99 research outputs found
Advances in automating analysis of neural time series data
No full textLes expériences d’électrophysiologie ont longtemps reposé sur de petites cohortes de sujets pour découvrir des effets d’intérêt significatifs. Toutefois, la faible taille de l’échantillon se traduit par une faible puissance statistique, ce qui entraîne un taux élevé de fausses découvertes et un faible taux de reproductibilité. Deux questions restent à répondre : 1) comment faciliter le partage et la réutilisation des données pour créer de grands ensembles de données; et 2) une fois que de grands ensembles de données sont disponibles, quels outils pouvons-nous construire pour les analyser ? Donc, nous introduisons une nouvelle norme pour le partage des données, Brain Imaging Data Structure (BIDS), et son extension MEG-BIDS. Puis, nous présentons un pipeline d’analyse de données électrophysiologie avec le logiciel MNE. Nous tenons compte des différents choix que l’utilisateur doit faire à chaque étape et formulons des recommandations standardisées. De plus, nous proposons un outil automatisé pour supprimer les segments de données corrompus par des artefacts, ainsi qu’un algorithme de détection d’anomalies basé sur le réglage des seuils de rejet. Par ailleurs, nous utilisons les données HCP, annotées manuellement, pour comparer notre algorithme aux méthodes existantes. Enfin, nous utilisons le convolutional sparse coding pour identifier les structures des séries temporelles neuronales. Nous reformulons l’approche existante comme une inférence MAP pour être atténuer les artefacts provenant des grandes amplitudes et des distributions à queue lourde. Ainsi, cette thèse tente de passer des méthodes d’analyse lentes et manuelles vers des méthodes automatisées et reproducibles.Electrophysiology experiments has for long relied upon small cohorts of subjects to uncover statistically significant effects of interest. However, the low sample size translates into a low power which leads to a high false discovery rate, and hence a low rate of reproducibility. To address this issue means solving two related problems: first, how do we facilitate data sharing and reusability to build large datasets; and second, once big datasets are available, what tools can we build to analyze them ? In the first part of the thesis, we introduce a new data standard for sharing data known as the Brain Imaging Data Structure (BIDS), and its extension MEG-BIDS. Next, we introduce the reader to a typical electrophysiological pipeline analyzed with the MNE software package. We consider the different choices that users have to deal with at each stage of the pipeline and provide standard recommendations. Next, we focus our attention on tools to automate analysis of large datasets. We propose an automated tool to remove segments of data corrupted by artifacts. We develop an outlier detection algorithm based on tuning rejection thresholds. More importantly, we use the HCP data, which is manually annotated, to benchmark our algorithm against existing state-of-the-art methods. Finally, we use convolutional sparse coding to uncover structures in neural time series. We reformulate the existing approach in computer vision as a maximuma posteriori (MAP) inference problem to deal with heavy tailed distributions and high amplitude artifacts. Taken together, this thesis represents an attempt to shift from slow and manual methods of analysis to automated, reproducible analysis
Contributions pour l'analyse automatique de signaux neuronaux
No full textElectrophysiology experiments has for long relied upon small cohorts of subjects to uncover statistically significant effects of interest. However, the low sample size translates into a low power which leads to a high false discovery rate, and hence a low rate of reproducibility. To address this issue means solving two related problems: first, how do we facilitate data sharing and reusability to build large datasets; and second, once big datasets are available, what tools can we build to analyze them ? In the first part of the thesis, we introduce a new data standard for sharing data known as the Brain Imaging Data Structure (BIDS), and its extension MEG-BIDS. Next, we introduce the reader to a typical electrophysiological pipeline analyzed with the MNE software package. We consider the different choices that users have to deal with at each stage of the pipeline and provide standard recommendations. Next, we focus our attention on tools to automate analysis of large datasets. We propose an automated tool to remove segments of data corrupted by artifacts. We develop an outlier detection algorithm based on tuning rejection thresholds. More importantly, we use the HCP data, which is manually annotated, to benchmark our algorithm against existing state-of-the-art methods. Finally, we use convolutional sparse coding to uncover structures in neural time series. We reformulate the existing approach in computer vision as a maximuma posteriori (MAP) inference problem to deal with heavy tailed distributions and high amplitude artifacts. Taken together, this thesis represents an attempt to shift from slow and