Skip to content

Version History

CaliAli 1.4.5 Release Notes — October 2025

  • 🧮 Smarter Defaults & Auto Sizing: batch_sz accepts 'auto', dynamically estimating chunk size from available RAM and data dimensions; gSig now defaults to 5 / spatial\_ds, keeping neuron footprints correct whether or not data are downsampled.
  • 🎨 Coloured, Actionable Logs: Adopted cprintf across the pipeline for clearer status and warning messages, including new progress bars in CNMF-E loops.
  • 🧹 Safer Preprocessing Pipeline: Non-negativity enforcement now happens after noise scaling in CaliAli_remove_background, reducing over-clipping; neuron projections are converted to uint8 using a 99.99th percentile anchor to tame hot pixels.
  • 🧭 Robust Alignment & File Handling: Inter-session alignment survives per-file frame size changes, honours frame counts per chunk, and natural-sorts session pieces even when filenames are out of order; HDF5 readers query metadata directly instead of loading datasets.
  • 🔍 Initialization Preview: Added Check_initialization_parameters , a quick sanity check that shows how many neurons will be seeded before you run CNMF-E.
  • 🐞 Bug Fixes: The parameter selection app now saves the chosen minimum correlation threshold, preventing overly permissive seeding; motion correction once again accepts ISXD/ISXD2 exports without extra steps; and CaliAli better handles per-session size mismatches when files were cropped outside the recommended workflow.
  • 🎥 Enhanced Demo: Demo_pipeline.mlx now simulates full-resolution data via the Simulate_Ca_Imaging_video toolkit, illustrating the benefits of downsampling and motion correction on configurable synthetic motion artifacts.

CaliAli 1.4 Release Notes — September 26th 2025

  • 🚀 Automatic Session Chunking: batch_sz now splits oversized sessions across motion correction, alignment, and projection steps. See Processing Large Sessions for the workflow.
  • 🔧 Batch Pipeline Helpers: Added create_batch_list() and pre_allocate_outputs() for chunk-aware processing while keeping templates consistent.
  • 💾 Memory & Resilience Tweaks: Leaner spatial updates and better fallback behaviour when GPU jobs fail.
  • ✂️ Interactive Cropping: Added CaliAli_crop() to preview multiple sessions, draw a shared region of interest, and apply on-disk cropping before motion correction for RAM-friendly processing.
  • ⚙️ Projection Refinements: Smarter aggregation per chunk so combined projections match full-session runs.
  • 🔍 Dev Utilities: check_version_sync() plus environment reporting to simplify release management.
  • 🎯 Migration: batch_sz = 0 preserves legacy behaviour; v1.3-style manual splitting can be replaced by the new chunking flow.
  • ⏳ Downsampling Automation: Batch-mode automation for downsampling/conversion is still on the roadmap—open an issue if you need it prioritized.

CaliAli 1.3 Release Notes

🗓️ August 1st 2025

  • 🚀 Introduced low-memory mode for processing sessions that exceed available RAM

    • Allows splitting large individual sessions into smaller files
    • Applies non-rigid alignment only across actual sessions, not within
    • Reduces computation time and avoids alignment redundancy
    • Demo file provided.

  • 🔧 Simplified BV detection pipeline

    • Replaced vignetting correction with faster Gaussian-based filtering

  • 🐛 Fixed bug in CaliAli_parameter

    • Default values now load correctly when no input is provided
    • Supports structure-based parameter updates:
      opt = CaliAli_parameter(opt, 'param_name', value);

  • 📊 Added demo for interactive parameter selection

  • Includes tools for tuning BV size and neuron size

  • 🧩 Developed App Designer tool for estimating optimal neuron size filter

CaliAli 1.2.2 Release Notes

🗓️ May 30th 2025

  • 🔧 The do_alignment flag has been split into do_alignment_translation and do_alignment_non_rigid, so you can perform either alignment operation—or skip both if both flags are set to false—replicating the original do_alignment behavior when neither is enabled.
  • ➕ Introduced the force_non_negative_tolerance parameter. Previously, any negative pixel values produced after detrending were automatically trimmed; now, small negative values are permitted.
  • ⚡ Substantially improved the efficiency of the create_similarity_matrix function used to compare extraction similarities across CNMF iterations.
  • 🎥 Optimized play_movie for more efficient playback handling.

