Attention Tracking Session Notes
OK. The dust from the Quantified Self Conference has settled and I'm still trying to take stock of everything I saw and did and all the amazing people I talked to. I think the best metaphor was it was the homebrew computer club of this decade. The things I saw and heard about there are what regular people will be hearing about ten years from now, but I'll save the remainder of my summary for another post. I'll have a longer write-up of things shortly, but sufficed to say it was an incredibly inspiring weekend. You'll have to forgive the long delay in writing this. The last 2 weeks in San Francisco have been a jam-packed schedule of work meetings and social meetings. I've also been dedicating a large amount of time to thesis writing, and my brain can only produce so many words per day.
I moderated a breakout session on attention tracking. It was a very last-minute job of pulling my resources together and I wasn't sure how useful or well-received it would be. Much to my relief, it ended up being a great success, not by my doing but the wealth of knowledge brought by the people that chose to attend. This same breadth of topic is also partially responsible for my tardiness in posting this summary. It's proven very difficult to distill an overview of all the points made during what ended up being a very eclectic breakout session. What follows is an attempt to both summarize and elaborate on that discussion.
Like the breakout, I'll start this overview with a bit of information about myself and my own interest in the topic. Like most Quantified Selfers, my interest in attention tracking has come out of a need to solve my own problems. I was diagnosed with attention deficit hyperactivity disorder (ADHD) in high school. While I have a prescription for medication, I've spent a significant amount of effort over the years trying to introspect how my brain works (and and doesn't work) with regards to my ability to focus on a given task.
Over time, I've come to the realization there is a fundamental difference in my brain manages my focus, compared to other people. This may seem obvious, but isn't exactly something they sit down and run you through when they diagnose you. I concluded that there are two classes of distractions I deal with on a regular basis: internal and external. External distractions are such things as people interrupting you, phone calls, noises, or movement in your visual field. While external distractions can easily reduce my focus to near-zero, they're pretty easy to control. Sit in a quiet room without a lot going on, and I can remove most of them.
However, what surprised me is that doing so didn't really decrease how distracted I felt. My brain generated a lot of distractions on its own. The severity depends on context, but most of what I do requires me to sit at a computer where the possibilities of distracting myself are effectively infinite. While I can distract myself without any external stimuli, I think the never-ending steam of new stimuli endemic of modern life is making most people behave more like me. Attention is no longer an ADHD problem, it's an everyone problem.
While not explicitly discussed in the session, the question of "What is attention and how do we define it?" presented itself in many forms. Personally, given my ADHD background, I tend to define attention as the absence of distractions, internal or external. A lower number and less frequency of occurrence is a good measure of my current level of focus.
Attention, however, is a multi-faceted issue. There are many other interpretations, such as only attending to one task, but without much intensity. In each circumstance, it's worth considering exactly how we're not paying attention and what that says about our relation to the subject matter.
Many of the participants had a more goal- or productivity-oriented definition of attention, more similar to efficiency or perhaps effectiveness. Discussion included the concept of work sprints, including the Pomodoro technique, for sustaining high level of focus for brief periods, then giving the mind a break. The discussion of such blocks of time and where distractions could be permitted lead to Paul Graham's maker schedule vs manager schedule and the acknowledgement that no number of 20-minute blocks of time could be put together to accomplish something that requires deep insight. I would consider the sprinting a guard against internal distractions, effectively making a contract with yourself that you can act on the distractions, as long as you do so later, while scheduling methods would be a way of guarding against external distractions. Another participant suggested the extreme programming practice of pair development where the social pressure of a colleague could help suppress urges of self-distraction.
Discussion on productivity also touched on Parkinson's law and how tasks expand in complexity to fill all available time. Several participants admitted that they were much more productive in crunch time, where the imminent deadline gave them no option but to focus on the task if they hoped to complete it in the allotted time.
All of these techniques address issues regarding our collective unwillingness or inability to operate with only one set of stimuli. There's a growing realization of exactly how poor our attention processes have really become and the level to which they affect our ability to complete tasks. Clifford Nass's paper Cognitive control in media multitaskers demonstrates that heavy media multitaskers are poor at multitasking and at a number of cognitive control processes. In effect they untrained their brains how to manage only one set of stimuli. The situation goes beyond issues of simple multitasking. Linda Stone coined the term continuous partial attention, which touches on the broader range of issues related to how we interact with connected devices:
Continuous partial attention is an always-on, anywhere, anytime, any place behavior that involves an artificial sense of constant crisis. We are always in high alert when we pay continuous partial attention. This artificial sense of constant crisis is more typical of continuous partial attention than it is of simple multi-tasking.
