Differentiable neural architecture search for optimal spatial/temporal brain function network decomposition

2021 ◽  
Vol 69 ◽  
pp. 101974
Author(s):  
Qing Li ◽  
Xia Wu ◽  
Tianming Liu
Keyword(s):  
Brain Function ◽  
Network Decomposition ◽  
Neural Architecture ◽  
Function Network

scholarly journals Hierarchical Spatio-Temporal Modeling of Naturalistic Functional Magnetic Resonance Imaging Signals via Two-Stage Deep Belief Network With Neural Architecture Search

2021 ◽  
Vol 15 ◽  
Author(s):  
Yudan Ren ◽  
Shuhan Xu ◽  
Zeyang Tao ◽  
Limei Song ◽  
Xiaowei He
Keyword(s):  
Magnetic Resonance Imaging ◽  
Deep Learning ◽  
Magnetic Resonance ◽  
Brain Function ◽  
Learning Models ◽  
Functional Magnetic Resonance ◽  
Resonance Imaging ◽  
Two Stage ◽  
Functional Activities ◽  
Neural Architecture

Naturalistic functional magnetic resonance imaging (NfMRI) has become an effective tool to study brain functional activities in real-life context, which reduces the anxiety or boredom due to difficult or repetitive tasks and avoids the problem of unreliable collection of brain activity caused by the subjects’ microsleeps during resting state. Recent studies have made efforts on characterizing the brain’s hierarchical organizations from fMRI data by various deep learning models. However, most of those models have ignored the properties of group-wise consistency and inter-subject difference in brain function under naturalistic paradigm. Another critical issue is how to determine the optimal neural architecture of deep learning models, as manual design of neural architecture is time-consuming and less reliable. To tackle these problems, we proposed a two-stage deep belief network (DBN) with neural architecture search (NAS) combined framework (two-stage NAS-DBN) to model both the group-consistent and individual-specific naturalistic functional brain networks (FBNs), which reflected the hierarchical organization of brain function and the nature of brain functional activities under naturalistic paradigm. Moreover, the test-retest reliability and spatial overlap rate of the FBNs identified by our model reveal better performance than that of widely used traditional methods. In general, our model provides a promising method for characterizing hierarchical spatiotemporal features under the natural paradigm.

scholarly journals The neural architecture of language: Integrative reverse-engineering converges on a model for predictive processing

2020 ◽  
Author(s):  
Martin Schrimpf ◽  
Idan Blank ◽  
Greta Tuckute ◽  
Carina Kauf ◽  
Eghbal A. Hosseini ◽  
...  
Keyword(s):  
Reverse Engineering ◽  
Language Processing ◽  
Brain Function ◽  
Computational Models ◽  
Predictive Processing ◽  
Neural Responses ◽  
Data Types ◽  
Neural Architecture ◽  
And Behavior ◽  
The Brain

AbstractThe neuroscience of perception has recently been revolutionized with an integrative reverse-engineering approach in which computation, brain function, and behavior are linked across many different datasets and many computational models. We here present a first systematic study taking this approach into higher-level cognition: human language processing, our species’ signature cognitive skill. We find that the most powerful ‘transformer’ networks predict neural responses at nearly 100% and generalize across different datasets and data types (fMRI, ECoG). Across models, significant correlations are observed among all three metrics of performance: neural fit, fit to behavioral responses, and accuracy on the next-word prediction task (but not other language tasks), consistent with the long-standing hypothesis that the brain’s language system is optimized for predictive processing. Model architectures with initial weights further perform surprisingly similar to final trained models, suggesting that inherent structure – and not just experience with language – crucially contributes to a model’s match to the brain.

scholarly journals Processing of Affective Pictures: A Study Based on Functional Connectivity Network in the Cerebral Cortex

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Zhongyang He ◽  
Kai Yang ◽  
Ning Zhuang ◽  
Ying Zeng
Keyword(s):  
Cerebral Cortex ◽  
Brain Function ◽  
Middle Temporal Gyrus ◽  
Emotional States ◽  
Eeg Signals ◽  
Brain Areas ◽  
Middle Temporal ◽  
Network Connections ◽  
The Difference ◽  
Function Network

