Age Estimation from fMRI Data Using Recurrent Neural Network

Finding a biomarker that indicates the subject’s age is one of the most important topics in biology. Several recent studies tried to extract a biomarker from brain imaging data including fMRI data. However, most of them focused on MRI data, which do not provide dynamics and lack attempts to apply recently proposed deep learning models. We propose a deep neural network model that estimates the age of a subject from fMRI images using a recurrent neural network (RNN), more precisely, a gated recurrent unit (GRU). However, applying neural networks is not trivial due to the high dimensional nature of fMRI data. In this work, we propose a novel preprocessing technique using the Automated Anatomical Labeling (AAL) atlas, which significantly reduces the input dimension. The proposed dimension reduction technique allows us to train our model with 640 training and validation samples from different projects under mean squared error (MSE). Finally, we obtain the correlation value of 0.905 between the predicted age and the actual age on 155 test samples. The proposed model estimates the age within the range of ±12 on most of the test samples. Our model is written in Python and is freely available for download.

Threat Vigilance and Intrinsic Amygdala Connectivity

A well documented amygdala-dorsomedial prefrontal circuit is theorized to promote attention to threat (‘threat vigilance’). Prior research has implicated a relationship between individual differences in trait anxiety/vigilance, engagement of this circuitry, and anxiogenic features of the environment (e.g. through threat-of-shock and movie-watching). In the present study, we predicted that—for those scoring high in self-reported anxiety and a behavioral measure of threat vigilance—this circuitry is chronically engaged, even in the absence of anxiogenic stimuli. Our analyses of resting-state fMRI data (N=639) did not, however, provide evidence for such a relationship. Nevertheless, in our planned exploratory analyses, we saw a relationship between threat vigilance behavior (but not self-reported anxiety) and intrinsic amygdala-periaqueductal gray connectivity. Here, we suggest this subcortical circuitry may be chronically engaged in hypervigilant individuals, but that the amygdala-prefrontal circuitry may only be engaged in response to anxiogenic stimuli.

Technologies for studying functional neural networks of the human brain based on data of nuclear functional magnetic tomography

Nuclear functional magnetic resonance imaging (fMRI) is one of the most popular methods for studying the functional activity of the human brain. In particular, this method is used in medicine to obtain information about the state of the functional networks of the patient’s brain. However, the process of processing and analysis of experimental fMRI data is complex and requires the selection of the correct technique, depending on the specific task. Practice has shown that different processing methods can give slightly different results for the same set of fMRI data. There are a number of alternative specialized software packages for processing and analysis, but the methodology still needs improvement and development. We are working in this direction: we analyze the effectiveness of existing methods; we develop our own methods; we create software services for processing and analysis of fMRI data on the basis of the distributed modular platform “Digital Laboratory”, with the involvement of the supercomputer NRC “Kurchatov Institute”. For research we use experimental fMRI data obtained on the scanner Siemens Verio Magnetom 3T at the Kurchatov Institute. One of our tasks within the framework of this project is to improve the technology for studying large-scale functional areas of the cerebral cortex at rest. To build a hierarchical model of interaction of large-scale neural networks, a verified binding of functional areas to anatomy is required. Today, there are a number of generally accepted atlases of the functional areas of the human cerebral cortex, which, nevertheless, are constantly being finalized and refined. This article presents the results of our study of the Glasser atlas for the consistency of voxels within one region and the connectivity metrics of voxel dynamics.

Embedding Functional Brain Networks in Low Dimensional Spaces Using Manifold Learning Techniques

Background: fMRI data is inherently high-dimensional and difficult to visualize. A recent trend has been to find spaces of lower dimensionality where functional brain networks can be projected onto manifolds as individual data points, leading to new ways to analyze and interpret the data. Here, we investigate the potential of two powerful non-linear manifold learning techniques for functional brain networks representation: (1) T-stochastic neighbor embedding (t-SNE) and (2) Uniform Manifold Approximation Projection (UMAP) a recent breakthrough in manifold learning.Methods: fMRI data from the Human Connectome Project (HCP) and an independent study of aging were used to generate functional brain networks. We used fMRI data collected during resting state data and during a working memory task. The relative performance of t-SNE and UMAP were investigated by projecting the networks from each study onto 2D manifolds. The levels of discrimination between different tasks and the preservation of the topology were evaluated using different metrics.Results: Both methods effectively discriminated the resting state from the memory task in the embedding space. UMAP discriminated with a higher classification accuracy. However, t-SNE appeared to better preserve the topology of the high-dimensional space. When networks from the HCP and aging studies were combined, the resting state and memory networks in general aligned correctly.Discussion: Our results suggest that UMAP, a more recent development in manifold learning, is an excellent tool to visualize functional brain networks. Despite dramatic differences in data collection and protocols, networks from different studies aligned correctly in the embedding space.

An examination of intolerance of uncertainty effects on subjective, psychophysiological, and neural markers during uninstructed and instructed threat acquisition and extinction training

Individuals who score high in self-reported Intolerance of Uncertainty (IU) tend to find uncertainty aversive. Prior research has demonstrated that under uncertainty individuals with high IU display difficulties in updating learned threat associations to safety associations. Importantly, recent research has shown that providing contingency instructions about threat and safety contingencies (i.e. reducing uncertainty) to individuals with high IU promotes the updating of learned threat associations to safety associations. Here we aimed to conceptually replicate IU and contingency instruction-based effects by conducting a secondary analysis of self-reported IU, ratings, skin conductance, and functional magnetic resonance imaging (fMRI) data recorded during uninstructed/instructed blocks of threat acquisition and threat extinction training (n = 48). Self-reported IU was not associated with differential responding to learned threat and safety cues for any measure during uninstructed/instructed blocks of threat acquisition and threat extinction training. There was some tentative evidence that higher IU was associated with greater ratings of unpleasantness and arousal to the safety cue after the experiment and greater skin conductance response to the safety cue during extinction generally. Potential explanations for these null effects and directions for future research are discussed.