Discovering functional connectivity features characterizing multiple sclerosis phenotypes using explainable artificial intelligence.

Multiple sclerosis (MS) is a neurological condition characterized by severe structural brain damage and by functional reorganization of the main brain networks that try to limit the clinical consequences of structural burden. Resting-state (RS) functional connectivity (FC) abnormalities found in this condition were shown to be variable across different MS phases, according to the severity of clinical manifestations. The article describes a system exploiting machine learning on RS FC matrices to discriminate different MS phenotypes and to identify relevant functional connections for MS stage characterization. To this end, the system exploits some mathematical properties of covariance-based RS FC representation, which can be described by a Riemannian manifold. The classification performance of the proposed framework was significantly above the chance level for all MS phenotypes. Moreover, the proposed system was successful in identifying relevant RS FC alterations contributing to an accurate phenotype classification.

AIforCOVID: Predicting the clinical outcomes in patients with COVID-19 applying AI to chest-X-rays. An Italian multicentre study.

Recent epidemiological data report that worldwide more than 53 million people have been infected by SARS-CoV-2, resulting in 1.3 million deaths. The disease has been spreading very rapidly and few months after the identification of the first infected, shortage of hospital resources quickly became a problem. In this work we investigate whether artificial intelligence working with chest X-ray (CXR) scans and clinical data can be used as a possible tool for the early identification of patients at risk of severe outcome, like intensive care or death. Indeed, further to induce lower radiation dose than computed tomography (CT), CXR is a simpler and faster radiological technique, being also more widespread. In this respect, we present three approaches that use features extracted from CXR images, either handcrafted or automatically learnt by convolutional neuronal networks, which are then integrated with the clinical data. As a further contribution, this work introduces a repository that collects data from 820 patients enrolled in six Italian hospitals in spring 2020 during the first COVID-19 emergency. The dataset includes CXR images, several clinical attributes and clinical outcomes. Exhaustive evaluation shows promising performance both in 10-fold and leave-one-centre-out cross-validation, suggesting that clinical data and images have the potential to provide useful information for the management of patients and hospital resources.

Exploring the Embodiment of a Virtual Hand in a Spatially Augmented Respiratory Biofeedback Setting.

Enhancing the embodiment of artificial limbs-the individuals' feeling that a virtual or robotic limb is integrated in their own body scheme-is an impactful strategy for improving prosthetic technology acceptance and human-machine interaction. Most studies so far focused on visuo-tactile strategies to empower the embodiment processes. However, novel approaches could emerge from self-regulation techniques able to change the psychophysiological conditions of an individual. Accordingly, this pilot study investigates the effects of a self-regulated breathing exercise on the processes of body ownership underlying the embodiment of a virtual right hand within a Spatially Augmented Respiratory Biofeedback (SARB) setting. This investigation also aims at evaluating the feasibility of the breathing exercise enabled by a low-cost SARB implementation designed for upcoming remote studies (a need emerged during the COVID-19 pandemic). Twenty-two subjects without impairments, and two transradial prosthesis users for a preparatory test, were asked (in each condition of a within-group design) to maintain a normal (about 14 breaths/min) or slow (about 6 breaths/min) respiratory rate to keep a static virtual right hand "visible" on a screen. Meanwhile, a computer-generated sphere moved from left to right toward the virtual hand during each trial (1 min) of 16. If the participant's breathing rate was within the target (slow or normal) range, a visuo-tactile event was triggered by the sphere passing under the virtual hand (the subjects observed it shaking while they perceived a vibratory feedback generated by a smartphone). Our results-mainly based on questionnaire scores and proprioceptive drift-highlight that the slow breathing condition induced higher embodiment than the normal one. This preliminary study reveals the feasibility and potential of a novel psychophysiological training strategy to enhance the embodiment of artificial limbs. Future studies are needed to further investigate mechanisms, efficacy and generalizability of the SARB techniques in training a bionic limb embodiment.

A Neuromorphic Prosthesis to Restore Communication in Neuronal Networks.

Recent advances in bioelectronics and neural engineering allowed the development of brain machine interfaces and neuroprostheses, capable of facilitating or recovering functionality in people with neurological disability. To realize energy-efficient and real-time capable devices, neuromorphic computing systems are envisaged as the core of next-generation systems for brain repair. We demonstrate here a real-time hardware neuromorphic prosthesis to restore bidirectional interactions between two neuronal populations, even when one is damaged or missing. We used in vitro modular cell cultures to mimic the mutual interaction between neuronal assemblies and created a focal lesion to functionally disconnect the two populations. Then, we employed our neuromorphic prosthesis for bidirectional bridging to artificially reconnect two disconnected neuronal modules and for hybrid bidirectional bridging to replace the activity of one module with a real-time hardware neuromorphic Spiking Neural Network. Our neuroprosthetic system opens avenues for the exploitation of neuromorphic-based devices in bioelectrical therapeutics for health care.

