I am performing simultaneous patch clamp electrophysiology and widefield, 1p voltage imaging. To synchronize my recordings from these two channels, I am using NI Labview to send a TTL pulse to my Hamamatsu Fusion BT to start my sequence acquisition at the same time as my patch clamp electrode amplifier starts recording.
I have noticed a perplexing artifact when trying these experiments. Only when I externally trigger the camera to start, I get an artifact in which the pixel intensities (on average) start low, gradually increase for 1.5-2s, then decrease. This occurs even when the lens cap is on the camera and no light should be hitting the sensor.
Here is an example of what the mean intensity values look like for each frame across a 5000 frame recording, taken at 1kHz, with the lens cap on. Internally staring the recording via HCImageLive is on top, whereas externally triggering the start via TTL pulse is on bottom:
I have been troubleshooting this for a number of days now, and I am completely stumped as to what could be causing this. Some notes from my troubleshooting:
The effect is independent of the duration of the TTL pulse (1ms, 100ms, 5000ms all produce the same effect)
It is independent of recording from the full sensor or a sub-array
It is independent of how the images are stored (memory, disk, temporary buffer, etc.)
The effect is time-dependent, not frame-dependent (first ~2000 frames are affected when recording at 1kHz, first ~150 frames are affected when recording at 90Hz)
I am interested if anyone has plausible explanations, or of course, potential fixes. Eager to discuss with folks – thank you for your time!
This is not going to be a satisfying explanation, but perhaps a welcome, somewhat similar experience, with some possible corrections (if the problem is similar, and with a good deal of effort).
I’ve seen similar, triggering mode specific, charge accumulation artifacts in previous Hamamatsu models, which i documented here:
The accumulation curve looks similar, (but note the x axis in fig 5 there is not the same as time, more complicated setup). Also, if yours happens with lens cap it suggests a combination of charge accumulation paired with high dark current to begin with (which is also a bit surprising assuming the camera is normally cooled and exposures are moderately low)
If we can assume this is similar phenomenon, even through yours is a newer camera, you can see above for possible strategies in correction. But perhaps it’s unrelated in which case I wish you luck in finding the answer
I don’t have an answer either but I definitely noticed similar behaviour on our Fusion BTs as well (the camera was always triggered and I never compared with free running mode). In the end, for our application, the solution was to discard the first images. It is not ideal and not every experimennt allows it obviously.
Curious to hear what other people say about the subject.
@talley, thank you so much for the contribution! This does look strikingly similar to what I am seeing in my setup. Charge accumulation does seem like it could be the culprit here.
The next question that pops into my head, if that is the case, is why it would be trigger-dependent. Unless I am misunderstanding my triggering settings (which is entirely possible), using “start trigger mode” should use the same exposure/readout/reset settings as the internal start.
Like you mention, it is possible that it is an issue that is specific to my camera, such as unusually high dark current. I am in contact with an engineer from Hamamatsu and have shared some image sequences. Hopefully they will pick that up if that’s the case.
Thank you again for your suggestion! If this ends up being the case, I will follow up.
@Omnistic, thank you, it is good to know that I am not alone or going crazy!
Like you said, although it is not perfect, I was considering the same solution of discarding the first ~2s of recording. I am also considering just using the internal start and sending an output TTL pulse when the recording starts so that the signals from my patch amplifier and camera can be synchronized post-hoc.
If Hamamatsu comes back to me with useful info, I’ll post here.
in which case I wish you luck in finding the answer