Mice take more pellets from FED3 during free-feeding vs. FR1 modes

[Lex]

I've noticed something interesting in our data, which is that mice remove ~20% more pellets from FED3 in free feeding mode vs. FR1 mode. We've also gotten data from two other researchers that were consistent with this (average chronogram from all 3 labs below). 

We have noted that pellet "hoarding" (removing pellets without eating them) occurs during free-feeding and not during FR1, which may account for some of the difference.  It's also possible that mice do in fact eat more in free-feeding mode, or eat pellets less efficiently in free-feeding mode such that they create more dust and don't get as many calories from each pellet. 

At any rate, I'm curious if anyone else has data on this point that they would like to share to get a more robust answer on this point.  If you have raw FED3 datafiles on free-feeding and FR1 from the same mice that you would like to share please reach out to me here or via direct email! I plan to share the results of this analysis here in this forum, and we can also discuss authorships on a potential publication for any labs that share data.  

Thanks!  

-Lex

[Matias Andina]

Hi Lex,
Really curious about this! My first impression is that free feeding is ad-lib and ad-lib mice gain weight. So the FR1 might be a small "tax" to keep them at bay. 

Not sure if it makes sense though....do FR5 mice eat the same as FR1? Is there any literature on FR1 mice gaining less weight than ad-lib mice? I would be willing to go down this rabbit hole with others.
We have been running animals in free feeding for a long time now and haven't noticed substantial hoarding (yet again we can't really compare with FR1 because we don't use it). 

When we see pellets in the cage, it's more of a "that weird animal" phenomenon. We have noticed that the hoarding extends to the area around the FED3, so we have not seen that mice have a special hiding place for pellets. I think it might be related to mice accidentally (?) dropping the pellets. Because we use ITI of 3 seconds, maybe mice that drop pellets go for the pellet already available in the well instead of the one in the floor (have no hard data about this). 

I know it is tough to determine, but maybe repeating FreeFeed vs FR1 with a longer ITI, say ITI = 20 sec, can change behavior?
Another thing that got into my mind is that we have different models of prints (some older, more open design, some newer with the well that is more closed). Do you have any info on whether a particular design is better/worse?
Best,
Matias

--
You received this message because you are subscribed to the Google Groups "FEDforum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fedforum+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAPM2s1TYABSTpbKBm_348fBaq5%2BHGX4wWfYnuztqXhi%3DuQfCeQ%40mail.gmail.com.

[Lex]

Thanks for your thoughts Matias!  I've now seen 3 (actually 4 if I count two datasets from our lab) datasets that all show the same thing - mice take ~20% more pellets from FED3 when it is a free-feeding session than an FR1 session (does anyone else on here have data that speaks to this - please email me!).  However, I haven't noticed mice gaining more weight on free-feeding, although it's tough to see weight differences on grain pellets. 

Despite taking them from FED3, I don't think free-feeding mice are *consuming* 20% more pellets than FR1 mice - I think the difference comes from a combination of 1) eating more (possibly 5-10%?), 2) taking pellets out of FED3 without eating them (which we call hoarding), and 3) grinding pellets differently between the two modes, such that they may eat more efficiently in FR1 and make more dust or drop half-eaten pellets more often with free feeding.  I found this cool paper on grinding from John Speakman's lab that suggests the grinding can account for food intake measurement errors of up to ~30%.

We have quantified "hoarding" (leftover pellets) in one group of 10 mice - and we noticed it only happens in free-feeding mode.  Here is a histogram of daily pellets left over from free-feeding vs. FR1 sessions (same mice).  We also noticed the "strange mouse" thing, where most of the high numbers come from the same mouse - we even counted 125 pellets in the bottom of the cage one day!  But once we moved the mice to FR1 this behavior stopped completely. 

We also see pellets dropped mostly around the FED3 device - so maybe "hoarding" isn't the best word for it because that implies they're intentionally saving them up for later.  But this might reflect them just being more careless during free feeding because they know they have a unlimited food source.  Or perhaps the FR1 group does the same amount of pellet dropping but ends up eating them later on instead of doing a nose poke?

Curious if you or anyone else on here has any more thoughts or data on this! 

