S1 | the sofi brain - initial findings

We have been pouring over the first bits of data from our different pioneer journeys with Valerian during the S1 pioneer program. With the initial build (S1), we have mainly been focusing on turning on various elements of the sofi system - a minimum lovable product. With our proof of concept now firmly underway (and many Pioneers being restocked), we will start turning our attention to improving individual elements of the sofi system. We are looking at the electronics (i.e. firmware, Bluetooth bridge), mechanical (the trigger, spray angle, bottle orientation), and digital (app flow aspects, sofi “brain”) in the next S2, S3 and S4 build ahead of sofi’s public release. One of the super exciting elements of the sofi system that has been activated (although very much in its basic form right now) is the “sofi brain”.

To understand how much Valerian affects your sleep, we effectively build mathematical models that learn to predict your sleep time and quality based on your use of the sofi pod (this is what we refer to as the “sofi brain”). This mathematical model is also given access to your sleep quality and time in the past (your baseline), just in case that’s helpful information for the sofi brain to consider while embarking on this learning adventure. If the model does sufficiently well, we then ask sofi what she looked at while trying to learn to predict your sleep time and quality. If she “decided” to ignore the spray data and focused on previous sleep patterns to predict future sleep, it means that no useful information could be extracted from the spray data as of yet. On the other hand, if the sofi brain “decided” to look at your spray data and has learned to predict your sleep time or quality well enough, your pod use directly affects your sleep!

We are considering sharing some of these initial findings with you here, and @alexwalkerjones will begin to ask individual Pioneers if they are happy for us to post what the initial sofi brain (let’s call it version 0.1!) was able to observe thus far. It would also be great to hear from each of the Pioneers to see how their experience compared to what the sofi brain could sense. We know we have a lot of work to do here, and we plan to share our progress on this and complete the work together. xx


Amazing insight sounds very complex can’t wait to get the results


Wow, that’s interesting :star_struck: absolutely impressive! Can’t wait to read more about this! :+1:


Wow! So happy to be a part of the pioneer programme! Can’t wait to see my own results. It looks fascinating :star_struck: Thankyou for this wonderful opportunity Sofi Team! And keep up the brilliant work :clap:


Sounds interesting but unfortunately for me my sleep patterns have been badly affected over the last week or so by a bout of flu. Even so this would not be an unusual occurrence for people to cope with and I would love to know if my data changed during this period.


Well I thought I should begin with sharing my data with the community here … as a Pioneer I was pretty good with my consistency but as I started to get to my third bottle (worried about running out ahead of our refill program) I started to use my sofi stick less and less often (had to save some for meetings as well to show how it worked) so my data over the third bottle was very inconsistent. You can see this reflected in my sofi score …

If you look at my spray data I was good with finishing one bottle than the next but then with the final one (N.02 - blue bottle) I was a bit less consistent …

When we ran the data through the model, @enricovarano made these observations from the graph below:

We observed that to learn to predict how much I had slept each night, the model generally looked at how much I slept over the previous 1-3 nights (that’s what y_t–x means: x is the number of days it looked back, left column of graphs). For my sleep quality, the model just couldn’t find a pattern but looked at spray data more closely (t–3, bottom right). We got close, so perhaps if I had been more consistent at journalling (even without the bottle) we would have been able to make more solid observations (or perhaps Passiflora will have more impact - I am enrolled in the S2 program so will find out!).

Now we have just finished compiling analysis of a few dozen pioneers so if you are up I will share the results here and would encourage each Pioneer to share their experience with the plant. Remember that the sofi brain only sees part of your experience and your qualitative feedback will be super important!

