Graph-Based Audio Classification Using Pre-Trained Models and Graph Neural Networks.

Sound classification plays a crucial role in enhancing the interpretation, analysis, and use of acoustic data, leading to a wide range of practical applications, of which environmental sound analysis is one of the most important. In this paper, we explore the representation of audio data as graphs in the context of sound classification. We propose a methodology that leverages pre-trained audio models to extract deep features from audio files, which are then employed as node information to build graphs. Subsequently, we train various graph neural networks (GNNs), specifically graph convolutional networks (GCNs), GraphSAGE, and graph attention networks (GATs), to solve multi-class audio classification problems. Our findings underscore the effectiveness of employing graphs to represent audio data. Moreover, they highlight the competitive performance of GNNs in sound classification endeavors, with the GAT model emerging as the top performer, achieving a mean accuracy of 83% in classifying environmental sounds and 91% in identifying the land cover of a site based on its audio recording. In conclusion, this study provides novel insights into the potential of graph representation learning techniques for analyzing audio data.

Soundscape Characterization Using Autoencoders and Unsupervised Learning.

Passive acoustic monitoring (PAM) through acoustic recorder units (ARUs) shows promise in detecting early landscape changes linked to functional and structural patterns, including species richness, acoustic diversity, community interactions, and human-induced threats. However, current approaches primarily rely on supervised methods, which require prior knowledge of collected datasets. This reliance poses challenges due to the large volumes of ARU data. In this work, we propose a non-supervised framework using autoencoders to extract soundscape features. We applied this framework to a dataset from Colombian landscapes captured by 31 audiomoth recorders. Our method generates clusters based on autoencoder features and represents cluster information with prototype spectrograms using centroid features and the decoder part of the neural network. Our analysis provides valuable insights into the distribution and temporal patterns of various sound compositions within the study area. By utilizing autoencoders, we identify significant soundscape patterns characterized by recurring and intense sound types across multiple frequency ranges. This comprehensive understanding of the study area's soundscape allows us to pinpoint crucial sound sources and gain deeper insights into its acoustic environment. Our results encourage further exploration of unsupervised algorithms in soundscape analysis as a promising alternative path for understanding and monitoring environmental changes.

Automatic Detection of Focal Cortical Dysplasia Using MRI: A Systematic Review.

Focal cortical dysplasia (FCD) is a congenital brain malformation that is closely associated with epilepsy. Early and accurate diagnosis is essential for effectively treating and managing FCD. Magnetic resonance imaging (MRI)-one of the most commonly used non-invasive neuroimaging methods for evaluating the structure of the brain-is often implemented along with automatic methods to diagnose FCD. In this review, we define three categories for FCD identification based on MRI: visual, semi-automatic, and fully automatic methods. By conducting a systematic review following the PRISMA statement, we identified 65 relevant papers that have contributed to our understanding of automatic FCD identification techniques. The results of this review present a comprehensive overview of the current state-of-the-art in the field of automatic FCD identification and highlight the progress made and challenges ahead in developing reliable, efficient methods for automatic FCD diagnosis using MRI images. Future developments in this area will most likely lead to the integration of these automatic identification tools into medical image-viewing software, providing neurologists and radiologists with enhanced diagnostic capabilities. Moreover, new MRI sequences and higher-field-strength scanners will offer improved resolution and anatomical detail for precise FCD characterization. This review summarizes the current state of automatic FCD identification, thereby contributing to a deeper understanding and the advancement of FCD diagnosis and management.

Influence of Patient-Specific Head Modeling on EEG Source Imaging.

Electromagnetic source imaging (ESI) techniques have become one of the most common alternatives for understanding cognitive processes in the human brain and for guiding possible therapies for neurological diseases. However, ESI accuracy strongly depends on the forward model capabilities to accurately describe the subject's head anatomy from the available structural data. Attempting to improve the ESI performance, we enhance the brain structure model within the individual-defined forward problem formulation, combining the head geometry complexity of the modeled tissue compartments and the prior knowledge of the brain tissue morphology. We validate the proposed methodology using 25 subjects, from which a set of magnetic-resonance imaging scans is acquired, extracting the anatomical priors and an electroencephalography signal set needed for validating the ESI scenarios. Obtained results confirm that incorporating patient-specific head models enhances the performed accuracy and improves the localization of focal and deep sources.

Solving the EEG inverse problem based on space-time-frequency structured sparsity constraints.

We introduce STOUT (spatio-temporal unifying tomography), a novel method for the source analysis of electroencephalograpic (EEG) recordings, which is based on a physiologically-motivated source representation. Our method assumes that only a small number of brain sources are active throughout a measurement, where each of the sources exhibits focal (smooth but localized) characteristics in space, time and frequency. This structure is enforced through an expansion of the source current density into appropriate spatio-temporal basis functions in combination with sparsity constraints. This approach combines the main strengths of two existing methods, namely Sparse Basis Field Expansions (Haufe et al., 2011) and Time-Frequency Mixed-Norm Estimates (Gramfort et al., 2013). By adjusting the ratio between two regularization terms, STOUT is capable of trading temporal for spatial reconstruction accuracy and vice versa, depending on the requirements of specific analyses and the provided data. Due to allowing for non-stationary source activations, STOUT is particularly suited for the localization of event-related potentials (ERP) and other evoked brain activity. We demonstrate its performance on simulated ERP data for varying signal-to-noise ratios and numbers of active sources. Our analysis of the generators of visual and auditory evoked N200 potentials reveals that the most active sources originate in the temporal and occipital lobes, in line with the literature on sensory processing.