Topic: recurrent neural network

Cognitive maps confer animals with flexible intelligence by representing spatial, temporal, and abstract relationships that can be used to shape thought, planning, and behavior. Cognitive maps have been observed in the hippocampus, but their algorithmic form and the processes by which they are learned remain obscure. Here, we employed large-scale, longitudinal two-photon calcium imaging to record activity from thousands of neurons in the CA1 region of the hippocampus while mice learned to efficiently collect rewards from two subtly different versions of linear tracks in virtual reality. The r

Efficiently transmitting information from dynamic environments necessitates sensory systems that rapidly and reversibly adapt to changes in input statistics. Single neurons adjust their input-output gains to adaptively normalize their stimulus-driven response variance (Fairhall et al. 2001). In neural populations, statistical whitening and related adaptations have been observed (Dan et al. 1996; Benucci et al. 2013; Wanner and Friedrich 2020), whereby joint statistics of neurons are decorrelated in addition to being re-scaled to have equal variance. Existing models of neural population adapt

Animals from flying insects to running humans thrive in uncertain and noisy environments. They smoothly navigate, swiftly track targets and escape predators because their sensorimotor systems are precisely fine-tuned for these tasks. Millisecond-precise spike timing codes, prevalent in both sensory and motor systems, can prompt large changes in muscle forces especially when several muscles coordinate to control a biomechanical action. Such coordinated spike timing changes, modulated by task-relevant sensory encoding over long timescales, can significantly contribute to the effectiveness of mot

Motor adaptation is a widely-used paradigm for understanding short-term learning. However, it is unknown how existing skillsets acquired through the long-term learning process that begins in utero can affect motor adaptation. Long-term learning likely causes changes in neural connectivity, which may shape the neural dynamics that can be produced. To understand the interaction between circuit connectivity constraints and a neural population’s ability to change its activity patterns, we modeled the neural dynamics of the motor cortex during skill learning and subsequent adaptation using a recurr

Experimental and computational studies suggest that motor cortex acts as a feedback controller, allowing for ‘on-the-fly’ movement corrections in response to afferent sensory feedback. However, it remains unclear whether feedback control relates to longer-term learning, and how this would be implemented in neural circuitry. Here, we tackled these questions by testing how a recurrent neural network (RNN) can use feedback to control its own output, and whether this process can enable learning. We built an RNN that received feedback signaling the error between its intended and observed output. An

Being able to identify key patterns of neural network connectivity that are specifically required to generate functional response patterns is a challenging and aspirational goal in neuroscience. The difficulty, in part, is because we are typically only able to probe a network with a limited number of stimulus conditions, and there exists huge degeneracies in the ways we can connect the neurons to reproduce the observed responses. Here we present a new geometric ensemble modeling approach to this problem. Inspired by whole-brain imaging approaches, we assume that we have access to a finite numb

Spike trains recorded from the cortex of behaving animals can be complex, highly variable from trial to trial, and therefore challenging to interpret. A fraction of cells exhibit trial-averaged responses with obvious task-related features such as pure tone frequency tuning in auditory cortex. However, a substantial number of cells (including cells in primary sensory cortex) do not appear to fire in a task-related manner and are often neglected from analysis. We recently used a novel single-trial, spike-timing-based analysis to show that both classically responsive and non-classically responsiv

Motor control is a complex, high-dimensional task. It has been suggested that the brain may reduce the dimensionality of this control problem by using a set of motor primitives (or muscle synergies) that encode frequently used patterns of muscle co-activations in the time and/or weight-domain (Bizzi et al., 1991; Kutch and Valero-Cuevas, 2012; Bizzi and Cheung, 2013; Alessandro et al., 2013). In this project we investigate how motor primitives may affect the learning of movement in motor cortex. We model motor cortex as a recurrent neural network (RNN) and compare two potential output scenario

The upper limbs are essential for performing everyday tasks that require a wide range of motion and precise coordination. Planning and timing are crucial to achieve coordinated movement. Sensory information about the target and current body state is critical, as is the integration of prior experience represented by prelearned inverse dynamics that generate the associated muscle activity. We propose a generative model that uses a recurrent neural network to predict upper limb muscle activity during various simple and complex everyday movements. By identifying movement primitives within the si

Visual search, the act of finding a target among multiple visually presented items, is a key paradigm in studying visual attention. While much is known about the brain networks underlying visual search, our understanding of the neural computations driving this behavior is limited, resulting in challenges in simulating such behavior in-silico. To address this gap, we developed an image-computable general-architecture visual search model (GVSM). We trained an artificial neural network to perform naturalistic visual search [Fig. A-B] using the same stimulus configuration as human psychophysics ex

Biological circuits adapt their behaviour through experience and ongoing learning in a non-trivial manner. Experimental evidence shows differing neuronal population responses to familiar and novel stimuli [1, 2], indicating that neural circuits operate in distinct processing modes depending on stimulus familiarity. In layer 2/3 of mouse primary visual cortex (V1) specifically, populations of pyramidal and vasoactive-intestinal-peptide-expressing (VIP) cells show pronounced responses to novel visual stimuli (novelty responses) and weaker responses to familiar ones. Although experimental and the