Next Pioneers we would like to share their initial journeys are @sonja.ekkers.77 , @dizzypuss , @jacquelinecannon1 , @Eevie , @tinaspain , @CjaJose , @chris.king5 , @jaxlou1982, @Aaron , @emmyginnis , @kaneary , @savanna , @lilywong16 , @kathleen.brown33 , @qudsiyahbano, @keith , @paul.ashton22 , @Ally … we are working on a next group after. But we will aim to share one a day and love your feedback. Happy Easter xx :pray:


I would love to see it but I know it works for me. While I was out sleep was very little and poorly. First night back on the spray :sleeping: slept like a baby :heart::heart: even on the spray that I didn’t feel worked for me the first time. I only had three sprays and I was out. :sleeping::sleeping::sleeping::sheep::sheep::sheep:


Nice to see you slept well :slight_smile:


I’m exactly the same when I run out my sleep definitely changed then top up came and very quickly my sleep resumed I have tried to keep journals so hopefully there is usuable data


Back to sleeping. COVID free and only two sheep left to lamb . What’s not to be happy about :heart_eyes::heart_eyes:

And I hate selfies :rofl::rofl::rofl: but taught I show you all who the crazy sheep lady is :rofl::rofl::rofl:I’m the one on the left in the picture :joy::joy: bell on the right. The top picture is forest gump :heart:


So lovely to meet you via selfies though it feels like we have met for some time now! See Forest and Belle are looking nice and happy! Hope you are having an excellent Easter. On the data that we will share is that we need to develop and adjust our sofi models to reflect now you feel. So every time there is a gap is an amazing opportunity for @enricovarano and team to begin to tweak and improve the sofi model! Will begin sharing more on the first Pioneers tomorrow. In the meantime wishing all a wonderful and restful break and Easter xx :pray:


Interesting findings! Here is my experience (sorry it’s a bit long! I wanted to be thorough). I’m 23 and struggle with anxiety in the form of worrying, overthinking, and not being able to quiet down my brain any time throughout the day or night, which then leads to sleeplessness. I was so excited for sofi and valerian to be able to help in this area, but unfortunately I haven’t noticed any difference.

My data may come up differently and will likely show an increase in better sleep recently compared to a month ago, but this is because I’ve had to take additional natural sleep supplements to help me, or else I just wouldn’t be able to sleep at all! FYI, my baseline was also recorded while taking additional supplements. The only difference I’ve noticed in my sleep has been through switching around these supplements. When I first started sofi I was taking certain ones including magnesium and melatonin (knowing that these two alone aren’t enough to help me sleep), and purposely leaving out something like GABA which has helped me in the past, because I wanted to see how I would fare with the addition of sofi. I didn’t see any positive improvements, even after bumping up my sprays to 4x each after dinner and before sleep. I paid close attention and even on extra anxious days, I unfortunately never felt sofi helping me calm down. I was able to fall asleep fairly quickly due to the melatonin, but had a pattern of waking up during the night and not being able to fall back asleep for hours.

At first I thought sofi was the thing helping me fall asleep more quickly in the evening, but for a few days I skipped magnesium and mostly just relied on sofi, and it ended up taking me a couple long hours to wind down instead of my usual half an hour or so. Recently I’ve added back GABA and have been sleeping well for the most part. Every time I switched my additional sleep supplements, I was very mindful of noticing effects, and was also ironically extra anxious because I didn’t want to mess up the sofi testing. I know my data isn’t as black and white as other testers because I had to take extra supplements, and I’ve worried about this every day since the start, but am hoping it’s still useful somehow!

All this to say, I’ve played around with additional (and lack of) supplements and sofi dosing, paying close attention to any changes and effects, but have found that valerian isn’t the right fit for me right now. It was fun and interesting to try and I’m grateful for the experience. I’ve had some positive results with passionflower in the past, so am really looking forward to trying it again in S2!


