Automated Accept/Reject Cells¶

Overview¶

The Automated Accept/Reject Cells tool is used to automatically classify cell statuses based on footprint and trace metrics.

Input Data¶

There are 2 main inputs to the Automated Accept/Reject Cells tool. They are:

Eventsets are optional arguments as only certain criteria (parameters) are dependent on event-related metrics.

Input Cell Set Files¶

Cell set files (also accepts series cell set files) that contain cell traces and footprints.

Input Event Set Files¶

Event set files (also accepts series event set files) that contain cell event timestamps and amplitudes. Each event set file is derived from a cell set file from the same recording.

Parameters¶

There are 6 input parameters to the Automated Accept/Reject Cells tool. Although parameters are optional, the user must enter value for at least 1 parameter for the tool to execute, i.e., we need at least 1 condition to filter cells and classify them as either accepted or rejected.

Depending on the values entered for the parameters, the Automated Accept/Reject Cells tool combines the conditions to filter cells, i.e., a cell is accepted only if all the conditions are satisfied. For example, if a user sets SNR Greater Than (>) to 2, SNR Less Than (<) to 5, and Cell Size Greater Than (>) to 10, a cell will be classified as accepted only if all the conditions are met. If any of the conditions is not met, the cell will be classified as rejected.

A detailed description of these parameters can be found below. For more information, see the documentation for Calcium Imaging Quality Control.

SNR¶

The signal-to-noise ratio (SNR) of a trace is a measure of how much larger the values associated with events are compared to the noise level. More formally, it is the median amplitude of the trace at event times divided by the median absolute deviation of the trace.

where the MAD is the median absolute deviation of a cell trace. Traces with higher SNRs are more cell-like because occasional bursts of action potentials should cause large relative rises in the trace from baseline noise. The SNR depends on the calcium indicator, calcium concentration dynamics of the imaged cell type, and on the state of the neural network.

SNR Greater Than (>)¶

Minimum SNR below and including which a cell will be classified as rejected.

SNR Less Than (<)¶

Maximum SNR below and including which a cell will be classified as rejected.

Cell Size¶

Cell diameter here refers to the maximum width of a cell, i.e., it is the longest distance between two points in the largest connected component’s contour in pixels. The cell size depends on the imaged cell type and on the spatial downsampling factor used at acquisition and at preprocessing.

Cell Size Greater Than (>)¶

Minimum cell diameter (in pixels) below and including which a cell will be classified as rejected.

Cell Size Less Than (<)¶

Maximum cell diameter (in pixels) above and including which a cell will be classified as rejected.

Event Rate¶

The event rate is a value in Hertz that describes how frequently calcium events occur. More formally, it is the number of events associated with a trace divided by the total duration of the trace in seconds.

Event Rate Greater Than (>)¶

Minimum event rate (in Hertz) below and including which a cell will be classified as rejected.

Event Rate Less Than (<)¶

Maximum event rate (in Hertz) above and including which a cell will be classified as rejected.

# of Components¶

The number of connected components in the image. The connected components are identified by finding the contours in the binarized version of 𝑆′′ using the algorithm described in ¹. A neuron is expected to be composed of only one spatial component. However, factors such as noise or poor results from PCA-ICA may result in a footprint with more than one component. Increasing the possible number of components (for example by setting the parameter to < 3) can compensate for noisy data but may result in more false positives.

# of Components Greater Than (>)¶

Minimum number of components in the cell footprints below and including which a cell will be classified as rejected.

# of Components Less Than (<)¶

Maximum number of components in the cell footprints above and including which a cell will be classified as rejected.

Circularity¶

The circularity is a measure of roundness for a cell and is approximated as follows

where the area is defined here. It is used to identify highly non-circular cells, which may be considered problematic and hence are candidates for rejection.

Circularity Greater Than (>)¶

Minimum circularity below and including which a cell will be classified as rejected.

Circularity Less Than (<)¶

Maximum circularity above and including which a cell will be classified as rejected.

Footprint Area¶

The footprint area is the number of pixels in each footprint, thresholded by 25% of the maximum intensity of that footprint.

Footprint Area Greater Than (>)¶

Minimum footprint area (in pixels) below and including which a cell will be classified as rejected.

Footprint Area Less Than (<)¶

Maximum footprint area (in pixels) above and including which a cell will be classified as rejected.

Trace Skew¶

The trace skew is the skewness of a trace. It tends to correlate with trace quality.

Trace Skew Greater Than (>)¶

Minimum trace skew below and including which a cell will be classified as rejected.

Trace Skew Less Than (<)¶

Maximum trace skew above and including which a cell will be classified as rejected.

Maximum Correlation To Other Traces¶

The maximum correlation to other traces is a measure how closely a cell trace relates to other cell traces. Large values of cell to cell trace correlation can arise due to over-segmentation, where footprint of a single cell is split into two over-segmented "cells", resulting in these two objects having highly correlated traces. Large correlations need not necessarily mean that these cells are over-segmented, but can identify cells that warrant closer inspection.

Maximum Correlation To Other Traces Greater Than (>)¶

Minimum maximum correlation to other traces rate below and including which a cell will be classified as rejected.

Maximum Correlation To Other Traces Less Than (<)¶

Maximum maximum correlation to other traces above and including which a cell will be classified as rejected.

Spearman Correlation¶

The Spearman correlation is a non-parametric measure of rank correlation and scales with how well the relationship between the cell trace and time can be described using a monotonic function. This metric captures non-linear trends just as well as linear trends, and does not require explicit model fitting. Large absolute values of the Spearman correlation indicate that the trace is monotonically increasing or decreasing with time.

Spearman Correlation Greater Than (>)¶

Minimum spearman correlation below and including which a cell will be classified as rejected.

Spearman Correlation Less Than (<)¶

Maximum spearman correlation above and including which a cell will be classified as rejected.

Goodness of Fit¶

The goodness of fit is a measure of how well cell traces are described by a decaying exponential. The equation used to fit the trace is the same as in the Movie trends figure. Cells with a large goodness of fit are well described by a decaying exponential, and are therefore candidates to be rejected.

Goodness of Fit Greater Than (>)¶

Minimum goodness of fit below and including which a cell will be classified as rejected.

Goodness of Fit Less Than (<)¶

Maximum goodness of fit above and including which a cell will be classified as rejected.

Median Decay¶

The median decay is the number of seconds for a trace to decline to 50% of its peak amplitude. This should match what is known about the biology of these cells and the kinetics of the calcium indicator.

Median Decay Greater Than (>)¶

Minimum median decay (in seconds) below and including which a cell will be classified as rejected.

Median Decay Less Than (<)¶

Maximum median decay (in seconds) above and including which a cell will be classified as rejected.

Output Data¶

The output of the tool execution of a cell set file with revised statuses of cells as either accepted or rejected based on the values entered for parameters.