Nicolas Farrugia

IMT Atlantique, Brest, France · Electronics

It is well established that auditory cueing improves gait in patients with idiopathic Parkinson’s disease (IPD). Disease-related reductions in speed and step length can be improved by providing rhythmical auditory cues via a metronome or music. However, effects on cognitive aspects of motor control have yet to be thoroughly investigated. If synchro...

Perceived regularity among events in the environment allows predictions regarding the "when" and the "what" dimensions of future events. In this context, one crucial question concerns the impact and the potentially optimizing effect, of regular temporal structure on the processing of "what", or formal, information. The current study addresses this...

Recent years have seen a growing interest in the neuroscience of spontaneous cognition. One form of such cognition is involuntary musical imagery (INMI), the non-pathological and everyday experience of having music in one's head, in the absence of an external stimulus. In this study, aspects of INMI, including frequency and affective evaluation, we...

Auditory stimulation via rhythmic cues can be used successfully in the rehabilitation of motor function in patients with motor disorders. A prototypical example is provided by dysfunctional gait in patients with idiopathic Parkinson's disease (PD). Coupling steps to external rhythmic cues (the beat of music or the sounds of a metronome) leads to lo...

Passive acoustic monitoring (PAM) is crucial for bioacoustic research, enabling non-invasive species tracking and biodiversity monitoring. Citizen science platforms like Xeno-Canto provide large annotated datasets from focal recordings, where the target species is intentionally recorded. However, PAM requires monitoring in passive soundscapes, crea...

We propose EEG-SimpleConv, a straightforward 1D convolutional neural network for Motor Imagery decoding in BCI. Our main motivation is to propose a simple and performing baseline that achieves high classification accuracy, using only standard ingredients from the literature, to serve as a standard for comparison. The proposed architecture is compos...

Political responses to the COVID-19 pandemic led to changes in city soundscapes around the globe.From March to October 2020, a consortium of 261 contributors from 35 countries brought togetherby the Silent Cities project built a unique soundscape recordings collection to report on local acousticchanges in urban areas. We present this collection her...

The intricate structural and functional architecture of the brain enables a wide range of cognitive processes ranging from perception and action to higher-order abstract thinking. Despite important progress, the relationship between the brain’s structural and functional properties is not yet fully established. In particular, the way the brain’s ana...

Humans can easily extract the rhythm of a complex sound, like music, and move to its regular beat, like in dance. These abilities are modulated by musical training and vary significantly in untrained individuals. The causes of this variability are multidimensional and typically hard to grasp in single tasks. To date we lack a comprehensive model ca...

Wind speed retrieval at the sea surface is of primary importance for scientific and operational applications. Besides weather models, in-situ measurements and remote sensing technologies, especially satellite sensors, provide complementary means to monitor wind speed. As sea-surface winds produce sounds that propagate underwater, underwater acousti...

Artificial neural networks are emerging as key tools to model brain processes associated with sound in auditory neuroscience. Most modelling works fit a single model with brain activity averaged across a group of subjects, ignoring individual-specific features of brain organisation. We demonstrate here the creation of personalised auditory artifici...

Deep learning has been widely used recently for sound event detection and classification. Its success is linked to the availability of sufficiently large datasets, possibly with corresponding annotations when supervised learning is considered. In bioacoustic applications, most tasks come with few labelled training data, because annotating long reco...

Humans can easily extract the rhythm of a complex sound, like music, and move to its regular beat, for example in dance. These abilities are modulated by musical training and vary significantly in untrained individuals. The causes of this variability are multidimensional and typically hard to grasp with single tasks. To date we lack a comprehensive...

BCI Motor Imagery datasets usually are small and have different electrodes setups. When training a Deep Neural Network, one may want to capitalize on all these datasets to increase the amount of data available and hence obtain good generalization results. To this end, we introduce a spatial graph signal interpolation technique, that allows to inter...

Wind speed retrieval at sea surface is of primary importance for scientific and operational applications. Besides weather models, in-situ measurements and remote sensing technologies, especially satellite sensors, provide complementary means to monitor wind speed. As sea surface winds produce sounds that propagate underwater, underwater acoustics r...

Decoding cognitive processes from recordings of brain activity has been an active topic in neuroscience research for decades. Traditional decoding studies focused on pattern classification in specific regions of interest and averaging brain activity over many trials. Recently, brain decoding with graph neural networks has been shown to scale at fin...

Graph Signal Processing is a promising framework to manipulate brain signals as it allows to encompass the spatial dependencies between the activity in regions of interest in the brain. In this work, we are interested in better understanding what are the graph frequencies that are the most useful to decode fMRI signals. To this end, we introduce a...

It is very common to face classification problems where the number of available labeled samples is small compared to their dimension. These conditions are likely to cause underdetermined settings, with high risk of overfitting. To improve the generalization ability of trained classifiers, common solutions include using priors about the data distrib...

