alarmingly high ΔF/F in the filtered raw data trace

[zylu1109]

Hello everyone,
I'm using the CaImAn-MATLAB-0.7 version matlab code. I'm not sure where to find the release version but it was downloaded in 2024 May.
I use it to analyze mouse hippocampal neuron activities. The ΔF/F I obtained seems quite high (I expected the value to be less than 3). I pasted an example result figure and my parameters initialization below.
I'm not sure if I set some number wrong about the code, or my ΔF/F is actually normal. I'm not an expert in either the code or the neuroscience part so any insights are welcome. Thanks in advance!

K = 30;
tau = 9; % std of gaussian kernel (half size of neuron in pixels)
p = 2;
options = CNMFSetParms(...
'd1',d1,'d2',d2,... % dimensionality of the FOV
'p',p,... % order of AR dynamics
'gSig',tau,... % half size of neuron
'merge_thr',0.85,... % merging threshold
'nb',1,... % number of background components
'min_SNR',1,... % minimum SNR threshold
'space_thresh',0.5,... % space correlation threshold
'cnn_thr',0.5... % threshold for CNN classifier
);

[zylu1109]

Another quick follow-up question is my trace baselines are never at 0 but always slightly higher, does anyone know why? Thanks!