Monday, September 24, 2012


Quantified Self Conference Notes

I attended the Quantified Self Conference 2012 last weekend, it was a total blast! I learned a lot, and I have a huge list of todo's from the thing. Small flickr set. I'll post my complete notes below, but some serious highlights for me: And so, so much more, I learned a lot! My full notes (sorry they may be kind of impenetrable, they are notes I made for me):
Quantified self conference, Sept 15th, 2012, Day 1

Coffee is protective against depression?
Minutes of meditation per day - good proxy for whether he is having a good day.

When you skip a day, or more, you loose confidence - academically called "loss of self efficecy"
Gary plotted inverse of meditation plot - days missed, longest skip, per month: turns out his skipping is seasonal!  Now he doesn't loose confidence!

What is a computer for? The answer has been changing, it's now very personal!

It's not the bigness of the data that matters, it the ourness of the data.  The outliers are the most interesting bits, they are not mistakes!

What if it's data all the way in? Data all the way down.

Technologies for mindfulness, Nancy Dougherty.
Tracking your emotions really changes them!  Aka, mindfulness really works.
She build an EMG device that activates a string of LEDs when she smiles - real smiles, she put the electrode up by her eyes.
"help sparkle me out of whatever funk I'm in"
She wants her sparkles to make her more like her dog - bringing happiness to herself and others everytime she meets someone.

Dave Asprey, talking about using software to exercise your mind, grow your focus
Dual n-back, he claims 10 iq points if you can stick with it, but super frustrating!
Calmness generates better thinking, recommends heart math 
Some claims that you should NOT use chunking on dual n-back, be more intuitive
He claims the benefit persists even if you stop doing it.  But it will make you super angry while you are doing it.  You need to let go in order to do well.
He also recommends feedback as you play, rather than only getting feedback at the end.
All of the things that improved his IQ he found super difficult and emotional - something about this stuff really pushes your buttons.
Active vs passive mode of your brain - more like a slider switch, how can we learn to better remember what our brain does in passive mode?  He has a counting app to be used while blindfolded with audio that forces your brain into alpha mode.
I3mindware - fun but doesn't think it works very well.  Brain workshop is the open source one that is good.
He recommends walking meditation, breathing, increasing your heart rate variability.
He is writing another book, partially about "threat response" stuff in your brain.
The heart math training allows you to turn off the threat response stress system, that's what the training is for!
The things that you pay attention to, you get better at.  He's impressed with the smile detecting device.
Neural programmer 2: software with sound training.
There is EMDR software.
Emotional n-back - squares have faces, words have emotional content - it's supposed to improve your emotional regulation.
Toba score - a measure of your ability to pay attention.
Most surprising - gratitude and forgiveness really required to get IQ improvements.  Not just mentally - has to be more than you pretending.
Biocybernaut testing - 10 day $20k retreat to do neural feedback training.

Ignite talks
A factory line of awesome
Doctors - reactive and paternalistic

Habit session
The muggle chasm - QS right now is for male geek wizards.  How do we cross the chasm to regular people.
He thinks "behaviorism 2.0" is the best thing for habit design.
Conditioning the environment to support the new behavior - shoes by the side of the bed
Exponential fee structure if you don't follow through with your habit
He no longer believes that gamification can lead to behavior change - it ultimately backfires, especially anything involving unbroken trends, etc.
Triggers like timers and alarms as things to hang habits on, eg calling friends.
B.j.fogs behaviors model
Book: the power of habit
A habit follows a que, then a routine occurs, and then there must be a reward.  Without all three, the habit will not last
Book: the will power instinct
"the science of small wins"
Negative habits - replace the routine, but not the que.  E.g. Eating cookies, keep the hunger queue, but replace cookie with socializing in cafeteria: same reward of getting out and seeing people.
Anchor habits - the big things in your life that happen every day.  Many routines are triggered based on the structure of the day, e.g. Getting up to go to work

Qs show and tell
600 nights of sleep data
Sleep charts are like fingerprints
Zeo is her mechanical mom
Best thing she learned - see the hidden part of you! (when you are asleep) - she tells how to visualize your sleep like she does - lines of color

Stephen Cartwright 
Visualizing data using mechanical systems with ardunio - cool programmable bar chart and 3d chart thing.
Really cool vizualizations of where he has been - GPS data for 15 years!

Seth Roberts
Brain tracking: what I've learned recently
Humans enjoy making skilled movements
People are innately specialists.  He has a theory of evolution about how hobbies were the stepping stone in our evolution from everyone making a living the same way (like any other species of animals) to everyone making a living in a different way.
A sign that you are omega3 deficient: bad gums

Body transformation
Wanted to live a more brauny life
Three types of motivations: vanity (looking good naked), performance (eg at sports) and health.  He claims health is less common!

Adafruit sells a $30 polar chest strap, new this week

Sunny Bo, with misfit ?.  Amar works with MCten
Sunny is looking at accelerometers, pressure sensors, GSR.  Things that you clip on your clothing, and things that you would wear but not strap on - no skin contact.  One of the great possibilities for wearables is real-time feedback.  E.g. Real time GSR sound feedback to overcome elevator phobia.

Amar is thinking mostly about the power that the sensors take.  Stickers, band aids, tattoos.  Conformativity to the skin is super important for them.  Looking at preventing injuries in athletes, e.g. Looking at hydration.  They want to eliminate the compliance issue. They plan to scavenge power from your cell phone.