Closed-Loop Systems and In Vitro Neuronal Cultures: Overview and Applications.

One of the main limitations preventing the realization of a successful dialogue between the brain and a putative enabling device is the intricacy of brain signals. In this perspective, closed-loop in vitro systems can be used to investigate the interactions between a network of neurons and an external system, such as an interacting environment or an artificial device. In this chapter, we provide an overview of closed-loop in vitro systems, which have been developed for investigating potential neuroprosthetic applications. In particular, we first explore how to modify or set a target dynamical behavior in a network of neurons. We then analyze the behavior of in vitro systems connected to artificial devices, such as robots. Finally, we provide an overview of biological neuronal networks interacting with artificial neuronal networks, a configuration currently offering a promising solution for clinical applications.

Soft brain-machine interfaces for assistive robotics: A novel control approach.

Robotic systems offer the possibility of improving the life quality of people with severe motor disabilities, enhancing the individual's degree of independence and interaction with the external environment. In this direction, the operator's residual functions must be exploited for the control of the robot movements and the underlying dynamic interaction through intuitive and effective human-robot interfaces. Towards this end, this work aims at exploring the potential of a novel Soft Brain-Machine Interface (BMI), suitable for dynamic execution of remote manipulation tasks for a wide range of patients. The interface is composed of an eye-tracking system, for an intuitive and reliable control of a robotic arm system's trajectories, and a Brain-Computer Interface (BCI) unit, for the control of the robot Cartesian stiffness, which determines the interaction forces between the robot and environment. The latter control is achieved by estimating in real-time a unidimensional index from user's electroencephalographic (EEG) signals, which provides the probability of a neutral or active state. This estimated state is then translated into a stiffness value for the robotic arm, allowing a reliable modulation of the robot's impedance. A preliminary evaluation of this hybrid interface concept provided evidence on the effective execution of tasks with dynamic uncertainties, demonstrating the great potential of this control method in BMI applications for self-service and clinical care.

Investigating the impact of electrical stimulation temporal distribution on cortical network responses.

BACKGROUND: The brain is continuously targeted by a wealth of stimuli with complex spatio-temporal patterns and has presumably evolved in order to cope with those inputs in an optimal way. Previous studies investigating the response capabilities of either single neurons or intact sensory systems to external stimulation demonstrated that stimuli temporal distribution is an important, if often overlooked, parameter. RESULTS: In this study we investigated how cortical networks plated over micro-electrode arrays respond to different stimulation sequences in which inter-pulse intervals followed a 1/f ß distribution, for different values of ß ranging from 0 to ∞. Cross-correlation analysis revealed that network activity preferentially synchronizes with external input sequences featuring ß closer to 1 and, in any case, never for regular (i.e. fixed-frequency) stimulation sequences. We then tested the interplay between different average stimulation frequencies (based on the intrinsic firing/bursting frequency of the network) for two selected values of ß, i.e. 1 (scale free) and ∞ (regular). In general, we observed no preference for stimulation frequencies matching the endogenous rhythms of the network. Moreover, we found that in case of regular stimulation the capability of the network to follow the stimulation sequence was negatively correlated to the absolute stimulation frequency, whereas using scale-free stimulation cross-correlation between input and output sequences was independent from average input frequency. CONCLUSIONS: Our results point out that the preference for a scale-free distribution of the stimuli is observed also at network level and should be taken into account in designing more efficient protocols for neuromodulation purposes.

In vitro studies of neuronal networks and synaptic plasticity in invertebrates and in mammals using multielectrode arrays.

Brain functions are strictly dependent on neural connections formed during development and modified during life. The cellular and molecular mechanisms underlying synaptogenesis and plastic changes involved in learning and memory have been analyzed in detail in simple animals such as invertebrates and in circuits of mammalian brains mainly by intracellular recordings of neuronal activity. In the last decades, the evolution of techniques such as microelectrode arrays (MEAs) that allow simultaneous, long-lasting, noninvasive, extracellular recordings from a large number of neurons has proven very useful to study long-term processes in neuronal networks in vivo and in vitro. In this work, we start off by briefly reviewing the microelectrode array technology and the optimization of the coupling between neurons and microtransducers to detect subthreshold synaptic signals. Then, we report MEA studies of circuit formation and activity in invertebrate models such as Lymnaea, Aplysia, and Helix. In the following sections, we analyze plasticity and connectivity in cultures of mammalian dissociated neurons, focusing on spontaneous activity and electrical stimulation. We conclude by discussing plasticity in closed-loop experiments.

Closed-loop neuro-robotic experiments to test computational properties of neuronal networks.

