Physician & Neuroimaging Researcher

I build reproducible neuroimaging workflows across structural MRI, task and resting-state fMRI, diffusion MRI, MEG, EEG, and PET.

My work emphasizes transparent processing, quality control, cross-software validation, open science, and technically documented analysis pipelines.

Reproducible single-subject diffusion MRI workflow using public MPI-LEMON data.

• MP-PCA denoising and Gibbs-ringing correction
• Motion, eddy-current, and outlier correction with FSL Eddy
• Diffusion tensor fitting with FA, MD, AD, and RD maps
• Single-shell constrained spherical deconvolution
• Probabilistic iFOD2 whole-brain tractography
• SIFT filtering and track-density imaging
• Automated quality-control figures and quantitative summaries

Single-subject analysis of the classic SPM Auditory dataset, implemented independently in SPM25 and FSL FEAT.

• Contrast: Listening > Rest
• Motion quality control
• First-level GLM
• Statistical maps and cluster tables
• Cross-software comparison

Ten-subject ADHD-200 preprocessed resting-state workflow using the Schaefer-100 atlas.

• Nuisance regression and temporal filtering
• Subject-level Pearson connectivity matrices
• Fisher-z group averaging
• Minimum-spanning-tree plus 15% density graph construction
• NetworkX graph metrics, communities, hubs, and small-world analysis
• BRAPH-ready matrix and label exports

Reproduction of the Brainstorm median-nerve workflow.

• Sensor-level event-related fields
• Early somatosensory topography at approximately 20 ms
• Minimum-norm dSPM source reconstruction
• Morlet time-frequency analysis from 4 to 80 Hz

Single-subject T1-weighted structural MRI workflow.

• Bias-field correction
• Grey matter, white matter, and CSF segmentation
• Spatial normalization and modulation
• CAT quality-control assessment
• Global tissue-volume extraction
• Preparation of smoothed grey matter for VBM

Single-subject cortical reconstruction and morphometry using OpenNeuro ds000114.

• Complete recon-all cortical reconstruction
• Parallel execution with -openmp 8
• Subcortical and global volumetric statistics
• Cortical thickness, area, and volume by Desikan–Killiany region
• Coronal, sagittal, and axial Freeview quality-control figures

Completed, quality-controlled single-subject research portfolio demonstration using public OpenNeuro dataset ds002898, subject sub-01.

• 225 dynamic FDG PET frames selected from approximately 30 to 90 minutes
• 2.8 mm isotropic resampling and FSL MCFLIRT motion correction
• Static PET construction, SPM12 tissue segmentation, and PET-to-T1 coregistration
• PETPVE12 Müller–Gärtner partial-volume correction at nominal 4, 5, 6, and 8 mm PSF values
• Bilateral cerebellar-cortex SUVR and Desikan–Killiany regional summaries
• 84 primary grey-matter ROIs with median PSF CV 1.488% and maximum PSF CV 7.116%
• Final workflow validation passed
• Research portfolio demonstration only; not clinical or diagnostic software

These repositories describe planned or scaffold workflows rather than completed scientific analyses. A Zenodo DOI is added only after a completed, quality-controlled release.

Neuroimaging

SPM · FSL/FEAT · FreeSurfer · CAT · Brainstorm · MRtrix3 · PETPVE12 · BRAPH · EEGLAB · FieldTrip

Programming and reproducibility

Python · MATLAB · Bash · Git · BIDS · statistical modelling · scientific visualization

• Structural and functional brain connectivity
• Diffusion MRI and tractography
• Reproducible neuroimaging
• Multimodal integration
• Neuroimaging quality control
• Source imaging
• Graph-theoretical analysis
• Clinical and translational neuroscience

• ORCID: 0000-0001-5627-579X
• Google Scholar
• LinkedIn

Raw neuroimaging data are generally excluded from the repositories. Each completed project documents its data source, processing decisions, software environment, outputs, and limitations.