arXiv subjects

Cristiano Capone

Publications and source records attributed to Cristiano Capone.

At least 19 recordsLinked to original sources

arXiv2026

Optimal stimulation sites are not the most affected: personalised models of resting-state fMRI in Alzheimer's disease

Resting-state functional connectivity (FC) is altered in Alzheimer's disease (AD), widely regarded as a distributed network process; whether its signature reduces to a few focal sites has not been tested causally, a question central to targeted neuromodulation. We fit subject-specific, cross-subject-identifiable models whose free-running dynamics reproduce those of each individual patient. The fitted model parameters classify AD from controls at modest accuracy, below that of structural atrophy; we build on the functional model nonetheless, because dynamics, not tissue loss, are what stimulation can act on. Changing a virtual patient's model connectivity toward the control template reverts its AD classification, establishing in silico that the disease signature is correctable, yet the required correction is intrinsically distributed: a coordinated, multi-site change of the model connectivity that no single-node edit reproduces. Where, then, should a physically realisable focal drive act? A single-site drive at the node whose connectivity is most altered fails to revert the classification even at supra-physiological amplitudes, whereas selecting each patient's site by its effect on the disease discriminant achieves complete, individualised reclassification from one site, and a real-time closed-loop controller reaches comparable efficacy at lower dose using only causally available information. Optimal targets are cortical and heterogeneous: the site to stimulate is not where connectivity is most altered but where the network is most therapeutically responsive.

q-bio.NC↗

arXiv2026

Bilinear gating of motor primitives: a principle linking dendritic computation to rapid goal-directed adaptation

Movement requires the motor cortex to specify both \emph{what} action to produce and \emph{which goal} it serves, yet how individual neurons separate these factors is not understood. Here we show that in macaque motor cortex the \emph{burst fraction} of a neuron, the proportion of its spikes emitted in high-frequency bursts, encodes reach direction far more selectively than its overall firing rate. This dissociation is highly consistent: it holds in every one of 12 recording sessions spanning three animals and two laboratories (all $p<10^{-12}$) and survives controls that remove any contribution of firing rate, showing that goal information is concentrated specifically in bursts. We then show that this coding signature is the predicted consequence of dendritic coincidence detection in layer-5 pyramidal neurons: when a goal-related apical input coincides with a state-related basal drive the neuron bursts, so burst probability computes the product of goal and state, a bilinear gate $G(g)\,Y(s)$. A minimal two-compartment spiking model reproduces the effect, and the same multiplicative gate, embedded in a reinforcement-learning agent, supports zero-shot generalisation to new goals and rapid online adaptation, providing a computational rationale for segregating goal information into bursts. These results identify burst fraction as a goal-selective code in motor cortex, tie it to a concrete cellular mechanism, and show that the mechanism confers a learning advantage.

q-bio.NC↗

arXiv2026

Unified Policy Value Decomposition for Rapid Adaptation

Rapid adaptation in complex control systems remains a central challenge in reinforcement learning. We introduce a framework in which policy and value functions share a low-dimensional coefficient vector - a goal embedding - that captures task identity and enables immediate adaptation to novel tasks without retraining representations. During pretraining, we jointly learn structured value bases and compatible policy bases through a bilinear actor-critic decomposition. The critic factorizes as Q = sum_k G_k(g) y_k(s,a), where G_k(g) is a goal-conditioned coefficient vector and y_k(s,a) are learned value basis functions. This multiplicative gating - where a context signal scales a set of state-dependent bases - is reminiscent of gain modulation observed in Layer 5 pyramidal neurons, where top-down inputs modulate the gain of sensory-driven responses without altering their tuning. Building on Successor Features, we extend the decomposition to the actor, which composes a set of primitive policies weighted by the same coefficients G_k(g). At test time the bases are frozen and G_k(g) is estimated zero-shot via a single forward pass, enabling immediate adaptation to novel tasks without any gradient update. We train a Soft Actor-Critic agent on the MuJoCo Ant environment under a multi-directional locomotion objective, requiring the agent to walk in eight directions specified as continuous goal vectors. The bilinear structure allows each policy head to specialize to a subset of directions, while the shared coefficient layer generalizes across them, accommodating novel directions by interpolating in goal embedding space. Our results suggest that shared low-dimensional goal embeddings offer a general mechanism for rapid, structured adaptation in high-dimensional control, and highlight a potentially biologically plausible principle for efficient transfer in complex reinforcement learning systems.

