the door to a detailed understanding of the role of dynamics in performing computation and ultimately driving behavior.

neuroscience↗