Functional connectivity fingerprints of the FEF and IFJ

Citation

Soyuhos, O., & Baldauf, D. (2023). Functional connectivity fingerprints of the frontal eye field and inferior frontal junction suggest spatial versus nonspatial processing in the prefrontal cortex. European Journal of Neuroscience, 57(7), 1114–1140. https://doi.org/10.1111/ejn.15936

---

Repository Structure

rMEG-FEF-vs-IFJ/
├── Analysis_Pipeline/
│   ├── AnalysisPipeline_Part1.m        # Preprocessing entry point
│   ├── AnalysisPipeline_Part2.m        # Full connectivity analysis entry point
│   └── helper_fun/
│       ├── load_path.m                 # Directory configuration (edit before running)
│       ├── clean_data.m                # MEG artifact removal (ICA, bad channels/segments)
│       ├── scouts_timeSeries.m         # Source reconstruction via Brainstorm (dSPM2018)
│       ├── functional_connectivity.m   # Per-subject connectivity computation (oPEC/iCOH/dwPLI/PDC)
│       ├── results_connectivity.m      # Per-subject connectivity results, saves .mat and .xlsx
│       ├── group_average_topK_forSelectedROIs.m    # Group average for iCOH / dwPLI (seed-based)
│       └── group_average_topK_wholeConnectivity.m  # Group average for oPEC / PDC (whole-brain)
└── Visualize_Results/
    ├── fsaverage/                      # fsaverage surface templates (multiple smoothing levels)
    ├── results/                        # Group-level connectivity matrices (see Data Availability)
    └── scripts/
        ├── RUNME_visualizeResults_functionConn.m               # Functional connectivity visualization
        ├── RUNME_visualizeResults_functionConn_collapseTargets.m  # Lateralization-collapsed version
        ├── RUNME_visualizeResults_effectiveConn.m              # Effective (directional) connectivity
        ├── helper_saveFig_functionalConn.m                     # Batch figure saving (functional conn)
        ├── helper_saveFig_effectiveConn.m                      # Batch figure saving (effective conn)
        └── helper_fun/
            ├── load_path.m                              # Directory configuration (edit before running)
            ├── helper_functionConn_fsaverage.m          # fsaverage visualization for functional conn
            ├── helper_functionConn_fsaverage_freqCollapsed.m   # Frequency-collapsed version
            ├── helper_effectiveConn_fsaverage.m         # fsaverage visualization for effective conn
            ├── helper_functionConn_circularGraph.m      # Circular graph visualization
            ├── helper_save_figure.m                     # Figure saving utility
            ├── helper_violinplot.m                      # Violin plot visualization
            ├── tutorial_nwa_connectivityviewer.m        # 3D connectivity viewer (from FieldTrip)
            ├── cmap_rbw.m                               # Red-blue-white colormap (from Brainstorm)
            ├── fdr_bh.m                                 # FDR correction
            ├── circularGraph/                           # Circular graph toolbox (third-party)
            └── Violinplot-Matlab-master/                # Violin plot toolbox (third-party)

---

Dependencies

Analysis Pipeline (Part 1 — Preprocessing)

• MATLAB R2012b or later
• FieldTrip r10442
• megconnectome 3.0

Analysis Pipeline (Part 2 — Connectivity)

• MATLAB R2020a or later
• FieldTrip 20210411
• Brainstorm (version 28-May-2021 or later)

Visualization

• MATLAB R2020a or later
• FieldTrip 20210411
• Third-party MATLAB toolboxes (included in helper_fun/):
  • fdr_bh
  • circularGraph
  • Violinplot-Matlab

---

Setup

Before running any script, fill in the directory paths in the load_path.m files:

• Analysis Pipeline: Analysis_Pipeline/helper_fun/load_path.m
• Visualization: Visualize_Results/scripts/helper_fun/load_path.m

Each file contains clearly labeled variables for the locations of toolboxes, raw data, and output directories.

---

Usage

Step 1 — Preprocessing (AnalysisPipeline_Part1.m)

Cleans raw resting-state MEG recordings from the HCP dataset and segments them into epochs.

1. Open Analysis_Pipeline/AnalysisPipeline_Part1.m
2. Fill in subjectids with your subject IDs (e.g., {'111514', '140117'})
3. Set trialDuration to 2, 5, or 10 (seconds)
4. Run the script

Output: cleaned, epoched data saved per subject under cleandata/.

Step 2 — Connectivity Analysis (AnalysisPipeline_Part2.m)

End-to-end seed-based functional connectivity analysis at the group level.

1. Open Analysis_Pipeline/AnalysisPipeline_Part2.m
2. Fill in subjectids
3. Select analysis options:
   • trialDuration: 2, 5, or 10
   • connMetric: 'oPEC', 'iCOH', 'dwPLI', or 'PDC'
4. Run the script

The pipeline runs three steps automatically:

Step	Function	Description
2	scouts_timeSeries	Source reconstruction (Brainstorm + dSPM2018 inverse, Glasser atlas)
3	functional_connectivity	Computes connectivity per subject per frequency band
4	results_connectivity / group_average	Aggregates and saves group-level connectivity results

Seed regions: L/R FEF, L/R IFJa, L/R IFJp (Glasser atlas)
Frequency bands: delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), gamma (30–100 Hz)

Ground-truth analysis — oPEC (orthogonalized Power Envelope Correlation):

Step 3 — Visualization (Visualize_Results/scripts/)

Applies statistical masking (Wilcoxon signed-rank test + FDR correction) and renders results on the fsaverage surface.

Script	Description
RUNME_visualizeResults_functionConn.m	Functional connectivity per hemisphere (fsaverage + circular graph)
RUNME_visualizeResults_functionConn_collapseTargets.m	Same, collapsed across lateralization
RUNME_visualizeResults_effectiveConn.m	Directional connectivity (PDC); green = seed→target, magenta = target→seed

Configurable options in each script: band_name, seed_target, conn_metric, statistics.alpha, FDR correction, and figure export.

---

Data Availability

The MEG data used in this study are publicly available through the Human Connectome Project (HCP 1200 Subjects Release).

---

License

This repository is licensed under the GNU General Public License v3.0. See LICENSE for details.

Portions of the code are adapted from:

• megconnectome 3.0 — Copyright (C) 2011–2014 Human Connectome Project, WU-Minn Consortium
• FieldTrip — tutorial_nwa_connectivityviewer.m
• Brainstorm — cmap_rbw.m