This is very thorough and excellent feedback as you suggest there is so much more going on to be able to isolate effect would be difficult without your help. Really keen to see if the different pathways than passiflora will rather in the s2 program will make help you observe any differences. Also adding @pamelaspence here as well as I certainly have come across some literature in relation to the other supplements and keen to hear what her experience has been of these in combination with Valerian. Will look into your data and share it tomorrow see if we can get a better picture together … as always super grateful for being on this journey with us and can’t wait to capture a positive signal with one of the plants on the roadmap! We issuing you a lovely and restful weekend and Easter. xx :pray:


Ps. Welcome to the treehouse! What an awesome first contribution … we will work hard to be worthy of it! :pray:


I thought was very consistent in my journaling and sprays but my data score does not reflect that @ only 53% :pensive:. That frustrates me as I’m very competitive with myself… and if there was a way to correct/edit the spraying I wonder if that data score would reflect better on me - or if it would allow me to notice a more positive outcome with the valerian :thinking: I had a lot of trouble getting the sprays to register but felt I did the journaling consistently through the entire 3 pod journey. I chose not to continue with the sprays after the 3rd pod to see if stopping the valerian will be noticeable since I only felt a noticeable difference at the beginning of each pod (placebo? Concentration level? :woman_shrugging:t2:). So far since I’ve stopped the valerian I’ve had to resume melatonin to get to sleep but am waking and unable to get back to sleep. Really hoping the Passiflora will make a noticeable difference for me!


Glad I didn’t overwhelm you with my feedback, haha! Thanks for your kind responses!


It’s not just you @kaneary … looks like the first iteration of the sofi brain was a tough marker and gave even the best pioneers a maximum score of 78 … now in the UK that’s a first … or a distinction … but in the US that’s not even an A! Look at these different sofi scores beginning with mine … these are all amazing Pioneers but no one seemed to be able to get to that 90 score that we love to see everyone who is consistent getting to, most plateauing in the 70s …

So with the help of @enricovarano we are going to recalibrate the sofi brain a little bit so it adjusts our marks a little bit higher! xx :pray:


Hi @savanna and welcome to the treehouse!

Thanks so much for such a detailed log of your experiences with sofi and the other supplements so far. That’s really helpful. I’m sorry that valerian is not ‘your herb’ but I think that is most likely the case and we absolutely know that one size does NOT fit all - in fact it’s a hashtag I use a lot on my social media accounts!

Clearly at the moment the magnesium is helping you with your wind down phase (which makes perfect sense as it is a muscle relaxant as well as having a calming effect on the brain) and the melatonin is able to act as the ‘sleep switch’ which is basically how I think of it.

Will be interested to see how you get on with Passiflora (and all the other plants we have lined up to follow after!) to see which one best fits your situation.

Keep feeding back - it is so useful!

Thanks and hope that was helpful!

Pamela :herb:


Hi Pamela, thank you for your response! :slight_smile: Do you happen to know why some supplements work well for me certain nights and not at all on other nights? Sometimes I have to keep taking different natural remedies all throughout the night in order to go back to sleep, it’s crazy. It’s like they’re bouncing off me and just doing nothing to quiet down my brain. The previous night I might have been fine, and I can’t work out why this is.

A couple other questions that maybe someone else on the team is better suited to answer? I have another member of my household who is interested in joining sofi for S2.

  1. Is this allowed in the program, since we live in the same house?
  2. He deals with poor sleep issues as well, and likely wouldn’t be able to sleep solely with sofi’s passionflora. He takes other sleep aids as well, like melatonin, but still struggles. Is this additional supplementation (like I used) allowed, as long as he establishes his baseline while still taking these? Then he can see how his sleep improves with the addition of sofi on top of the other supplements.

I want you to have accurate data and know it would be simpler to see sofi’s effects without these additional supplements — but on the other hand, it might be interesting to study this aspect if you eventually introduce sofi to public consumers, who potentially take other sleep aids as well. Please let me know your thoughts on this and if my data was helpful, knowing the other factors! :smiley:


Hi Savanna,

Unfortunately why some things work sometimes and not others is one of the big mysteries! What I would say is that if your routine doesn’t help you most nights then you haven’t yet found the solution.

Very much hope we can help you with that! :pray:

As for your other questions, I’m tagging @alexwalkerjones and @Kaveh who may be able to help :herb: