Computational analysis of longitudinal electroencephalograms after theta burst stimulation over left DLPFC using hierarchical dynamic causal modelling.

Steps:

Data Preprocessing

1. Specifying rest eeg trials (remove TEP epochs)
2. BandPass (1 - 50) and downsample to 258 (see tbs_rseeg_preprocess)
3. 2 second epoch for each trial (trial 1: Res Pre 1, trial 2: Rest pre 2, trial 3: Rest post 1, trial 4: Rest post 2, trial 5: Rest post 3)
4. Auto rejection of badepochs (see tbs_rseeg_trialrejection function)
5. Manual rejection of trials and epochs (see tbs_rseeg_cleaning)
6. ICA using runica
7. inspection of ICA (Auto labeling with manual rejection)
8. Rerefrence to average
9. Final inspection
10. change fieldtrip to spm (save spm object)

Data Analysis

1. Loading data to SPM using fieldtrip conversion
2. Source localization: using individual MRI and real-time chanel locs
3. Model specification: spectral DCM, Conductance-based Canonical Microcircuit Model (cmm-NMDA)
4. Model estimation
5. Explained Variance: More than 95 is desirable
6. Model Selection: Using model variation and selecting the best model with free energy criteria
7. First Level Parametric empirical bayes
8. Peb of Peb: The influence of TMS protocols on connectivity parameters
9. Cross-Validation

---

IDS team 23: Supervisor: Prof. Ali Motie Nasrabadi Mentor: Armin Toghi Members: Ghazale ghaffaripour, Hamed moghtaderi, Babak Aliyari