fMRI Study of Cognitive Control in the Flanker Task

We explore brain responses to cognitive conflicts using functional MRI (fMRI) data from the Eriksen flanker task, trying to locate areas that exhibit significant differences in activation between the congruent and incongruent conditions.

Project Overview

1. Data Quality Control
2. Preprocessing
3. First-level Analysis
4. Second-level Analysis
5. Third-level Group Analysis
6. ROI Analysis

1. Data Quality Control

• Checked T1 structural images for artifacts.
• Visually inspected functional data for motion artifacts.
• Categorized runs: No visible motion (34), slight motion (11), excessive motion (7).

2. Preprocessing

Skull Stripping

Removed non-brain tissue from structural and functional images using a threshold of 0.2.

Motion Correction

Used MCFLIRT algorithm to correct for subject head motion.

Smoothing

Applied 5mm Gaussian smoothing kernel to reduce noise and increase signal-to-noise ratio.

Registration

Aligned functional data to MNI152 2mm template for group analysis.

3. First-Level Analysis

• Created design matrix with regressors for congruent and incongruent trials.

• Fit General Linear Model (GLM) to get parameter estimates and contrasts. The original data (time series at one voxel) vs fitted model.

• Produced thresholded statistical maps (Z > 3.1). The significant brain activations after thresholding for COPE1 (incongruent)

4. Second-Level Analysis

• Combined data from the two runs for each subject using fixed effects model.

Sample results for COPE3 (incongruent - congruent activations) after the 2nd level analysis.

5. Third-Level Group Analysis

• Combined data across all subjects using mixed effects model (FLAME 1)
• Applied cluster-based thresholding for group activation maps. Sample results for COPE3 (incongruent - congruent activations) after the 3rd level analysis.

6. ROI Analysis

Examined activations in specific regions:

• Lateral occipital cortex
• Paracingulate gyrus
• Insular cortex

Used both anatomical masks and 5mm spherical ROIs centered on activation peaks. Spherical Mask around the Paracingulate Gyrus Anatomical Mask of the Paracingulate Gyrus

Key Findings

• Greater activation for incongruent vs congruent trials, particularly in:
  • Lateral occipital cortex: involved in object recognition and visual processing.
  • Paracingulate gyrus: associated with conflict management.
  • Insular cortex: part of the salience network, involved in detecting relevant stimuli.

These results are consistent with regions known to be involved in conflict processing and visual attention.

Important Note

For a step-by-step walkthrough, detailed results and explanations, check the fMRI-Study-of-Cognitive-Control-in-the-Flanker-Task.pdf in the repository .

The Data Analysis Toolkit Used

• FSL (FMRIB Software Library) for fMRI data processing and analysis.

Acknowledgments

This project was completed as part of the Neuroimaging course supervised by Dr. Meena M. Makary at Cairo University.

The project and scripts were inspired by and builds upon the knowledge gained from the "Andy's Brain Book fMRI Short Course" by Andrew Jahn

Course Link: Andy's Brain Book fMRI Short Course