Version 1.11.0

Happy new year everyone!

This releases brings a new network burst detection method (called "Threshold") that you can access in the advanced burst detection settings. There has also been some behind-the-scene changes to some features that we will announce once they are more formalized.

Other changes:

• We added support for axion data conversion when there is missing / disabled wells
• Added single channel burst metrics

Other minor changes:

• added error warning if genotype column in your input csv file is empty
• fixed auto-filling of DIV
• some network metrics will be NaN if the network size is smaller than the specified threshold
• fixed saving of .fig figures

Enjoy! Thanks again for using / trying out MEA-NAP, and please raise any issues that you have discovered.

Tim

Version 1.10.2

Major changes:

• Axion layout is now corrected, previous version may caused the layout to be rotated. Please double check if the new layout, including channel numbers, matches what you see in the acquisition programme. If not, please raise a github issue.
• Single channel burst detection calculation are not fixed. The parameters are currently not fine-tuned, but may be of interest. Future release will include single channel burst statistic plots.
• Added checking for STTC.c code, and if not compile, will revert to STTC.m function. Note: users have reported issue with this when running on Mac OS, so the current advice for mac users is to make sure the STTC.c code works.

Minor changes:

• Linear discriminant analysis (LDA) now handles rank-deficient matrices
• Added checking for sampling rates below 10000Hz for Wid parameter in spike detection settings
• active node definitions is now updated to be based on firing rate rather than based on correlation values
• added handling of linear-mixed effects model error in cases where there is zero variance for a particular feature

Version 1.10.1

This release includes minor bug fixes that was caused by the recent new feature of allowing user to subset nodes based on their activity levels. There is also better handling of empty adjacency matrices and cases where a csv output cannot be generated because no data is available.

Version 1.10.0

Major changes:

• GUI is updated with the tabs being shown on the left, the aim for this is to allow users to quickly view all available tabs and be able to jump between them
• We've added a bunch of network plot customisations. You can find them under the "Plotting" tab. Users can now set to maximum node size, as well as how the node size scales with node metrics. Users can also choose between whether nodes are plotted according to their physical location, or via force-directed graph layouts to allow for easier visualisation of communities in the network

Minor fixes:

• h5 conversion now skips files that fail to convert, instead of crashing
• Plots by lag are now improved with scatter points on top of lines, which make it visible when users are only using one lag

Version 1.9.2-alpha

This is a minor release with the following fixes:

• fixed node cartography failing when node value lies exactly at the boundary
• fixed issue when using previous analysis but (some) previous analysis file were not found

Version 1.9.1

This is a minor release with the following fixes:

• fixed issue with node by group and node by age plots not being dependent on the lag value (issue introduced in v1.9.0)
• fixed burst detection crashing when the total number of inter-spike-intervals across the entire recording is less than 10
• fixed issue with running MEANAP using spike detected data from older versions (before v1.9.0)

Version 1.9.0

Version 1.9.0 includes an important bug fix to an issue in version 1.8.0 which can cause the pipeline to crash when running MEANAP with data from multiple groups, we therefore recommend users of version 1.8.0 to upgrade to version 1.9.0

User interface changes:

• "Raw data folder" is now renamed "MEA Data Folder" to avoid confusion with ".raw" files
• User can now specify the output folder name. In previous versions, it will always be a folder named "OutputData" followed by the date
• Previous analysis folder is now the folder where a previous MEA-NAP run was performed, including the subfolder. Therefore, the previous analysis date field is now removed.

Feature improvements:

• users can now generate batch csv file for both Axion and multichannel system files (Under the File Conversion tab)
• Enhanced detection of issues with provided spreadsheet, including empty or non-numerical values in the DIV column
• Custom group order / default group orders are now sensitive to the spreadsheet range. This means that running a subset of data in a spreadsheet will result in only those subset of groups being plotted in group by DIV and group by genotype plots
• In line with the user interface changes, users can now continue analysis in the same output folders as a previous output data folder. Eg. You can now run step 1 and 2 on a specific output folder, and continue running step 2 onwards in that folder. (NOTE: This feature currently only works up to step 2, I am still working on allowing users to re-run step 4 within the same folder)

Bug fixes

• Fix bug causing pipeline to crash if at least one network does not have enough nodes for analysis, resulting in participation coefficient (PC) not being calculated for that network
• Fix bug causing pipeline to crash during creation of electrode-level data in step 4, this can occur when there are recordings with too few nodes for network-level analysis

Experimental features

• By default, if the user specified that they are using previous analysis, the pipeline will now check if previous network metrics were already computed in the previous run, and skip computation of those metrics. Future updates will ensure this only happens when the current parameters for analysis are the same as the previous analysis parameters, in cases where users want to re-run the analysis. You can disable this behaviour by going to the "Connectivity" tab and selecting "Recompute metrics", in which case the pipeline will ignore previously analysed metrics and re-compute all of the metrics

Version 1.8.0

IMPORTANT NOTE: There is a bug in version 1.8.0 that causes MEANAP to crash when running data with more than one group, this is fixed in version 1.9.0 and we therefore recommend users to switch to that / the latest version

This release introduces changes that make the pipeline neater, and make it easier for users to test and view outputs from the pipeline.

Major changes:

• Users can now download example data and test out the pipeline directly from the GUI. After launching MEA-NAP with MEApipeline, click on the "Test Pipeline" button and the application will download example raw matlab data, and run through the pipeline. First time users will find it helpful to run through this after installation to make sure all required packages and code are installed properly
• Users can now view figure outputs from the pipeline from the application. Click on the "View Outputs" button either before or after running the pipeline, and you will see a new launched application that allows you to view the plots from a selected output folder. From this viewer, you can read the corresponding figure title and legends that we provide in the documentation

Minor changes:

• added activeChannelcolumn to node level network activity csv
• moved messages from command prompt to application status area
• you can now close the pipeline window without getting an error message
• we are now sticking with a single numerical versioning system, instead of having a date-based versioning system for quick fixes
• step 4 plots that are grouped by age or by "group" (eg. genotype) now have a figure suffix to indicate so
• MEANAP version number now displayed in the MEANAP app window
• MEA-NAP now has a logo! You can see it in the GUI and the README
• When you select the raw data folder in the GUI, it will print out the number of mat files found in that folder, just to help you check you have selected the correct folder

Known issues:

• The pipeline may crash if running on datasets with more than one group, this issue is fixed in version 1.9.0

Version 1.7.0

Major changes

• MEA-NAP now checks if you are in the latest version of the pipeline, and warns you if not
• GUI now supports more file conversion options (eg. Axion)
• MEA-NAP can now run as a function with the input being the path to the parameter file

Minor changes:

• defaults to calculating average and modal controllability
• better handling of NaNs and empty networks

Version 1.6.0

This version provides a bunch of minor fixes, and now allows step 4B and step 4A to be run separately. In the future we plan to apply this to other steps to make the pipeline more modular.

Major changes

• within step 4, step 4B (summary statistics across all recordings) is now performed first, this makes it faster to have an overview of the results before individual network plots are made
• step 4A and 4B can now be run separately, you can specify this in the GUI / in the settings file by setting startAnalysisStep as '4A' or '4B', setting it to 4 will run both

Minor fixes:

• advanced settings file is now renamed from biAdvancedSettings.m to AdvnacedSettings.m
• Allow for zero variance in within-class variables in linear discriminant analysis
• STTC lag unit in the GUI is corrected from seconds (s) to milliseconds (ms)
• fixed title placement for spike detection check plot
• and many others (see commit history)