cs.LG↗

arXiv2026

Training slow silicon neurons to control extremely fast robots with spiking reinforcement learning

Air hockey demands split-second decisions at high puck velocities, a challenge we address with a compact network of spiking neurons running on a mixed-signal analog/digital neuromorphic processor. By co-designing hardware and learning algorithms, we train the system to achieve successful puck interactions through reinforcement learning in a remarkably small number of trials. The network leverages fixed random connectivity to capture the task's temporal structure and adopts a local e-prop learning rule in the readout layer to exploit event-driven activity for fast and efficient learning. The result is real-time learning with a setup comprising a computer and the neuromorphic chip in-the-loop, enabling practical training of spiking neural networks for robotic autonomous systems. This work bridges neuroscience-inspired hardware with real-world robotic control, showing that brain-inspired approaches can tackle fast-paced interaction tasks while supporting always-on learning in intelligent machines.

cs.RO↗

arXiv2026

Lightweight Test-Time Adaptation for EMG-Based Gesture Recognition

Reliable long-term decoding of gestures from surface electromyography (EMG) is hindered by signal drift caused by electrode displacement, muscle fatigue, and/or posture changes. Although modern models achieve high intra-session accuracy, their performance often degrades substantially across recording sessions. Existing approaches to mitigate this problem typically rely on large training datasets or computationally intensive pipelines that are unsuitable for energy-efficient wearable devices. We propose a lightweight test-time adaptation framework for EMG decoding. The framework includes three complementary adaptation strategies: (i) causal adaptive batch normalization for online statistical alignment, (ii) Gaussian Mixture Model alignment with experience replay to mitigate forgetting, and (iii) meta-learning for rapid few-shot calibration. We evaluate these methods on the multi-session NinaPro DB6 dataset. All approaches substantially improve inter-session robustness relative to a non-adaptive baseline while maintaining low computational overhead. Replay-regularized statistical alignment provides the most stable adaptation under limited data, while meta-learning achieves the highest accuracy when sparse calibration labels are available. Overall, our self-supervised test-time adaptation methods reach up to 82% inter-session accuracy, significantly improving upon prior approaches while maintaining resource-efficient operation. These results demonstrate that lightweight test-time adaptation can enable robust, long-term EMG decoding for wearable or prosthetic applications.

cs.LG↗

arXiv2025

Multiscale dynamical characterization of cortical brain states: from synchrony to asynchrony

The cerebral cortex spontaneously displays different patterns of activity that evolve over time according to the brain state. Sleep, wakefulness, resting states, and attention are examples of a wide spectrum of physiological states that can be sustained by the same structural network. Furthermore, additional states are generated by drugs (e.g., different levels of anesthesia) or by pathological conditions (e.g., brain lesions, disorders of consciousness). While the significance of understanding brain states in relation to brain dynamics and behavior has become increasingly evident over the past two decades, a unified definition of brain states remains elusive. In this review, we focus on two extremes of this spectrum: synchronous versus asynchronous states. These functional states predominantly underlie unconsciousness and consciousness, respectively, although exceptions exist. Our aim is to integrate data from different levels into a multiscale understanding, ranging from local circuits to whole-brain dynamics, including properties such as cortical complexity, functional connectivity, synchronization, wave propagation, and excitatory-inhibitory balance that vary across states and characterize them. Experimental and clinical data, as well as computational models (at micro-, meso-, and macrocortical levels) associated with the discussed brain states, are made available to readers.

q-bio.NC↗

arXiv2025

Adaptive behavior with stable synapses

Behavioral changes in animals and humans, as a consequence of an error or a verbal instruction, can be extremely rapid. Improvement in behavioral performances are usually associated in machine learning and reinforcement learning to synaptic plasticity, and, in general, to changes and optimization of network parameters. However, such rapid changes are not coherent with the timescales of synaptic plasticity, suggesting that the mechanism responsible for that could be a dynamical network reconfiguration. In the last few years, similar capabilities have been observed in transformers, foundational architecture in the field of machine learning that are widely used in applications such as natural language and image processing. Transformers are capable of in-context learning, the ability to adapt and acquire new information dynamically within the context of the task or environment they are currently engaged in, without the need for significant changes to their underlying parameters. Building upon the notion of something unique within transformers enabling the emergence of this property, we claim that it could also be supported by input segregation and dendritic amplification, features extensively observed in biological networks. We propose an architecture composed of gain-modulated recurrent networks that excels at in-context learning, showing abilities inaccessible to standard networks.