[Lindsey Czarnecki]

This is super interesting and I have similar observations as Matias.  Not all of our free feed animals "hoard" and those that do don't tend to do it everyday.  It also seemed to occur more often within the first week(ish) of them using the devices.  When they do hoard, I also observe the pellets to be close to the magazine.  I have also noticed a considerable amount of dust when this happens, so thank you for sharing the paper on grinding!  I will keep tabs on this thread and update while keeping a closer eye on my free feeding cohort. I'm not sure if in my current configuration I can set up a camera, but maybe I can do that at the end of my experiment and/or switch them to FR1 to see what happens.

[Kristen Delevich]

Hi all,

Just following up on this thread that we've observed something similar by analyzing grams of pellets consumed normalized to bodyweight across different task modes (free feeding, FR1, and progressive ratio). When we analyze the data this way, it's interesting to see that age and gonadal status have different effects on food intake, for instance, gonadectomy decreases effortful food intake but not under free feeding conditions! We have not carefully quantified food hoarding or grinding behavior yet.

-Kristen

https://www.sciencedirect.com/science/article/pii/S0018506X2300048X

--
You received this message because you are subscribed to the Google Groups "FEDforum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fedforum+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/4f046c3d-121b-4b6d-8258-1ba0927b74efn%40googlegroups.com.

--

Kristen Delevich, PhD

UC Berkeley

[Zane Andrews]

Hi All - just adding my 2 cents here as well. 

In response to Matias’ question

“Not sure if it makes sense though....do FR5 mice eat the same as FR1? Is there any literature on FR1 mice gaining less weight than ad-lib mice? I would be willing to go down this rabbit hole with others.“

We looked at this and mice with free access to FEDs for ~3 days at FR5 eat less than mice under same conditions at FR1. So if you make it harder collect a pellet they will eat less.

Cheers

Zane

Sent from my iPhone

On 30 Nov 2023, at 8:50 am, Kristen Delevich <kde...@gmail.com> wrote:



Hi all,

Just following up on this thread that we've observed something similar by analyzing grams of pellets consumed normalized to bodyweight across different task modes (free feeding, FR1, and progressive ratio). When we analyze the data this way, it's interesting to see that age and gonadal status have different effects on food intake, for instance, gonadectomy decreases effortful food intake but not under free feeding conditions! We have not carefully quantified food hoarding or grinding behavior yet.

-Kristen

https://www.sciencedirect.com/science/article/pii/S0018506X2300048X

<image.png>

On Tue, Nov 28, 2023 at 5:45 AM Lindsey Czarnecki <linds...@gmail.com> wrote:

This is super interesting and I have similar observations as Matias.  Not all of our free feed animals "hoard" and those that do don't tend to do it everyday.  It also seemed to occur more often within the first week(ish) of them using the devices.  When they do hoard, I also observe the pellets to be close to the magazine.  I have also noticed a considerable amount of dust when this happens, so thank you for sharing the paper on grinding!  I will keep tabs on this thread and update while keeping a closer eye on my free feeding cohort. I'm not sure if in my current configuration I can set up a camera, but maybe I can do that at the end of my experiment and/or switch them to FR1 to see what happens.

On Monday, November 27, 2023 at 10:50:45 PM UTC-5 lex.k...@gmail.com wrote:

Thanks for your thoughts Matias!  I've now seen 3 (actually 4 if I count two datasets from our lab) datasets that all show the same thing - mice take ~20% more pellets from FED3 when it is a free-feeding session than an FR1 session (does anyone else on here have data that speaks to this - please email me!).  However, I haven't noticed mice gaining more weight on free-feeding, although it's tough to see weight differences on grain pellets. 

Despite taking them from FED3, I don't think free-feeding mice are *consuming* 20% more pellets than FR1 mice - I think the difference comes from a combination of 1) eating more (possibly 5-10%?), 2) taking pellets out of FED3 without eating them (which we call hoarding), and 3) grinding pellets differently between the two modes, such that they may eat more efficiently in FR1 and make more dust or drop half-eaten pellets more often with free feeding.  I found this cool paper on grinding from John Speakman's lab that suggests the grinding can account for food intake measurement errors of up to ~30%.

We have quantified "hoarding" (leftover pellets) in one group of 10 mice - and we noticed it only happens in free-feeding mode.  Here is a histogram of daily pellets left over from free-feeding vs. FR1 sessions (same mice).  We also noticed the "strange mouse" thing, where most of the high numbers come from the same mouse - we even counted 125 pellets in the bottom of the cage one day!  But once we moved the mice to FR1 this behavior stopped completely. 

