Exploring Neural Mechanisms of Reward Processing Using Coupled Matrix Tensor Factorization: A Simultaneous EEG-fMRI...

Area of Science:

• Neuroscience
• Cognitive Science
• Neuroimaging

Background:

• Understanding neural feedback and cognitive decision-making in reward processing is crucial.
• Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) offer a powerful approach for studying brain function.
• This study pioneers the use of tensor decomposition in combined EEG-fMRI for reward processing research.

Purpose of the Study:

• To investigate neural mechanisms underlying reward processing during a gambling task.
• To explore cognitive decision-making processes associated with wins and losses.
• To demonstrate the utility of coupled matrix tensor factorization (CMTF) in analyzing complex neuroimaging data.

Main Methods:

• EEG data represented as third-order spectrogram tensors for frequency feature extraction.
• Joint decomposition of EEG and fMRI data using coupled matrix tensor factorization (CMTF).
• Graph-structured clustering applied for optimal model selection based on quantitative indices.

Main Results:

• Identified activation in expected regions of interest (ROIs) and novel areas like the olfactory cortex and fusiform gyrus.
• Orbitofrontal cortex and insula showed activation for both winning and losing stimuli.
• Distinct patterns of activation observed in regions like the superior orbital frontal gyrus, anterior cingulate cortex, inferior frontal gyrus, and medial superior frontal gyrus for winning versus losing stimuli.

Conclusions:

• The study provides novel insights into reward processing pathways in the brain.
• It deepens the understanding of brain function related to decision-making under uncertainty.
• This research establishes a new methodological approach for investigating neurovascular coupling using CMTF.