q-bio.NC↗

arXiv2024

Neuromorphic dreaming: A pathway to efficient learning in artificial agents

Achieving energy efficiency in learning is a key challenge for artificial intelligence (AI) computing platforms. Biological systems demonstrate remarkable abilities to learn complex skills quickly and efficiently. Inspired by this, we present a hardware implementation of model-based reinforcement learning (MBRL) using spiking neural networks (SNNs) on mixed-signal analog/digital neuromorphic hardware. This approach leverages the energy efficiency of mixed-signal neuromorphic chips while achieving high sample efficiency through an alternation of online learning, referred to as the "awake" phase, and offline learning, known as the "dreaming" phase. The model proposed includes two symbiotic networks: an agent network that learns by combining real and simulated experiences, and a learned world model network that generates the simulated experiences. We validate the model by training the hardware implementation to play the Atari game Pong. We start from a baseline consisting of an agent network learning without a world model and dreaming, which successfully learns to play the game. By incorporating dreaming, the number of required real game experiences are reduced significantly compared to the baseline. The networks are implemented using a mixed-signal neuromorphic processor, with the readout layers trained using a computer in-the-loop, while the other layers remain fixed. These results pave the way toward energy-efficient neuromorphic learning systems capable of rapid learning in real world applications and use-cases.

cs.AI↗

arXiv2024

Learning fast changing slow in spiking neural networks

Reinforcement learning (RL) faces substantial challenges when applied to real-life problems, primarily stemming from the scarcity of available data due to limited interactions with the environment. This limitation is exacerbated by the fact that RL often demands a considerable volume of data for effective learning. The complexity escalates further when implementing RL in recurrent spiking networks, where inherent noise introduced by spikes adds a layer of difficulty. Life-long learning machines must inherently resolve the plasticity-stability paradox. Striking a balance between acquiring new knowledge and maintaining stability is crucial for artificial agents. To address this challenge, we draw inspiration from machine learning technology and introduce a biologically plausible implementation of proximal policy optimization, referred to as lf-cs (learning fast changing slow). Our approach results in two notable advancements: firstly, the capacity to assimilate new information into a new policy without requiring alterations to the current policy; and secondly, the capability to replay experiences without experiencing policy divergence. Furthermore, when contrasted with other experience replay (ER) techniques, our method demonstrates the added advantage of being computationally efficient in an online setting. We demonstrate that the proposed methodology enhances the efficiency of learning, showcasing its potential impact on neuromorphic and real-world applications.

cs.NE↗

arXiv2023

Towards biologically plausible Dreaming and Planning in recurrent spiking networks

Humans and animals can learn new skills after practicing for a few hours, while current reinforcement learning algorithms require a large amount of data to achieve good performances. Recent model-based approaches show promising results by reducing the number of necessary interactions with the environment to learn a desirable policy. However, these methods require biological implausible ingredients, such as the detailed storage of older experiences, and long periods of offline learning. The optimal way to learn and exploit word-models is still an open question. Taking inspiration from biology, we suggest that dreaming might be an efficient expedient to use an inner model. We propose a two-module (agent and model) spiking neural network in which "dreaming" (living new experiences in a model-based simulated environment) significantly boosts learning. We also explore "planning", an online alternative to dreaming, that shows comparable performances. Importantly, our model does not require the detailed storage of experiences, and learns online the world-model and the policy. Moreover, we stress that our network is composed of spiking neurons, further increasing the biological plausibility and implementability in neuromorphic hardware.

cs.LG↗

arXiv2023

Comparing apples to apples -- Using a modular and adaptable analysis pipeline to compare slow cerebral rhythms across heterogeneous datasets

Neuroscience is moving towards a more integrative discipline, where understanding brain function requires consolidating the accumulated evidence seen across experiments, species, and measurement techniques. A remaining challenge on that path is integrating such heterogeneous data into analysis workflows such that consistent and comparable conclusions can be distilled as an experimental basis for models and theories. Here, we propose a solution in the context of slow wave activity ($<1$ Hz), which occurs during unconscious brain states like sleep and general anesthesia, and is observed across diverse experimental approaches. We address the issue of integrating and comparing heterogeneous data by conceptualizing a general pipeline design that is adaptable to a variety of inputs and applications. Furthermore, we present the Collaborative Brain Wave Analysis Pipeline (Cobrawap) as a concrete, reusable software implementation to perform broad, detailed, and rigorous comparisons of slow wave characteristics across multiple, openly available ECoG and calcium imaging datasets.