We also see pellets dropped mostly around the FED3 device - so maybe "hoarding" isn't the best word for it because that implies they're intentionally saving them up for later.  But this might reflect them just being more careless during free feeding because they know they have a unlimited food source.  Or perhaps the FR1 group does the same amount of pellet dropping but ends up eating them later on instead of doing a nose poke?

Curious if you or anyone else on here has any more thoughts or data on this! 

--
You received this message because you are subscribed to the Google Groups "FEDforum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fedforum+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/4f046c3d-121b-4b6d-8258-1ba0927b74efn%40googlegroups.com.

--

Kristen Delevich, PhD

UC Berkeley

--
You received this message because you are subscribed to the Google Groups "FEDforum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fedforum+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAGqp2FY%2BKQzRcAFQK6iEAqsykWzdyq5cNMvrCojpzd-WPiBypA%40mail.gmail.com.

[James McCutcheon]

Hi everybody

Hamid from our lab has provided Lex with data showing a similar finding re: FF eating more than FR1. The data will be included in the post and/or publication that Lex is preparing. Hamid also collected data on "hoarding" by counting the pellets in the bedding each day.

On the subject of "hoarding", which we've been researching in preparation for a more systematic study, literature on hoarding in hamsters shows that there is an inverted U relationship with foraging effort (Day & Bartness, 2003; https://pubmed.ncbi.nlm.nih.gov/12782221/). So at low efforts (wheel running in this case) you see a low amount of hoarding but when the foraging effort increases, hamsters begin to hoard more. Then it drops off as the effort becomes more intense. We're interested in whether we will see the same thing in mice with access to FEDs.

Thought this might be interesting. There are a lot of cool papers on hoarding in Siberian hamsters and the factors that influence it.

Jaime

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/5EF440F4-ECE3-4FB9-84DF-02A5FAA61017%40monash.edu.

[Alcantara, Ivan]

Hi all,

I’ve also seen mice hoarding pellets on free-feeding mode even after food deprivation (or activation of hunger neurons). I haven’t compared it to FR1 but I can try, unless someone has already done an experiment like this.

Ivan

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAEJTH%2B7A9TVhcn3V3KSiqennB2y4x5hFLN9C1WUS9QhJD%2BGxeQ%40mail.gmail.com.

[Lex Kravitz]

Hi Ivan!  I'm pretty sure they won't hoard with FR1 when hungry - we run these overnight in closed economy so they also get hungry at night and still don't hoard in our FR1 session. 

It's really interesting that they still hoard when you make them hungry - I had thought that perhaps on free feeding mode mice might grab pellets when they aren't hungry just because it's something to do, and then they drop them instead of eating them.  But it sounds like they still do this even when they are hungry, so it's not fully explained by non-hunger boredom behavior.

One aspect I don't understand is why on free-feeding they wouldn't just eat the pellets in front of the FED that they dropped, which seems easier than taking a new one out of FED.  But maybe they recognize that the floor is dirty or something so they'd prefer new pellets over previously discarded ones?  I'd act the same in that situation :)

[James McCutcheon]

Hi again,

One of the main drivers of hoarding shown in both hamsters and rats is food deprivation so if the food is readily available it didn't surprise me that you see more hosting under food restriction or activation of hunger neurons.

They also showed involvement of agrp, not, and leptin in the hamsters if I'm remembering correctly.

(On my phone so don't have access to the refs but most were from Tim Bartness's group)

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/e8033116-4ca2-49c2-bbeb-54cb8f7b9470n%40googlegroups.com.

[Mason Barrett]

Hi all,

I am a technician in Lex's lab, and I wanted to share some interesting findings from an analysis I conducted on free feeding versus FR1 data that other fed users sent us. The analysis consistently showed a significant effect where mice take more pellets during free feeding compared to FR1 schedules. While it's not entirely certain that all pellets are being consumed (as mice are known to hoard them), the data is clear that they collect more pellets during this feeding regimen. This data includes 212 FED files of 106 mice that were ran through Free and FR1 sessions. The data has been filtered to show the last 24 hours of the free session and the first 24 of the FR1 session. Thank you to all the labs that contributed data to this: Fobbs, Kravitz, Creed, Delevich, McCutcheon, Sutton, Maloney, Andrews!

I'd be interested to hear your thoughts or experiences on this result.

