MIRACL

<p align="center"> <img src="docs/gallery/images/icon.png" alt="alt text" width="400" height="250"/> </p>

---

MIRACL (Multi-modal Image Registration And Connectivity anaLysis) is a general-purpose, open-source pipeline for automated:

1. Registration of cleared and imaging data (ex. LSFM and MRI) to atlases (ex. Allen Reference Atlas)
2. 3D Segmentation and feature extraction of cleared data
3. Tract-specific or network-level connectivity analysis
4. Statistical analysis of cleared and imaging data
5. Comparison of dMRI/tractography, virus tracing, and connectivity atlases
6. Atlas generation and Label manipulation

---

**NEW WORKFLOW/FEATURE ANNOUNCEMENT**

We will soon release our Mapping Axonal Projections in Light-sheet Fluorescence Microscopy in 3D (MAPL3) workflow as part of the MIRACL platform. MAPL3 is an end-to-end deep learning pipeline for generalizable, brain-wide mapping of structural connectivity at the single-axon level using tera-voxel light sheet fluorescence microscopy (LSFM) dataset. It's built for scalability, precision, and biological insight and bridges the gap between raw LSFM data and interpretable brain-wide connectivity maps.

• SPECTRE (Spatial Patch Encoding with Convolutional TransfoRmEr) Network: A novel deep learning design that fuses convolutional precision with transformer-level context awareness, capturing both fine axonal details and global anatomical structure.
• Self-Supervised Generative Pretraining: Pretrained on ~22,000 3D sub-volumes using advanced self-supervised learning and unique patch-based augmentation strategies. Enhances model robustness across diverse imaging conditions and experimental setups.
• State-of-the-Art Performance: Extensively benchmarked against leading DL pipelines, MAPL3 consistently outperforms them in both in- and out-of-distribution datasets, from patch-level inference to full-brain reconstructions.
• Neuroscientific Discovery at Scale: MAPL3 reveals cell-type-specific axonal connectivity patterns previously unresolved by existing methods, enabling new insights into mesoscale and quantitative brain circuitry analysis.

---

**NEW WORKFLOW/FEATURE RELEASE**

We have released our AI-based Cartography of Ensembles (ACE) workflow, an end-to-end, automated pipeline that integrates cutting-edge deep learning segmentation models and advanced statistical methods to enable unbiased and generalizable brain-wide mapping of 3D alterations in neuronal activity, morphology, or connectivity at the sub-regional and laminar levels beyond atlas-defined regions.

ACE is now available. Tutorials and usage examples for ACE can be found in our docs. As of MIRACL version 2.4.2 the pre-trained DL models are publicly available and will automatically be downloaded during installation.

---

We recommend using MIRACL with the Docker or Apptainer containers we provide. Legacy instructions for installing MIRACL locally are available but we recommend against it. For more details, see our docs.

---

Attention: We changed the license for MIRACL from GPL-3.0 to CC BY-NC-ND 4.0 as of version 2.5.0. For more information read our LICENSE.md or go directly to the Creative Commons website.

---

Copyright (c) 2025 Maged Goubran

All Rights Reserved