manual methods of analysis to automated, reproducible analysis.Les expériences d’électrophysiologie ont longtemps reposé sur de petites cohortes de sujets pour découvrir des effets d’intérêt significatifs. Toutefois, la faible taille de l’échantillon se traduit par une faible puissance statistique, ce qui entraîne un taux élevé de fausses découvertes et un faible taux de reproductibilité. Deux questions restent à répondre : 1) comment faciliter le partage et la réutilisation des données pour créer de grands ensembles de données; et 2) une fois que de grands ensembles de données sont disponibles, quels outils pouvons-nous construire pour les analyser ? Donc, nous introduisons une nouvelle norme pour le partage des données, Brain Imaging Data Structure (BIDS), et son extension MEG-BIDS. Puis, nous présentons un pipeline d’analyse de données électrophysiologie avec le logiciel MNE. Nous tenons compte des différents choix que l’utilisateur doit faire à chaque étape et formulons des recommandations standardisées. De plus, nous proposons un outil automatisé pour supprimer les segments de données corrompus par des artefacts, ainsi qu’un algorithme de détection d’anomalies basé sur le réglage des seuils de rejet. Par ailleurs, nous utilisons les données HCP, annotées manuellement, pour comparer notre algorithme aux méthodes existantes. Enfin, nous utilisons le convolutional sparse coding pour identifier les structures des séries temporelles neuronales. Nous reformulons l’approche existante comme une inférence MAP pour être atténuer les artefacts provenant des grandes amplitudes et des distributions à queue lourde. Ainsi, cette thèse tente de passer des méthodes d’analyse lentes et manuelles vers des méthodes automatisées et reproducibles
Real-time machine learning of MEG: Decoding signatures of selective attention
No full textBrain--computer interfaces (BCIs) provide disabled patients with access to communication tools and control of prosthetic devices. Most BCIs employ a machine-learning algorithm which analyzes brain data in real time and provides users with feedback.
Magnetoencephalography (MEG) is a non-invasive method which records neuromagnetic signals from the brain at a high temporal resolution. This makes it particularly suitable for real-time analysis and machine learning. Developing tools that allow such analysis will have long-term benefits in using MEG for BCI approaches and exploring new experimental paradigms.
In this thesis, a real-time analysis pipeline for machine learning in MEG was developed with the goal to enable BCI in MEG systems. The implementation details of the pipeline were described in the thesis along with performance details. Additionally, pilot measurements to decode auditory attention were conducted. The spatio-temporal dynamics of the offline experiment were used to optimize the preprocessing steps required for the BCI application. In particular, the frequency range of 1.0--1.5 Hz was found to be particularly discriminative. Finally, simulating this pipeline in pseudo real-time mode demonstrated that a BCI to decode auditory attention is feasible in MEG
Author Correction to: Extreme Learning Machine Framework for Risk Stratification of Fatty Liver Disease Using Ultrasound Tissue Characterization
No full textThe original version of this article unfortunately contained a mistake. The family name of Rui Tato Marinho was incorrectly spelled as Marinhoe
Target of selective auditory attention can be robustly followed with MEG
No full textAbstract Selective auditory attention enables filtering of relevant acoustic information from irrelevant. Specific auditory responses, measurable by magneto- and electroencephalography (MEG/EEG), are known to be modulated by attention to the evoking stimuli. However, such attention effects have typically been studied in unnatural conditions (e.g. during dichotic listening of pure tones) and have been demonstrated mostly in averaged auditory evoked responses. To test how reliably we can detect the attention target from unaveraged brain responses, we recorded MEG data from 15 healthy subjects that were presented with two human speakers uttering continuously the words “Yes” and “No” in an interleaved manner. The subjects were asked to attend to one speaker. To investigate which temporal and spatial aspects of the responses carry the most information about the target of auditory attention, we performed spatially and temporally resolved classification of the unaveraged MEG responses using a support vector machine. Sensor-level decoding of the responses to attended vs. unattended words resulted in a mean accuracy of 79 % ± 2 % (N = 14) for both stimulus words. The discriminating information was mostly available 200–400 ms after the stimulus onset. Spatially-resolved source-level decoding indicated that the most informative sources were in the auditory cortices, in both the left and right hemisphere. Our result corroborates attention modulation of auditory evoked responses and shows that such modulations are detectable in unaveraged MEG responses at high accuracy, which could be exploited e.g. in an intuitive brain–computer interface
- …