Mason Barrett

[Zane Andrews]

Nice work Mason, it’s a pretty clear effect. Does the data look essentially the same if you’re a little looser on the filtering? Say if you averaged the 24 hour response over 3 days for both Free feed and FR1? - I guess it might depend on data availability?  Put two in a cage - one on free feeding and one of FR1 and see what happens? Just for the hell of it….😂

It would actually be interesting to get a poll going to see how many people use free feeding vs FR1 for normal home cage feeding. We typically havent used the FED3s for long term feeding experiments, more for short term feeding behaviours and/or coupled with photometry and opto.

Z

-- 
Zane B. Andrews, PhD
Professor, Department of Physiology

Monash Biomedicine Discovery Institute
Deputy Head, Metabolism, Diabetes and Obesity Program
Deputy Editor - Endocrinology (The Endocrine Society; Oxford Academic Journals)
President - Hypothalamic Neuroscience and Neuroendocrinology Australasia (HNNA)

Monash University
Rm F215, Building 13F 
26 Innovation Walk 
Victoria 3800
Australia

T: +61 3 9905 8165 
M: +61416 929 263
E: zane.a...@monash.edu
W:https://www.monash.edu/discovery-institute/andrews-lab/home

On 11 May 2024, at 8:54 am, Mason Barrett <masb...@gmail.com> wrote:

Hi all,

I am a technician in Lex's lab, and I wanted to share some interesting findings from an analysis I conducted on free feeding versus FR1 data that other fed users sent us. The analysis consistently showed a significant effect where mice take more pellets during free feeding compared to FR1 schedules. While it's not entirely certain that all pellets are being consumed (as mice are known to hoard them), the data is clear that they collect more pellets during this feeding regimen. This data includes 212 FED files of 106 mice that were ran through Free and FR1 sessions. The data has been filtered to show the last 24 hours of the free session and the first 24 of the FR1 session. Thank you to all the labs that contributed data to this: Fobbs, Kravitz, Creed, Delevich, McCutcheon, Sutton, Maloney, Andrews!

I'd be interested to hear your thoughts or experiences on this result.

Mason Barrett

<Study Sites.jpg><Pellet Count.jpg>

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/f68c85a0-40b9-4f22-b101-965c6a2bba4fn%40googlegroups.com.
<Study Sites.jpg><Pellet Count.jpg>

[Andrew Hardaway]

This is a very clear effect Mason! Thank you for sharing it. My lab is gravitating towards FR1 for most experiments to avoid the hoarding confound entirely. When you're trying to run an assay in both males and female mice and have a lack of transgene expression or bad optical fiber placement already as an issue that could lead to animal exclusion from a group, it's nice to eliminate the stochastic hoarding issue. We've seen some impressive hoarders too.

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/018D6FE9-69D2-4263-9880-F6DD9D4D6961%40monash.edu.

--

J, Andrew Hardaway
Assistant Professor

Department of Psychiatry & Behavioral Neurobiology
University of Alabama at Birmingham School of Medicine

he/him/his

[Tommy Kim]

Hello all,

Thank you so much for sharing! Hoarding has been a fundamental issue that we have also been facing as we try to decrease attrition rates in our experiments.

To this end, we are wondering if anybody has directly compared the microstructure of food intake between free feeding and FR1. This thread has helped us a lot, but we are curious to know if any other eating patterns besides volume are affected by the mode change.

Best,

Tommy Kim

[Lex]

Hi Tommy,

I'm not sure if this helps you but we have a lot of data from adult mice doing Free feeding or FR1 feeding.  This is the same data collected from 10 labs in the other thread BTW, so ~106 adult mice (a mix of both sexes) that did free and FR1.  These mice take (and likely eat) fewer pellets on FR1 vs. Free feeding. The inter-pellet interval graphs also show that the modal time between pellets extends from an average of ~24s between pellets for free feeding to ~42 pellets for FR1.  The poke action in FR1 takes a few seconds to complete but not 18s. So adding the additional nose-poke seems to slow down the feeding within bouts. 

I know you're interested in young mice but hopefully this helps!  If you want the raw data to analyze yourself let me know.  Best, -Lex

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/a678a473-6e56-4d9f-8d3b-29952546542an%40googlegroups.com.

[Lindsay Matthews]

Hi Lex

Nice data!

I'm curious about the data showing pellets delivered with an interpellet interval of about 0.03 minutes. 

I can imagine that this could occur on a free feeding schedule if the mice were scooping out successive deliveries without eating them.  