q-bio.NC↗

arXiv2023

NREM and REM: cognitive and energetic gains in thalamo-cortical sleeping and awake spiking model

Sleep is essential for learning and cognition, but the mechanisms by which it stabilizes learning, supports creativity, and manages the energy consumption of networks engaged in post-sleep task have not been yet modelled. During sleep, the brain cycles between non-rapid eye movement (NREM), a mainly unconscious state characterized by collective oscillations, and rapid eye movement (REM), associated with the integrated experience of dreaming. We propose a biologically grounded two-area thalamo-cortical plastic spiking neural network model and investigate the role of NREM - REM cycles on its awake performance. We demonstrate that sleep has a positive effect on energy consumption and cognitive performance during the post-sleep awake classification task of handwritten digits. NREM and REM simulated dynamics modify the synaptic structure into a sharper representation of training experiences. Sleep-induced synaptic modifications reduce firing rates and synaptic activity without reducing cognitive performance. Also, it creates novel multi-area associations. The model leverages the apical amplification, isolation and drive experimentally grounded principles and the combination of contextual and perceptual information. In summary, the main novelty is the proposal of a multi-area plastic model that also expresses REM and integrates information during a plastic dream-like state, with cognitive and energetic benefits during post-sleep awake classification.

q-bio.NC↗

arXiv2022

Simulations Approaching Data: Cortical Slow Waves in Inferred Models of the Whole Hemisphere of Mouse

Thanks to novel, powerful brain activity recording techniques, we can create data-driven models from thousands of recording channels and large portions of the cortex, which can improve our understanding of brain-states neuromodulation and the related richness of traveling waves dynamics. We investigate the inference of data-driven models and the comparison among experiments and simulations, through the characterization of the spatio-temporal features of cortical waves in experimental recordings and simulations. Inference is built in two steps: the inner loop that optimizes by likelihood maximization a mean-field model, and the outer loop that optimizes a periodic neuro-modulation by relying on direct comparison of observables apt for the characterization of cortical slow waves. The model is capable to reproduce most of the features of the non-stationary and non-linear dynamics displayed by the high-resolution recording of the in-vivo mouse brain obtained by wide-field calcium imaging techniques. The proposed approach is of interest for both experimental and computational neuroscientists.

q-bio.NC↗

arXiv2022

Beyond spiking networks: the computational advantages of dendritic amplification and input segregation

The brain can efficiently learn a wide range of tasks, motivating the search for biologically inspired learning rules for improving current artificial intelligence technology. Most biological models are composed of point neurons, and cannot achieve the state-of-the-art performances in machine learning. Recent works have proposed that segregation of dendritic input (neurons receive sensory information and higher-order feedback in segregated compartments) and generation of high-frequency bursts of spikes would support error backpropagation in biological neurons. However, these approaches require propagating errors with a fine spatio-temporal structure to the neurons, which is unlikely to be feasible in a biological network. To relax this assumption, we suggest that bursts and dendritic input segregation provide a natural support for biologically plausible target-based learning, which does not require error propagation. We propose a pyramidal neuron model composed of three separated compartments. A coincidence mechanism between the basal and the apical compartments allows for generating high-frequency bursts of spikes. This architecture allows for a burst-dependent learning rule, based on the comparison between the target bursting activity triggered by the teaching signal and the one caused by the recurrent connections, providing the support for target-based learning. We show that this framework can be used to efficiently solve spatio-temporal tasks, such as the store and recall of 3D trajectories. Finally, we suggest that this neuronal architecture naturally allows for orchestrating ``hierarchical imitation learning'', enabling the decomposition of challenging long-horizon decision-making tasks into simpler subtasks. This can be implemented in a two-level network, where the high-network acts as a ``manager'' and produces the contextual signal for the low-network, the ``worker''.

q-bio.NC↗

arXiv2022

Burst-dependent plasticity and dendritic amplification support target-based learning and hierarchical imitation learning