CaliAli 1.2.1 Release Notes

🗓️ April 16th 2025

  • 📚 Added documentation for processing split calcium imaging data files, including support for multiple video segments per session and TIFF files.
  • 🐛 Fixed a bug where the text progress bar did not display the final update.
  • 🐛 Fixed a bug where traces were not correctly displayed in postprocessing_app.
  • ⚠️ Disabled batch processing in v2uint16 and v2uint8 due to compatibility issues on some operating systems.
  • 🐛 Fixed a bug where the background component was not correctly handled in batch mode when using the SVG model.
  • 💾 Improved data handling when saving concatenated files, making the new code significantly faster.

CaliAli 1.2 Release Notes

🗓️ February 7th 2025

This update focuses on increasing the modularity of the CaliAli pipeline by restructuring how parameters and functions are handled. A central CaliAli_options structure now consolidates all settings required for video processing, enabling each module to operate more independently. This refactoring lays groundwork for future support of multiple calcium imaging data types (including both 2P and 1P) and specialized tools for dendritic signal extraction.

⚠️ Because of these significant changes, the previous workflow used to analyze data is no longer compatible, and older procedures will not work in the updated environment.

Other Changes
File / Group Change Summary
Downsample & ScanImageTiffReader (Downsample/…) Introduced CaliAli_downsample and batchConvertVideos; removed legacy Inscopix, AVI, temporal, and TIFF downsampling scripts; updated ISXD2h5 to omit downsampling; and added a new ScanImageTiffReader package with associated helpers and mex compilation adjustments for cpp‑tiff usage.
Motion Correction (Vessel_MC and other codes) Added new functions CaliAli_motion_correction and Rigid_mc; removed outdated MC_Batch and motion_correct_PV; introduced Non_rigid_mc and refinements in vessel–focused correction routines.
Other_codes (Utilities and Experimental) Added utilities for loading (CaliAli_load), saving (CaliAli_save, CaliAli_save_chunk), updating parameters (CaliAli_update_parameters, update_CaliAli_options), and various helpers (catpad_centered, concat_nan_centered, getSystemMemory, v2uint16/v2uint8 enhancements) with widespread improvements in memory management and file concatenation.
Postprocessing Modified functions for residual updates and manual corrections (get_seed, manually_update_residuals, mouse_click, postprocessDeconvolvedTraces, postprocessing_app, update_residual_custom_seeds) with refined parameter sourcing and GUI improvements.
Documentation (README.md & docs/…) Updated online documentation links; added numerous documentation files for new/updated functions (e.g. CNMFE_parameters, CaliAli_cnmfe, CaliAli_downsample, motion correction, postprocessing, etc.), enhancing details on syntax, inputs, outputs, and usage.

CaliAli 1.0.1 Release Notes

🗓️ April 22nd 2024

  • Introduced a new app for determining optimal blood vessel (BV) size.
  • Removed unnecessary or deprecated functions.
  • Several improvements to the documentation.

Full Changelog on GitHub.

CaliAli Stable Version 1.0 Release Notes

🗓️ April 19th 2024

We are excited to announce the first stable version of CaliAli, featuring significant enhancements and improvements. Below are the key changes in this release:

Changes:

  • Full Online Documentation: Access comprehensive documentation to guide you through using CaliAli efficiently.

  • Optimized Computational Performance: CaliAli is now optimized for low memory requirements, ensuring smoother and more efficient processing.

  • Improved Session Analysis: Analyze individual sessions and perform multisession concatenation seamlessly.

  • Enhanced Blood Vessel (BV) Extraction: Minimized vignetting artifacts for improved accuracy in BV extraction.

  • New BV Stability Metric: Evaluate tracking performance with a new stability metric integrated into BV extraction.

  • Bug Fixes in GUIs: Addressed several bugs in the graphical user interfaces (GUIs) for improved usability.

  • Enhanced Inter-Session Alignment: Further improvements to the inter-session alignment module for precise video session alignment.

Full Changelog on GitHub.

CaliAli Beta Release 1.0-beta

May 19th 2023

The first beta release of CaliAli is now available, offering advanced capabilities for extracting neural signals from one-photon calcium imaging data in free-moving conditions.

For details, refer to the BioRxiv preprint.

Explore CaliAli to analyze calcium imaging data with accuracy and efficiency, shaping the future of neural signal extraction in neuroscience research.