In this paper, we describe the results of a single subject study attempting at a better understanding of the subjective mental state during musical improvisation. In a first experiment, we setup an ecological paradigm measuring EEG on a musician in free improvised concerts with an audience, followed by retrospective rating of the mental state of th...

Alcohol use disorder (AUD) is widely associated with cerebellar dysfunction and altered cerebro-cerebellar functional connectivity (FC) that lead to cognitive impairments. Evidence for this association comes from resting-state functional magnetic resonance imaging (rsfMRI) studies that assess time-averaged measures of FC across the duration of a ty...

The application of graph theory to model the complex structure and function of the brain has shed new light on its organization, prompting the emergence of network neuroscience. Despite the tremendous progress that has been achieved in this field, still relatively few methods exploit the topology of brain networks to analyze brain activity. Recent...

In this paper, we describe the results of a single subject study attempting at a better understanding of the subjective state during musical improvisation. In a first experiment, we setup an ecological paradigm measuring EEG on a musician in free improvised concerts with an audience, followed by retrospective rating of the mental state of the impro...

Few-shot learning consists in addressing data-thrifty (inductive few-shot) or label-thrifty (transductive few-shot) problems. So far, the field has been mostly driven by applications in computer vision. In this work, we are interested in stressing the ability of recently introduced few-shot methods to solve problems dealing with neuroimaging data,...

The application of graph theory to model the complex structure and function of the brain has shed new light on its organization and function, prompting the emergence of network neuroscience. Despite the tremendous progress that has been achieved in this field, still relatively few methods exploit the topology of brain networks to analyze brain acti...

Human and animal brain studies bring converging evidence on a possible role for the cerebellum and the cerebro-cerebellar system in impulsivity. However, the precise nature of the relation between cerebro-cerebellar coupling and impulsivity is far from understood. Characterizing functional connectivity (FC) patterns between large-scale brain networ...

Converging evidence from human and animal studies predict a possible role of the cerebellum in impulsivity. However, this hypothesis has not been thoroughly investigated within the framework of functional connectivity (FC). To address this issue, we employed resting-state functional magnetic resonance imaging data and two self-reports of impulsivit...

In this paper, we tackle the problem of incrementally learning a classifier, one example at a time, directly on chip. To this end, we propose an efficient hardware implementation of a recently introduced incremental learning procedure that achieves state-of-the-art performance by combining transfer learning with majority votes and quantization tech...

Graph Signal Processing has become a very useful framework for signal operations and representations defined on irregular domains. Exploiting transformations that are defined on graph models can be highly beneficial when the graph encodes relationships between signals. In this work, we present the benefits of using Spectral Graph Wavelet Transform...

For the past few years, Deep Neural Networks (DNNs) have achieved state-of-art performance in numerous challenging domains. To reach this performance, DNNs consist in large sets of parameters and complex architectures, which are trained offline on huge datasets. The complexity and size of DNNs architectures make it difficult to implement such appro...

La transformée de Fourier sur graphe (TFG) pourrait être un outil essentiel pour l'analyse des signaux cérébraux. En ce sens, nous évaluons l'application du traitement de signal sur graphes (TSG) pour l'analyse des données de neuroimagerie. Ainsi, une approche basée sur le TSG est proposée et validée pour la classification des troubles du spectre a...

Graph Fourier Transform (GFT) could be a key tool for analyzing brain signals (Huang et al. 2018). In this work, we evaluate the application of Graph signal processing (GSP) for the analysis of neuroimaging data. More specifically, we characterize each fMRI time series of each brain subject by its standard deviation (STD). Then, we project these st...

Graph Fourier Transform (GFT) could be a key tool for analyzing brain signals. In this sense, we evaluate the application of Graph signal processing (GSP) for the analysis of neuroimaging data. Thus, a GSP-based approach is proposed and validated for the classification of autism spectrum disorder (ASD). More specifically, the resting state function...

Predicting the future of Graph-supported Time Series (GTS) is a key challenge in many domains, such as climate monitoring, finance or neuroimaging. Yet it is a highly difficult problem as it requires to account jointly for time and graph (spatial) dependencies. To simplify this process, it is common to use a two-step procedure in which spatial and...

Trait impulsivity is a component of personality that involves a tendency to display behavior with little to no forethought nor consideration of consequences. Impulsivity takes a part in the disease pattern of several neuropsychiatric disorders that exhibit alterations in cortico-cerebellar loops such as addiction and ADHD. We hypothesize that the f...

The relationship between musical and linguistic skills has received particular attention in infants and school-aged children. However, very little is known about pre-schoolers. This leaves a gap in our understanding of the concurrent development of these skills during development. Moreover, attention has been focused on the effects of formal musica...

