Masking Resonant Scanning Noise/Measuring It/We Don't Use Good Tone Frequencies?

[jmdelahanty]

The resonant scanner, operating at a fixed frequency of 8kHz (I'm pretty sure...) has a decent amount of noise that has a somewhat high pitch. After talking with Meenakshi briefly the other day, she mentioned that in other labs she's been in/seen they would not only use different tones (2kHz is apparently quite hard for mice to hear unless its really loud!) but they would try and mask the sound of the resonant scanning noise as well. This was in an auditory specific lab, but it's probably best to try and control those things too or at least measure them/know about them if you can.

Having a microphone or something in place that can measure this over time or at least some way of calibrating things within the box would be a good thing to do probably if only for our own understanding of the environment we're putting the mice in. Meenakshi mentioned that the tone produced by the scanning sounded something like 8kHz or so which is also super close to the 9kHz frequency used for signaling sucrose coming. This is another thing that could potentially confuse the mouse somewhat while they're being imaged since it's likely they can hear it.

Another interesting thing would be we could collect USVs while the mouse is under the scope and that would just be something kinda neat in general.

[jmdelahanty]

First helpful link illustrating this here.

Second, here's what Meenakshi's colleague, Kameron Clayton, PhD sent me:

We’ve covered our resonance scanner in acoustic foam, which does attenuate the 8 kHz tone substantially. This is all you can do short of an active cancellation strategy, which I think may be overkill.

There are a couple biological factors that work in your favor here: the brain (specifically auditory cortex) tends not to care about stationary sounds. In fact, I would hazard that you could barely decode whether a continuous 8 kHz tone was present or absent if you looked at spike rates in auditory cortex a few minutes after the sound is turned on. This is fairly consistent with behavior as well (with an 11.3 or 12 kHz tone, we don’t really see a difference between sound detection thresholds with 2P or behavioral only.

I would suggest considering different frequencies if at all possible though, for the reason Meenakshi mentioned. 2 kHz and to a lesser extent 9 kHz are to the far low extreme of mouse hearing. As such, you are giving the mice 2 cues: 1) loudness and 2) frequency, because a 2 kHz sound at the same sound pressure level as a 9 kHz tone will actually sound softer (ask Meenakshi about sensation level). I would recommend shifting the frequencies up. 12 kHz has worked well for us and avoids any concerns of high frequency hearing loss you can get when working with C57 mice that are 12+weeks old.

While I have heard this 8kHz tone since I started messing around with this scope and therefore honestly didn't think much of it, Austin heard it for the first time today.

Some of the mice that are performing very well in the training boxes aren't performing very well under the scope. The thoughts he has are:

• New context under scope, mice are scared
• Room is quite loud, mice are scared
• 8kHz tone is similar to 9kHz tone and confusing the mice

So finding a way to mask this scanning noise is a high priority.

[jmdelahanty]

Another thing: If Luca Godenzi PhD arrives in the lab he'll want to do USV recordings I think so would be good to measure this kind of thing...

[jmdelahanty]

I talked with postdoctoral fellow Teja Bollu, PhD from Dr. Martyn Goulding's lab briefly today about some things and he suggested the following (paraphrasing):

The 8kHz and 9kHz tone is probably confusing them, or at least it's not helping. Best things to do would be to both change the frequency to something like 20kHz and, to accomplish that successfully, both:

• Don't use the piezo speakers we have (which in the future we won't be)
• Generate the tones off a class D amplifier (which in the future we will be)

I also chatted with Deryn briefly about this and here's what she thought/said about the tones in use/frequencies (Paraphrasing):
She's altered the decibel levels for the HF boxes and the NLW scope to be in line with what helps the mice discriminate between each frequency.

I (Jeremy) told her that we have the speaker going about as loud as it can go without it corrupting the piezo speaker too badly on the older Bruker. She mentioned this (paraphrasing):
If it's that loud the tone can act as a negative stimuli and also damaging the mice hearing. It's good to know that the Bruker produces an 8kHz noise with the scanner but the speaker is less than a foot away from the mice. It's likely more of an issue with the noise level or the hearing of the mice than the frequency since 9kHz is a common tone used in behavior.

I then told her that Austin had asked members of Dr. Takaki Komiyama's lab about this, they said something along the lines that it might not be the only reason the mice are struggling under the scope but it's definitely not helping. She then mentioned (paraphrasing again):
It's not ideal, but could be something that's animal specific and not the task itself.

Tomorrow morning I'm going to finish a branch that's for running just a voltage recording (the previously removed behavior_only flag (I hope). Austin is going to use another mouse for testing tomorrow without the scanning going to see if their behavior improves. If it does, then we will have solved at least part of the mystery. If not, then who knows what's happening...

[jmdelahanty]

Many of the mice are running well under the Bruker scope now that the speaker is even closer than before and (I'm uncertain this is a good thing...) the speaker is literally as loud as it can be. The risk of hurting the mice ears/it being a negative stimulus is something I'm concerned still about...