Duarte, J. V. | bioRxiv Science

bioRxiv2022

A two-stage framework for neural processing of biological motion

It remains to be understood how biological motion is hierarchically computed, from discrimination of local "life motion" animacy to global dynamic body perception. Here, we addressed this functional separation of the correlates of the perception of life motion, defined as characteristic for the local motion of parts of living beings, from perception of global motion of a body. We hypothesized that life motion processing can be isolated, by using a single dot motion perceptual decision paradigm featuring the biomechanical details of local realistic motion of a single joint. To ensure that we were indeed tackling processing of biological motion properties we used a discrimination instead of detection task. We discovered using representation similarity analysis that two key early dorsal and two ventral stream regions (visual motion selective hMT+ and V3A, extrastriate body area EBA and a region within fusiform gyrus FFG) showed robust and separable signals related to encoding of life motion and global motion. These signals reflected two independent processing stages, as revealed by representation dissimilarity analysis and deconvolution of fMRI responses to each motion pattern. This study showed that higher level pSTS encodes both classes of biological motion in a similar way, revealing a higher-level integrative stage, reflecting scale independent biological motion perception. Our results reveal a two-stage framework for neural computation of biological motion, with an independent contribution of dorsal and ventral regions for the initial stage.

neuroscience↗

bioRxiv2021

Optimizing EEG source reconstruction with concurrent fMRI-derived spatial priors

Reconstructing EEG sources involves a complex pipeline, with the inverse problem being the most challenging. Multiple inversion algorithms are being continuously developed, aiming to tackle the non-uniqueness of this problem, which has been shown to be partially circumvented by also including fMRI-derived spatial priors in the inverse models. However, only task activation maps and resting-state networks (RSNs) have been explored so far, overlooking the recent, but already accepted, notion that brain networks exhibit dynamic functional connectivity (dFC) fluctuations. Moreover, there is no consensus regarding the inversion algorithm of choice, nor a systematic comparison between different sets of spatial priors. Using simultaneous EEG-fMRI data, here we compared four different inversion algorithms (MN, LORETA, EBB and MSP) under a Bayesian framework, each with three different sets of priors consisting of: 1) those specific to the algorithm (S1); 2) S1 plus fMRI task activation maps and RSNs (S2); and 3) S2 plus network modules of task-related dFC states estimated from the dFC fluctuations (S3). The quality of the reconstructed EEG sources was quantified in terms of model-based metrics, namely the free-energy and variance explained of the inversion models, and the overlap/proportion of brain regions known to be involved in the visual perception tasks that the participants were submitted to, and RSN templates, with/within EEG source components. Model-based metrics suggested that model parsimony is preferred, with the combination MSP+S1 exhibiting the best performance. However, optimal overlap/proportion values were found using EBB+S2 or MSP+S3, respectively, indicating that fMRI spatial priors, including dFC state modules, are crucial for the EEG source components to reflect neuronal activity of interest. Our results pave the way towards a more informative selection of the optimal EEG source reconstruction approach, which may be crucial in future studies.

neuroscience↗