Abstract
BackgroundDeep learning has revolutionized the field of computer vision, where convolutional neural networks (CNNs) extract complex patterns of information from large datasets. The use of deep networks in neuroscience is mainly focused to neuroimaging or brain computer interface -BCI- applications. In electroencephalography (EEG) research, multivariate pattern analysis (MVPA) mainly relies on linear algorithms, which require a homogeneous dataset and assume that discriminant features appear at consistent latencies and electrodes across trials. However, neural responses may shift in time or space during an experiment, resulting in under-estimation of discriminant features. Here, we aimed at using CNNs to classify EEG responses to external stimuli, by taking advantage of time- and space- unlocked neural activity, and at examining how discriminant features change over the course of an experiment, on a trial by trial basis. New methodWe present a novel pipeline, consisting of data augmentation, CNN training, and feature visualization techniques, fine-tuned for MVPA on EEG data. ResultsOur pipeline provides high classification performance and generalizes to new datasets. Additionally, we show that the features identified by the CNN for classification are electrophysiologically interpretable and can be reconstructed at the single-trial level to study trial-by-trial evolution of class-specific discriminant activity. Comparison with existing techniquesThe developed pipeline was compared to commonly used MVPA algorithms like logistic regression and support vector machines, as well as to shallow and deep convolutional neural networks. Our approach yielded significantly higher classification performance than existing MVPA techniques (p = 0.006) and comparable results to other CNNs for EEG data. ConclusionIn summary, we present a novel deep learning pipeline for MVPA of EEG data, that can extract trial-by-trial discriminative activity in a data-driven way.
Highlights
Multivariate pattern analysis (MVPA) is commonly used in the field of neuroscience to extract discriminative patterns of neural responses to external stimuli (Haynes and Rees, 2006)
Initially developed for functional magnetic resonance imaging, multivariate pattern analysis (MVPA) techniques have been adapted for the field of magneto- and electro-encephalography (M/EEG) (Grootswagers et al, 2017)
The high classification performance in the test set suggests that the trained networks could extract discriminant features of EEG responses to Repeated vs. Novel sounds, and generalize to Binary cross-entropy loss
Summary
Multivariate pattern analysis (MVPA) is commonly used in the field of neuroscience to extract discriminative patterns of neural responses to external stimuli (Haynes and Rees, 2006). Initially developed for functional magnetic resonance imaging (fMRI), MVPA techniques have been adapted for the field of magneto- and electro-encephalography (M/EEG) (Grootswagers et al, 2017) These are most commonly based on linear classifiers, which are applied on sensor-level topographic data, either aggregated across time (Tzovara et al, 2012) or on a time-point by time-point basis (King and Dehaene, 2014). This latter approach is most commonly implemented by training and testing one classifier at a given time-point within a trial (Castegnetti et al, 2020; Demarchi et al, 2019) and identifying time-points for which classification is above chance levels. Conclusion: In summary, we present a novel deep learning pipeline for MVPA of EEG data, that can extract trialby-trial discriminative activity in a data-driven way
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have