Triple mouse M Double glass window, no artificial dura Behaving in Simon's 2-mode detection task Visual expert 5/26/18 Objective at 12.5 PMT at 445 Task, starting audio trials only. Later set to 10% vision trials. At the end, collected some additional visual trials. Using normal LED "house lights" Imaging at vessel, anteriomedial of the window Depth 185 um Laser 33% Continuous imaging, signal looks good. Scanners at 2.5V. Movement seems ok. Chasing Z-drift: 3000 up 2 9000 up 2 12000 up 2 17000 up 2 21000 up 2 26000 up 2 28000 up 2 32000 down 5 seems like recording has very slow or no drift. will have to be careful to not over correct. Performance seems fine so far. ~75% audio after 100 trials. Shielding is good - no visible impact from visual stimulus. 40000 up 1 50000 up 1 Switching to vision after 400 performed trials to make sure we get enough non-trained responses. Should have a lot of audio responses by now. After 450 performed trials at switched to 50/50 audiovisual trials. Audio performance got worse towards the end of the session. Might leave some of those trials away. Overall great session. Audio ~75%, vision 50%, signal was good and stable and 500+ performed trials. =========================== 7/11/18 Objective at 6.2. Tried to adjust for tilt but now it is slightly angled in the other direction. Should be ok though. Both PMTs at 445 Task, starting audio trials only. Shielding is bad, no housing light, audio only. Some visual trials (~5) at the beginning. Laser at 35% depth is 175um Imaging at the anterior center of the window. Should match ALM based on coordinates / widefield. Imaging with green/red channel using D3228 texas red 3000MW as structural marker for vessels. Injected 50ul of 2.5% solution SC.Works pretty well - vessels are clearly visible. Signal is not amazing.See some neurons though. Had to quickly stop the acquisition at ~7500 collected frames. Restarted afterwards and continued streaming. Seems to be fine - written counter is going up accordingly. Hope this doesnt mess anything up. Based on structural channel, signal seems fairly stable. Will try to take a Z-stack at the end. Drift chase 55000 up 5 62000 slightly adjusted for xy drift. 65000 up 5 91000 up 10 Red channel has some honeycomb noise that appeared and disappeared sometime during the session. Behavior is ok. ~68% auditory in the end. Animal is not doing well anymore at ~370trials. Will keep recording until 400 performed trials and then stop. Manually recorded a frame stack of 40um. 20um above and below imaging plane in 2um steps. Saved in individual imaging files. =========================== 7/12/18 Objective at 8 - Should decrease angle for next recording. Angle is better in this but still not great. Shielding is not perfect but should be ok. Will do both audio and visual stimulation. House lights are on. Recording 2 channels again. Same as last time but D3228 concentration is 1.25% instead of 2.5%. Seems good enough to provide structural image. Changed sinusiodal correction from 159 to 160. Signal seems better now. Using 8 frame averaging to follow structural signal also helps. First imaging file contains a stack of 40um. 20um above and below imaging plane in 2um steps. Imaging depth is 200um. Lots of XY - movement at around frame 8000. Had to manualy move back to original plane. 35000: 10 up 45000: 6 down Hard to see the pupil in this one. Animal is squinting a lot at ~175 trials. Should change the camera angle soon... Recording quality seems to have improved over the course of the session. Weird :) 75000: up 4 105000: up 2 Behavior is ok: 68% audio. Switched to all visual after 465 performed trials. 130000: up 5 145000: down 10 Stopped at 600 performed trials. Should be a good recording. Seems like the Z-stack didnt get saved for some reason... Double check whether it is at the beginning of the current file 002. Don't think so though. Consider trying to find the same plane again later and take the stack then. Maybe it's close enough to be used anyway. Seems like the analog data was not saved as well so presumable this recording will be useless. Great... just great. If this recording is REALLY needed, it might be possible to identify visual stimuli by averaging whole frames in the red channel. There is a mild artefact from visual stimuli that you can use for alignment. Before getting into this, it might be better to just do more experiments though... Adjust objective angle to 6.9 after recording. =========================== 7/14/18 Objective at 6.9 - Angle looks better. Slightly touches edge of the implant but should be ok. Shielding is good. Will do both audio and visual stimulation. House lights are on. Right visual stimulus causes visible artefact but should be ok. Imaging depth is 225um. Using 50ul of D3228 at 1.25% for structural vessel signal. Looks good. Location should match ALM. Anterior center of the window. First file is a stack of imaging frames with 8frame averaging. +-20um around imaging plane in 2um steps. First frame is 20 um above sample, last frame is 20um below. Drift seems fine. 35000: 10 up Signal looks good, performance is not great. ~62% audio, chance for vision. Switched to all vision after 400 performed trials. Sometimes, there seems to be some bleedthrough of the green channel into the red when neurons are strongly activated. Stopped after 463 performed trials. =========================== 7/16/18 Angle seems a bit worse than usual and signal is much less good. Not clear what changed... Maybe the objective is touching the edge of the implant. Imaging depth is 390um. Shielding is good. Will do both audio and visual stimulation. House lights are on. Using 50ul of D3228 at 1.25% for structural vessel signal. Not much to see this time. Saw vessels at the surface but nothing deeper. Only recording Green channel and no stack in this one. SNR is much worse than in earler recordings. Laser power is 40%. Should measure laser later to check if something changed. Maybe implant quality has degraded. Location should match ALM. Really hard to see any drift in this one. It seems pretty stable. Behavior is not good. Almost at chance for audio and animal is not very motivated anymore after 300 trials. Overall not very good experiment :( It seems like there are much more cells active towards the end of the experiment. Don't think thats because of drift... Mabybe cholinergic modulation when the animal gets stressed? Strange... SO much more signal towards the end. Wonder if there was an air bubble under the objective. Happened on the 12th as well though. When going higher after the end of the recording, signal was overall much better. Presumably something changed in the setup - maybe there was a problem with air bubble. Check data quality, maybe don't use all the early trials. =========================== 7/16/18 Imaging depth is 360um. Location is close to last time, should match ALM. SNRA is not good. Can't see anything in the Red channel but record it anyway in case SNR gets better later. SNR is comparable to beginning of last session, so not good. See some cells though. Laser power is 40%. Drift seems fine. Shielding is ok, presenting both audio and visual stimuli. Performance is no good. Change for both. 108000 up 5 Switched to vision after 400 performed trials. Stopped after 500 performed trials. Detection at chance in both modalities. Red channel is presumably useless. Didnt take a stack for this one.