Any chance it's time for a nap?

As I have stated many times before on this site, sleep is really the only thing that new parents want. They want their baby to sleep and they want to be able to sleep themselves. And they want both to happen in big 8-10 hour chunks. (We still suffer from a sleep deficit stretching back 8 months.) In my quest to understand and try to crack Trixie’s sleep code, I introduce yet another way to quantify her sleep habits: The Sleep Probability Spectrum.

Trixie's Sleep Probability over 123 days

The Sleep Probability chart is simply an average of Trixie’s daily Sleep Maps for a given set of days. The averages are translated to a gray-scale spectrum. The stronger the contrast, and the larger the pattern, the more predictable her sleep habits for the time period examined. It’s designed to collapse a range of time into 10-minute increments on a 24 hour scale in order to figure out if there are any patterns developing.

How is the Sleep Probability Created?

The probability of Trixie being asleep at a particular time of day for the given range of days is expressed as a gray-scale value. Solid black represents a 100% probability that she was asleep. Solid white means a 100% probability that she was awake. The gray areas represent different degrees of uncertainty regarding her sleep schedule. 50% gray indicates a complete lack of schedule. It means there was an equal (or completely random) chance of her being awake or asleep.

Evolution of Sleep Probability over the Last 5 Months

What kind of predictive value do these Probability Spectrums have? Unfortunately not as much as I’d I hoped – or at least I haven’t figured out a way to apply them yet. But they are great for looking back at her progress over the past five months. Of course, when you start at the bottom, there’s nowhere to go but up.

This entry was posted in Behavior, Metrics, Sleep. Bookmark the permalink.

6 Responses to Any chance it's time for a nap?

  1. schaff says:

    Long time listener, long time caller. Just want to say thanks for the new hard data. Like many of your fans, the telemetry means the most to me, and I look forward to the day when I can use your expert code to make my own site. I want to be able to track Trixie’s sleep and diaper use on my own site, so please let me know when that will be possible.

  2. John says:

    I’m glad to know Colum’s not the only one who’s got issues before 1 and after 3. What’s up with that? Does Trixie need soothing when she wakes in the middle of the night? Colum does about half the time.

    You makes charts; I write poems. Here’s my sleep poem.

  3. Kaleberg says:

    It looks like your baby has a 3 hour sleep wake cycle with an an evening peak. You really should do a spectral analysis (FFT) and get a real spectrum. Your chart is the probability distribution, not a spectrum in the signals analysis sense.

    I may be off on the scale, so I have to estimate things on your sleep probability evolution series, but it looks as if this three hour pattern was present, but noisier, in the past, but is now becoming clearer as Trixie is starting to sleep more regularly.

    If I remember, REM sleep cyles on a 45 minute basis, with REM deep sleep alternating with a shallower sleep stage characterized by increased digestive activity. 45 minutes is in resonance with 3 hours, so this is suggestive.

    Do you have a sleep/poop chart? I’m fascinated.

    Also, to get some more insight on Trixie’s emerging sleep patterns, you might make a two dimensional chart of sleep probabilities. Each horizontal line would be the derived sleep probability integrated over four (or whatever) days. This would be a lot like your five month series, except the charts would be thinner and stacked more tightly.

    Do you have the raw data on your site?

    Have you read that neat Chekov story “I Want To Sleep”?

  4. benmac says:

    Hi Kaleberg-

    I like the sound of spectral analysis. I’ll make a url with the raw data stream and email you. I have already been looking at stacked series of sleep probabilities but just didn’t pull it into this story. thanks for your perspective-

  5. Holy Moley, this is the most amazing thing I’ve seen, and not just in the six weeks since our kid was born.

    I just posted it to my new weblog for new dads, And I’ll get my spectroscopy-savvy wife at NASA to take a look as well. Anything to figure out the Mystery of Sleep.

  6. enoch choi says:

    we did similar stuff in excel, but nowhere as sophisticated:
    the charts for sleep aren’t displayed…