Dynamic changes in the mental rotation network revealed by pattern recognition analysis of fMRI data

Area of Science:

• Neuroscience
• Cognitive Neuroscience
• Brain Imaging

Background:

• Investigating dynamic changes in brain networks during complex cognitive tasks is crucial for understanding brain function.
• Traditional fMRI analysis often requires predefined models of task components and hemodynamic responses, limiting the study of dynamic processes.
• Spatio-temporal Support Vector Machines (SVMs) offer a data-driven approach to analyze dynamic brain activity without task-specific models.

Purpose of the Study:

• To apply spatio-temporal SVMs to investigate the temporal dynamics and connectivity changes within the mental rotation network.
• To identify brain regions and their temporal activation patterns that discriminate between different mental rotation difficulties.
• To examine functional connectivity alterations between key brain areas during varying degrees of mental rotation.

Main Methods:

• Utilized spatio-temporal SVMs, a pattern recognition technique, to analyze functional Magnetic Resonance Imaging (fMRI) data.
• Trained SVMs on fMRI observations from two cognitive states (different angular disparities in mental rotation) within defined time windows.
• Generated spatio-temporal maps to identify discriminating voxels and time points, and assessed functional connectivity between identified regions.

Main Results:

• Spatio-temporal SVM successfully discriminated between varying angular disparities (0 vs. 20, 60, 100 degrees), with accuracy correlating to disparity.
• Frontal premotor regions showed earlier discrimination than superior parietal cortex; inferior parietal lobe discriminated earlier than superior parietal lobe.
• Identified a network with decreased BOLD response (posterior cingulate, frontal, superior temporal gyrus) during higher disparity, showing high discrimination; increased functional connectivity observed in activated regions.

Conclusions:

• Spatio-temporal SVM is effective for analyzing dynamic brain network changes during complex cognitive tasks like mental rotation.
• The study reveals a temporal hierarchy in brain region activation and discrimination during mental rotation, with distinct patterns for increasing task difficulty.
• Findings highlight dynamic changes in functional connectivity within the mental rotation network, differentiating between regions with increased and decreased BOLD responses.