Quantified Self Conference, Day 2, Sept 16th 2012

Cognitive measurement - by quantified mind people.
Yoni first got into this because he wanted to stop aging.
Trying to build stuff that is repeatable, fast and efficienct.
Did a year of research in psychometrics
They have also been doing brain scanning stuff (open source) using the emotiv.
They have been experimenting the doing tests while wearing it, and recording the data.
Have managed to increase alpha doing this.
The emotiv puts pressure on your head, Jacob says he can only wear it for about 15 minutes before it becomes too painful.
Yoni says that he can't think of a test on quantified mind that isn't heavily G loaded
Analogy between muscle testing and mental testing - its hard to measure your strength without depleting your muscles, and ultimately making them stronger.  But the mental change due to testing seems a lot faster than they change in muscles - and this does matter.  It complicates things if you do multiple tests in a row, does it matter in other ways?  There doesn't seem to be anything practical we can do about it.
Analogy to early days of AI - we are probably naive about how much will actually be possible with simple tests in terms of cognative measurements.  But continuous measurements like EEG offer some small hope...
Processing speed, executative function, working memory, several other things
He hasn't been able to make a repeatable test for creativity, despite lots of research...
Commenter suggests using mechanical Turk to get human judging of such a test...
Practice effect - he has a huge data set of 100,000 people doing tests repeatedly, and it turns out the practice effect is better modeled with plateau and burst than power law.
Can you use the practice effect itself as a test?  Totally, he has seen some stuff on this.
Quantified mind was built to measure, but he can get a lot more users by claiming its a tool to improve!

Ignite talks
Hind hobeika
Shutterfly swimming adapted pulse rate sensor, adapted from open source pulse sensor, uses rgb led to display in corner of vision.

Automated activity tracking
How to not drain battery?  Bursts, adaptive sampling, upload strategy

Natalie mckeever, internal worlds - other species

iPhone rhythm strip
Clinical implications of wireless and ubiquitous heart rate monitoring

Ellie's log
Quantified baby 

Human systems debugging

The quantified doctor
Started "my doctor" private practice
Needed "quant coach" on the medical team to help patients self track

Heart rate variability.
RR intervals - the time between beats of your heart
30 years of research on it, 5000 papers in NIH alone
There are studies comparing chest straps to real EKG in terms of HRV, polar and ANT+ are both good these days
HRV is created by tug of war between parasympathetic and sympathetic nervous system.  It's trying to keep a more constant blood pressure, and other things.
Vageltone corresponds to HRV, Vegas nerve
Even athletes with low resting heart rate but low HRV can have heart failure
HRV is measure of your adaptability
It has a circadian rhythm - lower at night
Age, gender and ethnicity dependent!  Declines with age, of course
HRV is also a proxy for will power!
Time domain, frequency domain and non-linear measures of it - a whole series of acronyms.
SDNN = standard deviation of NN which is similar to RR 250ms to 50, 25 is cutoff for sudden death, done on 5minute window
RMSSD root mean square
PNN50 percentage of intervals that are not different by more than 50 ms
Low LF  - .04 to .15 sympathetic
High HF  .15 to .4 parasympathetic
Looking at the balance of the nervous system - LF/HF ratio.  <2 is great. In morning, close to 1 or a little less than 1.  If you meditate, or do coherency training, it all goes into LF (.1hz, called the meditators peak), very little in HF, so it looks bad.
Pretty plots, fractal stuff, obviously different in different states.  Wavelet analysis.
There is actually research showing (dean ornish) that there is causal change in improving your HRV to improve your heart health.  So it isn't just an indicator
eustress - good stress that motivates us to do things - still associated with high HRV, you're in the zone
They have an app: Sweet Beat.
Open source data analysis - kubios.  It's a tool you can get and import HRV time series

Jo Beth Dow  - heart rate variability show and tell
green exercise - in the park 
Alcohol is terrible for her HRV and makes her resting heart rate higher too - even just a glass or two 
Quantified marriage breakdown
She thinks the weight loss helped lower her HRV 
HRV is low when you are exercising, cause the beat to beat time is so low already. 

Elliot Hedman, GSR during a concert
Noisy to quiet - all three people responded
He was the only person to respond to xylophone, he used to play it
Any transition, e.g. To whispering, gets peoples attention
Emotional stuff - look into core affect at MIT media lab.  

Lisa betts-LaCroix
What i learned about tracking by tracking
Behavioral tracking + weight tracking = success.  Weight tracking alone does not work!
Tracking weight is really just a proxy for mindfulness about your life.

Jonas on spaced repetition
Uses a scanner to grab sentences from things he reads, puts definition into SuperMemo.
Familiarity increases interest
12500 flashcards, 50 minutes a day, 90 minutes on the weekend
He is now using it to memorize poetry!
SuperMemo is not for learning - you need to learn it first, then review to remember with SuperMemo

Free posters at the registration desk!

Project lifeslice
The onion: 90% of time spent staring at glowing rectangles
What how learn next (these are the three QS questions + typical follow up :-)
Webcam snapshot photo
Taking photo/ screenshot once per hour is actually a good productivity hack!
Code is on github

Kevin Kelly: "your mission is to discover your mission"
The quantified century
Documentary: the century of the self
Extended self
He thinks the quantification is a step on the route to new senses - like North Paw
We'd like to just perceive the data, as a sensation
2x(n=1) > n=2, but probably this only extends into 100s, not thousands
Quantified community
New ways to measure, new things to better
Data = new gold (including bubbles and hype)
Information wants to be linked
Surveillance is proportional to personalization

Obesity is the usual cause of Pickwian syndrome which later on manifest as heart trouble symptoms. Through weight loss, this is minimized.

Jodie Simons
Post a Comment

Subscribe to Post Comments [Atom]

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]