Minian Pipeline Analysis Troubleshooting

[Sarvesh30]

Hi,

I am reaching out about troubleshooting the preprocessed/motion corrected video from the minian pipeline. The following parameters for denoise and motion correction are listed below:

param_denoise = {
"method": "gaussian",
"ksize": (3, 3),
"sigmaX": 0
}

param_estimate_motion = {"dim": "frame"}

The background removal step was not used as this step lowered the brightness of the final video and made it difficult to view individual cells compared to the original video. I have attached two preprocessed/motion corrected videos after the glow removal, denoise, and motion correction steps. The problem that I am facing is that there are only a few cells that can be clearly seen in the video. Are there any possible parameter values that can enhance the output? Thank you for any help. I really appreciate it.

Video1
Video2

Best regards,
Sarvesh

[austinbaggetta]

Hi Sarvesh,

As you mentioned, from the videos I can see some clear cells, but it looks generally pretty sparse. This isn't a Minian issue but could be due to a variety of factors at the surgery level and can be influenced by what brain region you're imaging from. Is the problem that you are not detecting these cells in the Minian output and have no spatial footprints for these cells, or that there are just a few cells to begin with? If you do not have spatial footprints for the cells that exist in the video, there could be a multitude of steps where they are being dropped. If the problem is that there are only a few cells, that is not a Minian issue.

You may try either logarithmic or exponential transformation to expand the range of low intensity pixel values or high intensity pixel values respectively, but I cannot promise that this will discover cells that aren't there to begin with.

[Sarvesh30]

Hi austinbaggetta,

Thank you so much for your response. We have been troubleshooting this issue and have come across the following problems. It would be really appreciative if you can please help us out in addressing these issues. I have attached the issue as a reference below:

#289

Best regards,
Sarvesh