1 February 2022 in Computer Methods and Programs in Biomedicine
The study of brain function has recently expanded from classical univariate to multivariate analyses. These multivariate, machine learning-based algorithms afford neuroscientists extracting more detailed and richer information from the data. However, the implementation of these procedures is usually challenging, especially for researchers with no coding experience. To address this problem, we have developed MVPAlab, a MATLAB-based, flexible decoding toolbox for multidimensional electroencephalography and magnetoencephalography data. The MVPAlab Toolbox implements several machine learning algorithms to compute multivariate pattern analyses, cross-classification, temporal generalization matrices and feature and frequency contribution analyses. It also provides access to an extensive set of preprocessing routines for, among others, data normalization, data smoothing, dimensionality reduction and supertrial generation. To draw statistical inferences at the group level, MVPAlab includes a non-parametric cluster-based permutation approach. A sample electroencephalography dataset was compiled to test all the MVPAlab main functionalities. Significant clusters (p<0.01) were found for the proposed decoding analyses and different configurations, proving the software capability for discriminating between different experimental conditions. This toolbox has been designed to include an easy-to-use and intuitive graphic user interface and data representation software, which makes MVPAlab a very convenient tool for users with few or no previous coding experience. In addition, MVPAlab is not for beginners only, as it implements several high and low-level routines allowing more experienced users to design their own projects in a highly flexible manner.
4 June 2020 in International Journal of Neural Systems
27 April 2026 in European Journal of Neuroscience
RFRoberto Fernandes‐Magalhaes
Roberto Fernandes‐Magalhaes
+3
Fibromyalgia involves widespread musculoskeletal pain and hypersensitivity, often accompanied by neurological, cognitive, and affective disturbances. Resting‐state electroencephalography studies have revealed abnormal brain activity in chronic pain conditions, with anxiety and symptom duration potentially exacerbating these alterations. This study applied multivariate pattern analysis to differentiate intertrial resting‐state electroencephalography signals between fibromyalgia patients and healthy controls across frequency bands associated with pain processing, incorporating state and trait anxiety scores. It also examined differences between patients with short‐ and long‐duration symptoms and identified the most relevant scalp regions contributing to the models. Fifty‐one female participants (25 fibromyalgia patients, 26 controls; aged 35–65) were included. Patients were classified into short‐term (12) and long‐term (13) groups. Normalized power spectral density values were extracted from electroencephalography data and used to train machine learning classifiers, with Haufe‐transformed weights computed to determine key scalp contributions. The models distinguished patients from controls with area under the curve values exceeding 0.75 across all frequency bands, reaching 0.99 in beta and gamma bands when anxiety was included. Symptom duration was also a relevant factor, as the model differentiated short‐ from long‐term fibromyalgia patients with area under the curve values up to 0.96 in beta and gamma bands. Alterations in theta power within frontal and parietal regions, along with frequency‐specific contributions, highlight disrupted pain processing in fibromyalgia and suggest cumulative effects of prolonged symptom duration. Future resting‐state studies leveraging multivariate pattern analysis may support the development of potential biomarkers to improve diagnosis and guide treatment strategies in clinical settings.
15 September 2022 in bioRxiv
Summary: Proactive cognition brain models are mainstream nowadays. Within these, preparation is understood as an endogenous, top-down function that takes place prior to the actual perception of a stimulus and improves subsequent behavior. Neuroimaging has shown the existence of such preparatory activity separately in different cognitive domains, however no research to date has sought to uncover their potential similarities and differences. Two of these, often confounded in the literature, are Selective Attention (information relevance) and Perceptual Expectation (information probability). We used EEG to characterize the mechanisms that pre-activate specific contents in Attention and Expectation. In different blocks, participants were cued to the relevance or to the probability of target categories, faces vs. names, in a gender discrimination task. Multivariate Pattern (MVPA) and Representational Similarity Analyses (RSA) during the preparation window showed that both manipulations led to a significant, ramping-up prediction of the relevant or expected target category. However, classifiers trained on data from one condition did not generalize to the other, indicating the existence of unique anticipatory neural patterns. In addition, a Canonical Template Tracking procedure showed that there was stronger anticipatory perceptual reinstatement for relevance than for expectation blocks. Overall, results indicate that preparation during attention and expectation acts through distinguishable neural mechanisms. These findings have important implications for current models of brain functioning, as they are a first step towards characterizing and dissociating the neural mechanisms involved in top-down anticipatory processing.