The brain can learn to solve a wide range of tasks with high temporal and energetic efficiency. However, most biological models are composed of simple single compartment neurons and cannot achieve the state-of-art performances of artificial intelligence. We propose a multi-compartment model of pyramidal neuron, in which bursts and dendritic input segregation give the possibility to plausibly support a biological target-based learning. In target-based learning, the internal solution of a problem (a spatio temporal pattern of bursts in our case) is suggested to the network, bypassing the problems of error backpropagation and credit assignment. Finally, we show that this neuronal architecture naturally support the orchestration of hierarchical imitation learning, enabling the decomposition of challenging long-horizon decision-making tasks into simpler subtasks.

q-bio.NC↗

arXiv2021

Error-based or target-based? A unifying framework for learning in recurrent spiking networks

Learning in biological or artificial networks means changing the laws governing the network dynamics in order to better behave in a specific situation. In the field of supervised learning, two complementary approaches stand out: error-based and target-based learning. However, there exists no consensus on which is better suited for which task, and what is the most biologically plausible. Here we propose a comprehensive theoretical framework that includes these two frameworks as special cases. This novel theoretical formulation offers major insights into the differences between the two approaches. In particular, we show how target-based naturally emerges from error-based when the number of constraints on the target dynamics, and as a consequence on the internal network dynamics, is comparable to the degrees of freedom of the network. Moreover, given the experimental evidences on the relevance that spikes have in biological networks, we investigate the role of coding with specific patterns of spikes by introducing a parameter that defines the tolerance to precise spike timing during learning. Our approach naturally lends itself to Imitation Learning (and Behavioral Cloning in particular) and we apply it to solve relevant closed-loop tasks such as the button-and-food task, and the 2D Bipedal Walker. We show that a high dimensionality feedback structure is extremely important when it is necessary to solve a task that requires retaining memory for a long time (button-and-food). On the other hand, we find that coding with specific patterns of spikes enables optimal performances in a motor task (the 2D Bipedal Walker). Finally, we show that our theoretical formulation suggests protocols to deduce the structure of learning feedback in biological networks.

q-bio.NC↗

arXiv2021

Thalamo-cortical spiking model of incremental learning combining perception, context and NREM-sleep-mediated noise-resilience

The brain exhibits capabilities of fast incremental learning from few noisy examples, as well as the ability to associate similar memories in autonomously-created categories and to combine contextual hints with sensory perceptions. Together with sleep, these mechanisms are thought to be key components of many high-level cognitive functions. Yet, little is known about the underlying processes and the specific roles of different brain states. In this work, we exploited the combination of context and perception in a thalamo-cortical model based on a soft winner-take-all circuit of excitatory and inhibitory spiking neurons. After calibrating this model to express awake and deep-sleep states with features comparable with biological measures, we demonstrate the model capability of fast incremental learning from few examples, its resilience when proposed with noisy perceptions and contextual signals, and an improvement in visual classification after sleep due to induced synaptic homeostasis and association of similar memories.

q-bio.NC↗

arXiv2021

Target spiking patterns enable efficient and biologically plausible learning for complex temporal tasks

Recurrent spiking neural networks (RSNN) in the human brain learn to perform a wide range of perceptual, cognitive and motor tasks very efficiently in terms of energy consumption and requires very few examples. This motivates the search for biologically inspired learning rules for RSNNs to improve our understanding of brain computation and the efficiency of artificial intelligence. Several spiking models and learning rules have been proposed, but it remains a challenge to design RSNNs whose learning relies on biologically plausible mechanisms and are capable of solving complex temporal tasks. In this paper, we derive a learning rule, local to the synapse, from a simple mathematical principle, the maximization of the likelihood for the network to solve a specific task. We propose a novel target-based learning scheme in which the learning rule derived from likelihood maximization is used to mimic a specific spiking pattern that encodes the solution to complex temporal tasks. This method makes the learning extremely rapid and precise, outperforming state of the art algorithms for RSNNs. We demonstrate the capacity of our model to tackle several problems like learning multidimensional trajectories and solving the classical temporal XOR benchmark. Finally, we show that an online approximation of the gradient ascent, in addition to guaranteeing complete locality in time and space, allows learning after very few presentations of the target output. Our model can be applied to different types of biological neurons. The analytically derived plasticity learning rule is specific to each neuron model and can produce a theoretical prediction for experimental validation.

q-bio.NC↗