, , , , , , , ,

People with mood disorders seem to discuss triggers incessantly because knowing your triggers and communicating them to others is a really valuable skill. I haven’t talked much about my triggers, other than sleep, but will get to it when I have something coherent to say.

But I do have another nifty tool provided by Moodscope to help better understand triggers. It’s called a Triggergram, “diagrammatic representations which may suggest what lies behind your changing moods.” It uses the text data that you can optionally enter as notes on mood scores (the digital version of a mood journal) to generate simple tag clouds with text size based on frequency, and allow click-through to see the notes that include those words. It will soon become part of a premium service while the core services will remain free.

As an unrepetant geek, I use something of a controlled vocabulary for these notes. It’s a lot like tagging and forces me to be a bit more reflective about the day’s events and emotions. Using fewer words means that more of them are included when aggregated, and the relative frequency becomes more meaningful. For example, instead of using both “anxiety” and “anxious”, I (try to) just use “anxious”. When I don’t do that, I get less useful words showing up.

I don’t like to enter a note for every day because it makes it harder for me to see at a glance on which days something important happened on my Moodscope graph if they are all highlighted to show that there’s a note. I do enter details for particularly high or low days, so my Triggergram has actually verified some intuitions about mood triggers. I enter some most of these notes after the fact based on my more detailed paper mood journal, when I can be more reflective about my mood states.

My Moodscope Triggergram for October 2011 through March 2012

The positive themes are shown in warm colors on the left, and negative ones in cool colors on the right. A few observations:

  • I can tell when I am (or have been) hypomanic, partly because I feel really excited.
  • Hypomania is often triggered by travel, which relates to the terms travel, London, social, conference, and meeting.
  • Social activities also make me happy. And possibly hypomanic. Or being hypomanic makes me more social. Either way.
  • Sunny weather might be a trigger for hypomania as well.
  • When I feel lousy, there’s a lot of tiredness, difficulty waking, grogginess, and anxiety.
  • Crying indicates feeling awful. Obviously. I end up in tears more than I would have expected.
  • Med changes tend to be tough. All the medications (except Lamictal) show up on the negative side, as do mg, change, med, meds.
  • Headaches make me feel rotten and are usually caused by med changes. The word headache shows up a lot more frequently because the change of meds only gets noted once, but the headaches sometimes go on for weeks. You can bet that if I had a med change in the last week, I’ve also had a headache all day every day for the last week.
  • Some of the words are easily clustered: adventure+hiking, submitting+manuscript, and postdoc+offer, for example. Most are one-time mentions. Meaningful, yes, but not particularly useful trends since they were singular events.

Many of these observations were things I already knew, but this helps verify it. The influence of sunny weather is notable; I hadn’t realized how much of a difference it makes. Compared to earlier versions of the Triggergram, recent med changes have been really difficult – previously, Lamictal showed up on the left and words like “mg” were evenly distributed – so this is a relevant shift.

There’s a bit of room for improvement in the representations, but this is a great start. I’d love to see options or refinements related to: stemming, 2-word phrases, click-to-remove, tagging vs annotation, non-elimination of contextually meaningful stop words like “up” and “good”, allowing selection of date ranges, and aggregating more than 10 days out of 180. Even so, the Triggergram is very helpful for developing new insights and verifying suspicions when it comes to mood triggers. It’s a great complement to mood charts and stats for learning about mood swings.