Parameters for CNMF-E¶
| Name | Description |
|---|---|
| Cell diameter | The average cell diameter of a representative cell in pixels. |
| Minimum correlation | The minimum correlation of a pixel with its immediate neighbors when searching for new cell centers. |
| Min peak-to-noise ratio | The minimum peak-to-noise ratio of a pixel when searching for new cell centers. |
| Background Spatial Subsampling | The spatial downsampling factor to use when estimating the background activity. |
| Ring size factor | The multiple of the average cell diameter to use for computing the radius of the ring model used for estimating the background activity. |
| Gaussian kernel size | The width in pixels of the Gaussian kernel used for spatial filtering of the movie before cell initialization. This can render cells more visible and more easily identifiable during initialization. |
| Closing kernel size | The size in pixels of the morphological closing kernel used for removing small disconnected components and connecting small cracks within individual cell footprints. A morphological closing operation consists of a dilation (expansion) followed by an erosion (shrinking). |
| Merge threshold | The temporal correlation threshold for merging cells that are spatially close. |
| Processing mode | The processing mode to use to run CNMF-E. See below for details |
| Number of threads | The number of threads to use for processing. When using the All in Memory processing mode, the entire field of view will be processed at once using the specified number of threads for operations where parallelization is possible. When using the Sequential Patch mode, each patch will be processed sequentially using the specified number of threads where parallelization is possible. When using the Parallel Patch mode, patches will be processed in parallel each using a single thread. |
| Patch size | The side length of an individual square patch of the field of view in pixels. This parameter is only available for sequential and parallel patch modes. |
| Patch overlap | The amount of overlap between adjacent patches in pixels. This parameter is only available for sequential and parallel patch modes. |
| Output Unit Type | The units of the output temporal traces.See below for details. |
Output Unit Types¶
dF¶
Temporal traces on the same scale of pixel intensity as the original movie. dF is calculated as the average fluorescence activity of all pixels in a cell, scaled so that each spatial footprint has a magnitude of 1.
dF over noise¶
Temporal traces divided by their respective estimated noise level.
It is common practice to standardize time series by subtracting the mean and dividing by the standard deviation, effectively converting each value to a Z-score. dF over noise uses a similar formula, with the exception that the traces are divided by the estimated noise level as opposed to the standard deviation. Since noise is estimated by averaging the high-frequency fluctuations along the trace, dF over noise is more robust to outliers than standard scores. More precisely, outliers would inflate the standard deviation measured and thus reduce the standard scores computed. The approach used to measure noise here is less susceptible to the influence of outliers.
Processing Mode¶
All in memory¶
Processes the entire field of view at once. Use this option if you have a lot of memory, processing time is not a concern, and you want to avoid potential border effects associated with patch mode.
Sequential patches¶
Breaks the field of view into overlapping patches and processes them one at a time using the specified number of threads where parallelization is possible. Use this option if you have a very limited amount of memory and processing time is less important.
Parallel patches¶
Breaks the field of view into overlapping patches and processes them in parallel using a single thread for each. The number of patches being processed at any given time is controlled by the specified number of threads. Use this option if you have a large amount of memory and want to reduce processing time as much as possible.