But an interpellet interval of about 0.03 sec seems a little short on a FR1 schedule....any thoughts?

best

Lindsay

--

--

Lindsay Matthews PhD

Editorial Board Animals

Associate Editor Frontiers in Veterinary science

Associate Editor NZ J Zoology

https://matthewsresearchinternational.com/

Matthews Research International LP

Animal behaviour, welfare and learning specialists

Matthews Research International Deutschland UG

Spezialisten für Tierverhalten, Tierschutz und Lernen

University of Auckland (Honorary Academic)

E: lin...@matthewsresearchinternational.com

T: +64 (0) 22 422 1510

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAPM2s1SYwDcP0vq_6vpfoqRO7rjuuc7cy8jHjSEhMh9obc6h8g%40mail.gmail.com.

[James McCutcheon]

Late response for me, but massive thanks Mason for putting all these data together!

With respect to meal microstructure, we (well mostly Hamid, a PhD student in my lab) have been looking into it in more detail recently and definitely find that average meal size appears to be a lot smaller when mice on are on FR1 mode. We are trying various parameters to distinguish between meals other than our standard ones, for example, looking at differently sized meals (i.e., meals vs. "snacks" vs. "feasts") and using different inter-pellet intervals as threshold.

If anybody is at FENS in Vienna then Hamid will have a poster there and I'm sure would love to chat about this more.

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAM9LoGO9QuhaukNVo2cg--NEdb4BMRfuO5a_3cUP-htVqkisKQ%40mail.gmail.com.

[Hamid Taghipourbibalan]

Hi Tommy,
I am replying in this thread, in connection with your other email,
I just took a quick look at the data I am dealing with at the moment and plotted some parameters in 10 Control Male mice chow-fed ad lib / 8 weeks old/ C57BL/6NRj / 24h session time FF and 24h FR.

, figures not polished.
A meal is defined as interpellet interval >60 sec and the minimum number of pellets is 2 before it is logged as a meal. 
As Lex mentioned and expected, they take fewer pellets on FR1 mode. Meal frequency drops on FR1 mode(Match with the KDE plot I believe), but to my surprise in this set of data,  meal size is a bit larger on FR1 mode (however this is not a clean set of data, i.e. mice could get a 50-50 mixture of 20% Casein and 5% Casein diet(normal vs low-protein diet), so they might try again to get the next pellet which may or may not be a 20% casein pellet).

So all in all, apparently taking food from the countertop vs opening the fridge door is somehow changing feeding behaviour.

Hamid Taghipourbibalan

Ph.D. Fellow in Behavioral Neuroscience, Mccutcheonlab, UiT, Tromsø, Norway

Contact:+4746566050

Skype ID: htbibalan_1

LINKEDIN

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/a678a473-6e56-4d9f-8d3b-29952546542an%40googlegroups.com.

[Lex]

Lindsay, there are a few events registered that fast (~2s between pellets) in both Free and FR1.  But overall it's rare: In our dataset of 106 mice ~2.5% of the pellets taken during Free had an inter-pellet interval of ~2s, vs. ~0.4% in FR1.  

For Free feeding these events with a ~2s inter-pellet interval can reflect instances where the mouse removes a pellet, drops it, and immediately removes another one.  I could also imagine a mouse removing a pellet briefly from the well as they grab it, FED registering that removal, and then the mouse dropping it right back in the well and picking up the same pellet again.  That would result in a pellet being double counted with a short duration between pellets.  Unfortunately we don't have any way in the FED hardware to detect if that's what happened, vs. the mouse taking a pellet, dropping it, and picking up a new pellet.  But overall it's a rare thing, one could filter these events out if they wanted to be conservative and assume that mice were not actually eating 2 pellets that quickly. 

[Lindsay Matthews]

Thanks Lex, sounds good to me.

[Lex]

"taking food from the countertop vs opening the fridge door is somehow changing feeding behaviour."  Indeed :)  

If we look at meal size in the large dataset of 106 mice we see the opposite from what you saw with respect to meal size.  The counting of "meals" is dependent on definitions, but with a minimum of 3 pellets to count as a meal and a minimum duration of 2min between meals, here's the histogram of meal sizes between Free and FR1.  Note the longer tail of meals >15 pellets that only occur in the Free condition.  Again, these are the same mice tested between the two conditions:

If I plot metrics like average meal size and # of meals/hour I also see lower values for both in the FR1 sessions:

Fun times :) 

-Lex

To view this discussion on the web visit https://groups.google.com/d/msgid/fedforum/CAPU4b_tNqdbb4ocV1BeA0Ot_LTDed68mcEXaDOSCEn3B7jLRJA%40mail.gmail.com.