Utility scripts for post-processing output from MEA-NAP, the MATLAB pipeline for analyzing microelectrode array (MEA) recordings. These scripts operate on MEA-NAP's CSV exports to handle specific downstream tasks: merging a run's per-condition exports, filtering recording groups, visualizing per-channel firing rate distributions, measuring how far stimulation moves each channel from its own baseline, and checking normality of selected metrics ahead of statistical testing.
Plots per-channel firing rate distributions from a MEA-NAP node-level CSV (FileName, Grp, Channel, FR). By default it opens an interactive viewer in your web browser, with dropdowns for group and organoid, hover tooltips identifying the source recording of each point, and a toolbar button that saves the current view as a high-resolution PNG. The viewer is a single self-contained HTML file; it loads the Plotly.js charting library from the internet the first time it runs on a machine. Static figures can also be rendered directly to a file.
Each stim type / condition in the file (e.g. stim1, stim3, base) is analysed as its own experiment, with its own boxes, organoids and slices computed from that condition's recordings alone. A CSV holding several conditions therefore produces several viewers/figures — one browser tab per condition, or one output file per condition when saving (-o viewer.html becomes viewer_stim1.html, viewer_stim3.html, ...). Use --stim to restrict the run to a single condition.
python3 fr_boxplots.py NeuronalActivity_NodeLevel.csv # interactive viewer (opens in browser)
python3 fr_boxplots.py data.csv --list # list stims/groups/organoids/slices
python3 fr_boxplots.py data.csv -o viewer.html # write the interactive viewer(s) to file(s) to share
python3 fr_boxplots.py data.csv --stim stim1 -o viewer.html # a single condition only
python3 fr_boxplots.py data.csv --grp BCTL --organoid CT7 -o bctl_ct7.png # save a static figure
Measures how far stimulation moves each channel from its own baseline, as a percentage, from a single node-level CSV holding every condition:
percentage difference = 100 x (stim FR - base FR) / base FR
A _stim recording is paired with a _base recording when the run ID and the organoid slice match, so R250929CT1A_DIV250_stim1 is compared against R250929CT1A_DIV250_base and never against another slice or another run. Not every slice was recorded under stimulation; a slice holding only a baseline (or only stimulation) has nothing to compare, so no percentage is computed and it gets no panel. --list reports which slices paired and which did not, and why.
The viewer is one HTML file: a panel per slice, laid out as a grid so every slice can be scanned at once, with a dropdown that zooms into a single slice full width. Each panel plots the percentage difference against channel, one colour per stimulation pattern, sharing one y-axis so a +10% slice cannot be mistaken for a +900% one (untick the box, or pass --per-panel-y, to let each panel scale to its own data). Clicking a legend entry hides that pattern in every panel at once.
Some channels report 0 Hz for reasons that have nothing to do with the organoid: the electrode was grounded, or it was the one delivering the stimulation and so recorded nothing while it fired. Both show up either as a baseline firing rate of 0, which leaves no percentage to compute, or as a baseline that is fine while every stimulation recording reads 0, which comes out as a flat −100%. Neither is a firing-rate change, so the channel is not plotted; its number is printed in red along that panel's x-axis instead, and --list names it.
python3 fr_diff.py NeuronalActivity_NodeLevel.csv # interactive viewer (opens in browser)
python3 fr_diff.py data.csv --list # which slices paired, and why the rest did not
python3 fr_diff.py data.csv -o diff.html # write the viewer to a file to share
python3 fr_diff.py data.csv --slice CT1A -o diff.html # open on one slice instead of the grid
Needs fr_boxplots.py beside it — it shares that script's CSV reader and file-name grammar — and its page template in viewers/, but unlike fr_boxplots.py it does not need matplotlib.
Merges two or more MEA-NAP CSV exports that share a column layout and a recording run into one file — typically the per-condition node-level exports of a single run (_base, _stim1, _stim3, _stimLR, _stimRL).
Nothing is written until three checks pass. Every file must carry the same set of columns (compared ignoring case and order; the merged file uses the first file's spelling and order); every row must come from the same run, identified by the leading R<digits> token of its FileName — R250929 in R250929CT7A_DIV250_stim1; and no recording may appear in more than one file. A file holding two runs, or one from a different run than the rest, stops the merge, as does an overlap between inputs — a FileName repeats within a file once per channel, which is expected, but the same FileName in two files would double that recording's rows. Rows are otherwise passed through untouched and in the order given: no de-duplication and no reordering. The inputs are never modified; the default output is NeuronalActivity_NodeLevel_base_stim_merged.csv, written beside the first input.
python3 merge_csv.py base.csv stim1.csv stim3.csv
python3 merge_csv.py stimLR.csv stimRL.csv -o combined.csv
Keeps or drops rows of a CSV by their Grp value. --grp is a keep list — --grp PreStim keeps only those rows and discards everything else — while --drop and --drop-prefix are drop lists, by exact value and by leading characters respectively. The two drop flags may be combined and the union is removed, but neither may be combined with --grp. All three take several values, comma-separated (--drop a,b) or by repeating the flag (--drop a --drop b). Matching ignores case and surrounding whitespace; --grp and --drop match the whole value, so prestim matches PreStim but not prestim2.
--list prints each distinct Grp value with its row count and exits — worth running first. Before anything is written, a per-Grp summary of what is kept and what is dropped is printed along with the recordings being lost, and confirmation is asked for; the input is never modified. Passing no selection flag is an error rather than a default, so nothing is filtered by accident.
python3 grp_filter.py input.csv --list # distinct Grp values and row counts
python3 grp_filter.py input.csv --grp PreStim,PostStim # keep only these groups
python3 grp_filter.py input.csv --drop CCTL --drop-prefix B # drop the union of both
python3 grp_filter.py input.csv --drop-prefix C -o filtered.csv # the old CCTL/CMOS/CMUT behaviour
Runs Q-Q plots, Shapiro-Wilk, and/or Kolmogorov-Smirnov tests on selected numeric columns, broken out by Grp. Q-Q plots are saved as PNGs; test statistics print to the console.
./assess_normality.R input.csv --columns FR,Burst_Rate
./assess_normality.R input.csv --columns FR --tests shapiro,ks --groups BCTL,BMOS
The Python scripts depend only on matplotlib. assess_normality.R uses base R and requires no additional packages.
With uv:
git clone https://github.com/mfang21/meanap-scripts.git
cd meanap-scripts
uv sync
With pip:
git clone https://github.com/mfang21/meanap-scripts.git
cd meanap-scripts
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
R scripts run with any current R installation: Rscript assess_normality.R ....
- Install Python (3.10 or later).
- Install R if you plan to run
assess_normality.R. - Download this repository: on this page, click Code → Download ZIP, then unzip it.
- Open Terminal (macOS) or Command Prompt (Windows) and navigate to the unzipped folder:
cd path/to/meanap-scripts - Install the required Python library:
pip install -r requirements.txt - Run a script:
On Windows, run the R script as
python3 fr_boxplots.py your_data.csvRscript assess_normality.R your_data.csv --columns FR.