Emotion plays an important role in people’s life. However, the existing researches do not give a unified conclusion on the brain function network under different emotional states. In this study, pictures from the international affective picture system (IAPS) of different valences were presented to subjects with a fixed frequency blinking frequency to induce stable state visual evoked potential (SSVEP). With the source location method, the cerebral cortex source signal was reconstructed based on EEG signals, and then the difference in SSVEP amplitudes in key brain areas under different emotional states and the difference in brain function network connections among different brain areas were analysed in cortical space. The results of the study show that positive and negative emotions evoked greater activation intensities in the prefrontal, temporal, and parietal lobes compared with those of neutral emotion. The network connections with a significant difference between emotional states mainly appear in the alpha and gamma bands, and the network connections with significant differences between positive emotion and negative emotion mainly exist in the right middle temporal gyrus and the superior frontal gyrus on both sides. In addition, the long-range connections play an important role in the process of emotional processing, especially the connections among frontal gyrus, middle temporal gyrus, and middle occipital gyrus. The results of this study provide a reliable scientific basis for revealing and elucidating the neural mechanism of emotion processing and the selection of brain regions and frequency bands in emotion recognition based on EEG signals.

Musical Disorders

2019 ◽  
pp. 759-775
Author(s):  
Isabelle Royal ◽  
Sébastien Paquette ◽  
Pauline Tranchant
Keyword(s):  
Intellectual Disability ◽  
Human Brain ◽  
Early Development ◽  
Brain Function ◽  
Daily Basis ◽  
Specific Role ◽  
Brain Areas ◽  
Neural Architecture ◽  
Auditory Dysfunction ◽  
Neurological Event

Nearly everyone is exposed to music on a daily basis and the human brain is equipped with the necessary neural architecture to naturally acquire musical abilities during early development. Despite the universality of music, a minority of individuals present with very specific musical deficits that cannot be attributed to a general auditory dysfunction, intellectual disability, or a lack of musical exposure. These musical deficiencies can either be present from birth (congenital amusia, beat finding disorder) or acquired following a neurological event (acquired amusia). The purpose of the present chapter is to provide an overview of these intriguing musical disorders, highlight their common and different underlying features, and to demonstrate how they represent a unique opportunity to study brain function and to isolate brain areas that play a specific role in musical processing.

Research on Modeling of Brain Function Network in Major Depressive Disorder

2014 ◽  
Vol 989-994 ◽  
pp. 2037-2042
Author(s):  
Li Min Niu ◽  
Hao Guo ◽  
Jun Jie Chen
Keyword(s):  
Major Depressive Disorder ◽  
Depressive Disorder ◽  
Brain Function ◽  
Evaluation Criteria ◽  
Real Data ◽  
Class Model ◽  
Major Depressive ◽  
Common Neighbor ◽  
Model Based ◽  
Function Network

In order to analyze the gap of function network between Major depressive disorder and health person, this paper studies with modeling approach. This paper analyzes the function network of Major depressive disorder with the model based on anatomical distance and the number of common neighbor. The result shows that the distribution of the optimal brain function network is linear in all volunteer. And the slope of the linear relationship in the patients is less than health, so we hope this point can be as secondary evidence to determine the person whether fall ill. And we also propose two models and those models of brain function are based on anatomical distance or the number of common neighbor. Create the evaluation criteria for select the optimal brain function model network in each class model based on select the maximum value in the proportion of the common edges of two network accounted all edges. Select the model that can simulate the real brain function network by comparison with real data fMRI network. Finally, the results show the best model only is based on anatomical distance .

scholarly journals Voxelwise-based Brain Function Network using Multi-Graph Model

2018 ◽  
Vol 8 (1) ◽  
Author(s):  
Zhongyang Wang ◽  
Junchang Xin ◽  
Xinlei Wang ◽  
Zhiqiong Wang ◽  
Yue Zhao ◽  
...  
Keyword(s):  
Brain Function ◽  
Graph Model ◽  
Function Network
Sign in / Sign up

Export Citation Format

Share Document