We present a publicly available dataset of 228 healthy participants comprising a young (N=154, 25.1±3.1 years, range 20–35 years, 45 female) and an elderly group (N=74, 67.6±4.7 years, range 59–77 years, 37 female) acquired cross-sectionally in Leipzig, Germany, between 2013 and 2015 to study mind-body-emotion interactions. During a two-day assessm...

The dataset enables exploration of higher-order cognitive faculties, self-generated mental experience, and personality features in relation to the intrinsic functional architecture of the brain. We provide multimodal magnetic resonance imaging (MRI) data and a broad set of state and trait phenotypic assessments: mind-wandering, personality traits,...

Convolutional Neural Networks (CNNs) are state-of-the-art in numerous computer vision tasks such as object classification and detection. However, the large amount of parameters they contain leads to a high computational complexity and strongly limits their usability in budget-constrained devices such as embedded devices. In this paper, we propose a...

Deep learning-based methods have reached state of the art performances, relying on a large quantity of available data and computational power. Such methods still remain highly inappropriate when facing a major open machine learning problem, which consists of learning incrementally new classes and examples over time. Combining the outstanding perfor...

Deep learning-based methods have reached state of the art performances, relying on large quantity of available data and computational power. Such methods still remain highly inappropriate when facing a major open machine learning problem, which consists of learning incrementally new classes and examples over time. Combining the outstanding performa...

Comparisons between involuntarily and voluntarily retrieved autobiographical memories have revealed similarities in encoding and maintenance, with differences in terms of specificity and emotional responses. Our study extended this research area into the domain of musical memory, which afforded a unique opportunity to compare the same memory as acc...

The dataset enables exploration of higher-order cognitive faculties, self-generated mental experience, and personality features in relation to the intrinsic functional architecture of the brain. We provide multimodal magnetic resonance imaging (MRI) data and a broad set of state and trait phenotypic assessments: mind-wandering, personality traits,...

Graph Signal Processing (GSP) is a promising method to analyze high-dimensional neuroimaging datasets while taking into account both the spatial and functional dependencies between brain signals. In the present work, we apply GSP with dimensionality reduction techniques to decode brain activity from real and simulated fMRI datasets. We introduce se...

The Battery for the Assessment of Auditory Sensorimotor and Timing Abilities (BAASTA) is a new tool for the systematic assessment of perceptual and sensorimotor timing skills. It spans a broad range of timing skills aimed at differentiating individual timing profiles. BAASTA consists of sensitive time perception and production tasks. Perceptual tas...

Training based on rhythmic auditory stimulation (RAS) can improve gait in patients with idiopathic Parkinson’s disease (IPD). Patients typically walk faster and exhibit greater stride length after RAS. However, this effect is highly variable among patients, with some exhibiting little or no response to the intervention. These individual differences...

Both musically trained and untrained adults can reproduce the tempo of familiar music with high precision. However, conflicting evidence exists as to how well representations of tempo are preserved within musical imagery. The present study investigated whether previous conflicting evidence might result from the use of different tasks to measure ima...

The study of spontaneous and everyday cognitions is an area of rapidly growing interest. One of the most ubiquitous forms of spontaneous cognition is involuntary musical imagery (INMI), the involuntarily retrieved and repetitive mental replay of music. The present study introduced a novel method for capturing temporal features of INMI within a natu...

Parkinson's disease (PD) is a neurodegenerative disorder that targets mainly dopaminergic neurons of the basal ganglia. PD is characterized by motor symptoms typically leading to dysfunctional gait. External rhythmic auditory cues have shown beneficial effects on gait kinematics in PD patients. These effects are likely to be mediated by a general-p...

The beneficial effect of auditory cueing on gait performance in Parkinson's disease (PD) has been widely documented. Nevertheless, little is known about the neural underpinnings of this effect and the consequences of auditory cueing beyond improved gait kinematics. The therapy relies on processing the temporal regularity in an auditory signal to wh...

Several studies have focused on coupling between perception and action (e.g., via finger tapping), and more generally on entrainment, in average musicians and non-musicians. Yet, little is known about the effects of entrainment on timing and movement kinematics in individuals exhibiting outstanding rhythmical abilities. In this study we examined th...

In this paper, we present a parallel architecture for fast and robust face detection implemented on FPGA hardware. We propose the first implementation that meets both real-time requirements in an embedded context and face detection robustness within complex backgrounds. The chosen face detection method is the Convolutional Face Finder (CFF) algorit...

We describe a High-Level Synthesis implementation of a parallel architecture for face detection. The chosen face detection method is the well-known Convolutional Face Finder (CFF) algorithm, which consists of a pipeline of convolution operations. We rely on dataflow modelling of the algorithm and we use a high-level synthesis tool in order to speci...