Home

#Welcome to the neurofinder wiki!

####Source output format:

• 3D datasets have 3 numbers in “coordinates”
• Bounding box can be user specified to be larger than max / min coordinates if desired
• Neuropil can be represented as either: negative numbers in “values”, or as two sources files with matched id numbers, one neuropil, one neurons.

---

####Blocks, merging, and transforms:

• Blocks size and merging function should be parameterized.
• Feed arbitrary list of block centers and sizes.
• Force block to return at most one source.
• Source extraction algorithms should be submitted with a transform function specified
• Transform function will do neuropil subtraction if desired
• Transform function will do temporal unmixing between overlapping sources if desired.
• String together feature, methods, merging, and transforms in a modular fashion.

---

####Ground truth

• Manually chosen centroids + ridge finding algorithm. One for filled neurons, one for donuts.

---

Evaluation metrics:

Metrics based on centroids, pixel overlap, and fluorescence trace.

Centroid:

• Distance to ground truth
• Binary value if under distance threshold

Pixel overlap:

• % overlap if the values of thresholded values
• Cross correlation if weighted values

Fluorescence trace evaluation:

• Correlation to ground truth fluorescence trace

Want to assign a quality metric to sources - high confidence, low confidence

Metric should treat negative values corresponding to neuropil appropriately

Evaluate different portions of the algorithms separately. i.e. did you return the right number of neurons per block i.e. starting with the right centers do you find the right boundaries

---

####For the future:

• Df/f to spike conversion
• Cross-day merging (do you find same neurons?)
• Finding axons / dendrites - block methods may work well, but need different merging function
• Incorporate motion correction into source extraction
• Incorporate detrending
• Estimate z-motion for timeseries