In this case we willfully distract ourselves because we're afraid we're missing out on something. There could be another email, another status update, another news article. On top of the distractibility, there is an added layer of anxiety that permeates so much of modern life. I think this issue is best summed up by, ironically, a tweet by Alain de Botton:
A chief effect of the internet is to boost the already unhelpfully strong sense that the answers are 'out there' rather than within.
There's a certain friction in much of what we do. In programming, writing, etc... you have to manipulate some form of language in your brain. You simply can't do this while interacting with another person^1. It was suggested a lot of our distractions come from social services and perhaps much of the issue is due to the fact that we're social creatures that are lonely doing solitary things.
The discussion also touched upon what mental state was most suited to a particular task. Interestingly, there are striking similarities, in terms of the quality of focus, between flow and the particular flavor of intense focus on meaningless activities often exhibited by people diagnosed with ADHD. While the ADHD-focus might switch slightly more often, in both cases the ego and external references of time and other stimuli fall away and only the subject matter remains. However, flow is deemed a desirable quality, while intense ADHD self-distraction is not. Clearly there's some complexity in labeling and assigning value to tasks and types of attention.
From there, the topic shifted to how best to engineer particular mental states and types of focus. There was a discussion of pharmaceuticals that can alter mental states, including some of the more common substances like Adderall and modafinil, as well as melatonin. There's a lot to say with respect to these, but I'll leave it at a short comment about Adderall, which I know best. I've come to realize that Adderall produces a singly-focused mental state, well-suited for well-defined, linear tasks but that can hamper your ability to make many loose connections and synthesize new information and novel concepts. Other methods of altering mental state included how fed you were. I noted that for my particular dietary choices^2, I tend to feel more focused after fasting for 10-12 hours. However, a Type I diabetic also suggested I might be slightly hypoglycemic, so who can say.
All of these thoughts beg the question "in the context of attention for a particular activity, is there one particular brain state best suited to all activities?" Personally, I would say "no". I find myself much better to focus on many small lots of small tasks like email replies before noon, while I'm best able to focus for long periods on a single task that requires deeper insight from 11pm to 3am. The process of discovering these blocks in my day has taken me a fair bit of time, exploring different routines. There's the issue of determining which time is best for which activity. But even before determining optimal times; there's the deeper issue of identifying and naming the different states your mind might be in. Cataloging the different mental modes has taken me time and a lot of introspection.
My suspicions on mental states seem to be at least partially supported by evidence. Tansy Brook, an employee of Neurosky, discussed her company's work on measuring alpha and beta waves to understand focus and relaxation. Alpha waves correlate to states of relaxation and beta waves relate to focus. While brainwaves only roughly correspond to specific brain activity, some differences in focus can be determined from the amplitude of waves – for example, math problem produce higher beta waves than reading. In traditional cognitive tasks, such as said math problems, subjects tend to show high focus but much lower relaxation. On the other hand, Olympic archers tend to be both very focused and very relaxed. With athletes in general, there is a mix of alpha and beta waves, suggesting both relaxation and cognitive engagement.
This line of inquiry leads to the last major point -- which didn't get much time in the discussion -- the issue of how you actually track attention and focus. EEGs provide one means of attempting to quantify the level of focus. But w;hat options are available that don't require a laboratory setting? Surprisingly, there isn't much I could find. Cognition and learning tests, including the sustained attention to response task (SART), offer a possible means of (intrusively) trying to quantify focus at any given instance. My friend Kyle of OpenYou also suggested the incorporation of other sensors like pedometers, webcams or Microsoft Kinect to try and build up a picture of our focus. However, most sensor-based methods would require some calibration against an external "gold standard" measure of attention and focus. I'm not sure what such a measure would look like, if it even exists.
As I mentioned in the session, I ended up quickly modifying a window tracking^3 on Github to provide detailed information about how long each window is in the foreground. Beau Gunderson ended up producing a similar logging app for Windows. These are rough sketches at best and provide attention data in only the rawest sense. To be meaningful, the output would need to be compared against some external measure of focus to correlate to something meaningful.
With that, I'll try and wrap this up because the post is already far too long. There are a lot of other topics I didn't get to touch on here. For more notes from the session, please have a look at:
If this topic interests you, I've set up an Attention tracking Google group to continue discussion. Things are just getting started -- mostly due to my lack of effort -- but should hopefully pick up shortly. Some topics I'd like to see explored further include:
- What aspect of attention tracking interests you most? Number of distractions? Focus/intensity? Productivity measures? Something else?
- How would measure it? Ideally, both on the computer and off?
- Given most people showed up wishing they did a better job at their respective aspect, what change or experiment would you propose for making use of your metric?
To conclude, a lot of understanding attention requires a deeper understanding of your own mind, both to introspect and identify its current state and to apply specific manipulations to alter that state. In all cases it's slow process of gaining insight and exercising a "muscle" that you never knew you had.