Information coding in the Central Nervous System (CNS) remains unexplored. There is mounting evidence that, even at a very low level, the representation of a given stimulus might be dependent on context and history. If this is actually the case, bi-directional interactions between the brain (or if need be a reduced model of it) and sensory-motor system can shed a light on how encoding and decoding of information is performed. Here an experimental system is introduced and described in which the activity of a neuronal element (i.e., a network of neurons extracted from embryonic mammalian hippocampi) is given context and used to control the movement of an artificial agent, while environmental information is fed back to the culture as a sequence of electrical stimuli. This architecture allows a quick selection of diverse encoding, decoding, and learning algorithms to test different hypotheses on the computational properties of neuronal networks.

Interfacing in silico and in vitro neuronal networks.

In this study, an in vitro cortical culture was stimulated according to the activity exhibited by its in silico counterpart. We connected the latter, an artificial spiking neural network (SNN), to the former, a biological network (BNN), via an open-loop (i.e. unidirectional) configuration, with the ultimate goal of establishing a closed-loop bidirectional connection in future studies. We detected network bursts in the SNN and utilized them as triggers for stimulus delivery to the BNN. We analyzed evoked BNN responses to evaluate whether the BNN activity was entrained by the SNN by correlating stimuli and BNN evoked network burst rates at different timescales. We found that this correlation was nearly constant at all timescales, but its magnitude depended on the efficacy of the stimulation source in entraining BNN activity.

Impact of stimuli distribution on neural network responses.

In this study we investigated the relationship between stimulus regularity and observed responses in a biological neural network of dissociated cortical rat neurons plated over Micro-Electrode Arrays (MEAs). In particular, the intervals between identical stimuli in our experiments followed a 1/f(ß) distribution, and regularity increased with the value of ß (values tested were 0, 0.5, 1, 1.5, ∞). Comparisons occurred on the correlation between low-passed (rectangular window, 0.1s in length) stimulation trains and network-wide spike trains. Our results show that cultures are largely unable to synchronize network-wide responses with regular stimulation at the considered stimulation rate (0.5Hz), while this occurs to a much higher degree for irregular stimulations.

In vitro large-scale experimental and theoretical studies for the realization of bi-directional brain-prostheses.

Brain-machine interfaces (BMI) were born to control "actions from thoughts" in order to recover motor capability of patients with impaired functional connectivity between the central and peripheral nervous system. The final goal of our studies is the development of a new proof-of-concept BMI-a neuromorphic chip for brain repair-to reproduce the functional organization of a damaged part of the central nervous system. To reach this ambitious goal, we implemented a multidisciplinary "bottom-up" approach in which in vitro networks are the paradigm for the development of an in silico model to be incorporated into a neuromorphic device. In this paper we present the overall strategy and focus on the different building blocks of our studies: (i) the experimental characterization and modeling of "finite size networks" which represent the smallest and most general self-organized circuits capable of generating spontaneous collective dynamics; (ii) the induction of lesions in neuronal networks and the whole brain preparation with special attention on the impact on the functional organization of the circuits; (iii) the first production of a neuromorphic chip able to implement a real-time model of neuronal networks. A dynamical characterization of the finite size circuits with single cell resolution is provided. A neural network model based on Izhikevich neurons was able to replicate the experimental observations. Changes in the dynamics of the neuronal circuits induced by optical and ischemic lesions are presented respectively for in vitro neuronal networks and for a whole brain preparation. Finally the implementation of a neuromorphic chip reproducing the network dynamics in quasi-real time (10 ns precision) is presented.

Modular neuronal assemblies embodied in a closed-loop environment: toward future integration of brains and machines.

Behaviors, from simple to most complex, require a two-way interaction with the environment and the contribution of different brain areas depending on the orchestrated activation of neuronal assemblies. In this work we present a new hybrid neuro-robotic architecture based on a neural controller bi-directionally connected to a virtual robot implementing a Braitenberg vehicle aimed at avoiding obstacles. The robot is characterized by proximity sensors and wheels, allowing it to navigate into a circular arena with obstacles of different sizes. As neural controller, we used hippocampal cultures dissociated from embryonic rats and kept alive over Micro Electrode Arrays (MEAs) for 3-8 weeks. The developed software architecture guarantees a bi-directional exchange of information between the natural and the artificial part by means of simple linear coding/decoding schemes. We used two different kinds of experimental preparation: "random" and "modular" populations. In the second case, the confinement was assured by a polydimethylsiloxane (PDMS) mask placed over the surface of the MEA device, thus defining two populations interconnected via specific microchannels. The main results of our study are: (i) neuronal cultures can be successfully interfaced to an artificial agent; (ii) modular networks show a different dynamics with respect to random culture, both in terms of spontaneous and evoked electrophysiological patterns; (iii) the robot performs better if a reinforcement learning paradigm (i.e., a tetanic stimulation delivered to the network following each collision) is activated, regardless of the modularity of the culture; (iv) the robot controlled by the modular network further enhances its capabilities in avoiding obstacles during the short-term plasticity trial. The developed paradigm offers a new framework for studying, in simplified model systems, neuro-artificial bi-directional interfaces for the development of new strategies for brain-machine interaction.