Sum Intensity Analysis

Sum intensity analysis measures normalized line intensity along a line-scan kymograph ROI and detects transient peaks from a functional reporter (like GCaMP). The ROI crop uses time along rows and space along columns.

Input data

Expects a line-scan kymograph and a rectangular ROI covering the spatial region used for the mean line-intensity trace. Peak detection uses normalized intensity (sum_intensity / spatial_pixel_count) so ROIs with different widths remain more comparable.

Signal pipeline

1. Compute row sums over the spatial dimension, with optional rolling averaging.
2. Normalize by spatial pixel count.
3. Optionally median-filter the normalized trace.
4. Optionally apply single-exponential detrending for photobleaching.
5. Estimate a scalar F0 baseline (percentile or manual).
6. Compute df/f0 and its time derivative.
7. Detect onsets (derivative threshold by default), enforce a refractory period, refine peaks, and measure fractional peak widths.

Programmatic use

from acqstore.acq_image import AcqImage
from acqstore.acq_image.analysis.sum_intensity_analysis.sum_intensity_analysis import (
    SumIntensityAnalysis,
)
from acqstore.acq_image.analysis.sum_intensity_analysis.sum_intensity_presets import (
    SumIntensityPresetName,
)
from acqstore.sample_data import ensure_sample_file

acq = AcqImage(str(ensure_sample_file('kymograph-diameter')))

channel = acq.images.channel_indices[0]
roi = acq.rois.create_rect_roi(name='sum_intensity')

detection_params = SumIntensityAnalysis.get_detection_preset_params(
    SumIntensityPresetName.MEDIUM
)

sum_intensity = acq.analysis_set.create_and_run(
    SumIntensityAnalysis,
    channel=channel,
    roi_id=roi.roi_id,
    detection_params=detection_params,
    replace_existing=True,
)
print(sum_intensity.result.summary)
acq.save()

Detection parameters

Built-in presets provide starting parameter sets:

Preset	Typical use
fast	Rapid transients
medium	General-purpose starting point
slow	Slower rise and decay kinetics

Presets are copied into the analysis when selected; later edits do not change the built-in preset registry. Manual F0 uses baseline_method="manual" and manual_f0_baseline. Percentile F0 uses baseline_method="percentile" and baseline_percentile.

For the full schema see the Sum Intensity Analysis API (get_detection_param_schema).

Results

For a source file my_file.tif:

my_file.tif.json
my_file.tif.sum_intensity.csv

Typical summary fields include num_peaks, f0_baseline, baseline_method, detection_source, peak_amplitude_mean, peak_amplitude_median, and peak_events. The CSV stores per-timepoint traces and onset/peak markers.

See also the Sum Intensity Analysis notebook.