Skip to content

Latest commit

 

History

History
51 lines (39 loc) · 2.26 KB

README.md

File metadata and controls

51 lines (39 loc) · 2.26 KB
Raw

Optosynth

Optosynth

Optosynth is a python package for generating synthetic voltage imaging ("Optopatch") data using empirical morphological reconstructions and paired electrophysiology measurements from Allen Brain single-neuron database.

Optosynth currently simulates the following aspects of voltage imaging data:

  • spatial variability in expression of archaerhodopsin
  • action potential dendritic backpropagation
  • voltage propagation decay
  • lowpass filtering of electrophysiology measurements according to the camera sampling rate
  • conversion of voltage to fluorescence
  • static and dynamic background fluorescence
  • variability in global fluorescence intensity of individual neurons (i.e. tissue depth)
  • Gaussian PSF
  • Poisson-Gaussian camera noise

The synthetic data generated by Optosynth is extremely realistic and can be used for benchmarking denoising, spike detection, and segmentation pipelines.

Future plans include:

  • multi-threading
  • removing redundant parameters
  • camera motion modeling (towards synthetic in vivo data)
  • better reconstruction of soma (currently cicular -- Allen SDK only provides center coordinates and radius)
  • 3D tissue/culture synthesis and 3D point spread function (currently, the reconstruction is purely 2D)

Requirements

Optosynth requires the following python modules:

  • allensdk
  • Pillow>7.0.0 (having the right version is extremely important -- Optosynth produces bad data with older versions of Pillow)
  • boltons
  • tqdm
  • torch
  • numpy
  • scipy
  • pandas

We will eventually package Optosynth into a proper python module. At the moments, the user must install the requirements manually.

Notebooks

  • download_allen_data.ipynb: downloads relevant data from Allen Brain
  • process_allen_morphology.ipynb: processes morphological reconstructions and generates binary masks for each neuron
  • process_allen_elecrophysiology.ipynb: processes electrophysiology data, retains only square pulses, and generates a manifest of retain sweeps along with useful metadata (stimulation current, number of spikes, etc)
  • optosynth_main.ipynb: generates and saves synthetic data, explores some aspect of the generated data