I am trying to detect calcium dynamics in boutons in one plane, but suite2p is missing many of them, even ones with high SNR. Another issue is that it is creating an ROI where there does not appear to be a bouton.
I tried using the default dendrite/axon settings on the GUI and have also adjusted several variables I will list below. Does anyone have any other advice on how to optimize bouton detection? I am including a sample image with and without masks. My images have boutons of different sizes and boutons that are part of a traversing axon and others that are not. I am wondering if this variability is affecting the output.
I tried decreasing threshold scaling and it does increase # of ROIs but only to a certain point. It plateaus at about 0.7. Shown in my image below.
I have tried increasing max_iterations up to 40 and it does not make a difference.
I tried running the file with and without nonrigid registration and does not make a difference.
Any ideas would be greatly appreciated!
Here's some info about other settings.
I'm using fs=10 because I'm imaging at 10Hz in one plane.
For tau, I've tried, 1-1.5. I'm using GCaMP6s.
These are 2P recordings, each movie has 1500 frames.
Hello,
I am trying to detect calcium dynamics in boutons in one plane, but suite2p is missing many of them, even ones with high SNR. Another issue is that it is creating an ROI where there does not appear to be a bouton. I tried using the default dendrite/axon settings on the GUI and have also adjusted several variables I will list below. Does anyone have any other advice on how to optimize bouton detection? I am including a sample image with and without masks. My images have boutons of different sizes and boutons that are part of a traversing axon and others that are not. I am wondering if this variability is affecting the output.
I tried decreasing threshold scaling and it does increase # of ROIs but only to a certain point. It plateaus at about 0.7. Shown in my image below. I have tried increasing max_iterations up to 40 and it does not make a difference. I tried running the file with and without nonrigid registration and does not make a difference.
Any ideas would be greatly appreciated!
Here's some info about other settings. I'm using fs=10 because I'm imaging at 10Hz in one plane. For tau, I've tried, 1-1.5. I'm using GCaMP6s. These are 2P recordings, each movie has 1500 frames.