Tutorial - Input Source Selection

Quick Overview

📦 Make sure GEM-pRF is installed (see the Installation page)
⬇️ Download the GEMpRF DemoKit
▶️ Run the interactive demo script
🔍 Choose examples 001–005 to explore different input source configurations

⚠️ Ensure that you have installed GEM-pRF package (follow the Installation page).

Step-by-Step Guide

Step 1. Download the DemoKit

This repository contains sample fMRI/MRI data and example configurations.

git clone https://github.com/siddmittal/GEMpRF-DemoKit.git
cd GEMpRF-DemoKit

Step 2. Run the interactive demo

The demo script guides you through multiple examples. To explore input sources, select options 001–005.
```
python run_gemprf_demo.py
```

Step 3. What happens during the demo

Interactive menu: choose examples showing different input styles
Automatic path fixing: XML configs update to your local DemoKit paths
GPU memory check: prevents running models that won’t fit
Full GEM-pRF run: executes gp.run() using the selected configuration

How Input Sources Work

GEM-pRF accepts inputs from multiple sources: raw filesystem paths or structured BIDS datasets. Everything is controlled through XML configuration files, not by modifying GEM-pRF code.

Filesystem paths: direct references to your fMRI and stimulus files
BIDS datasets: GEM-pRF locates the correct files based on subject/session

Examples You Will See (001–005)

001: BIDS + prfprepare (surface), individual runs
002: BIDS + fMRIPrep (surface), individual runs
003: BIDS + fMRIPrep (volume), individual runs
004: BIDS + fMRIPrep (surface + volume), individual runs
005: Fixed paths (non-BIDS) + prfprepare (surface)

Notes

Ensure GEM-pRF and required dependencies (e.g., CuPy for GPU support) are installed. When adapting examples to your own data, check the XML comments to see which fields control input behaviour.