*whisper* printed shirts

Finally, special shout out to my coauthors @nessa-ikani.bsky.social and non-BS Jana Vrijsen and Noa Magusin (n/n)

More contextual research can help explain these findings a bit better, but you can also find the article here: www.nature.com/articles/s41...

Finally, we checked whether participants' mood at recall influenced answers, and found this only in the never-depressed group. This shows the importance of context in measuring bias. but also the idea of only negative bias in depression may be inaccurate: perhaps its depressive realism? (4/n)

We also asked them to try remembering what their mood was during pervious timepoints to see how mood memory may differ, expecting depressed patients to report more negative memory of mood, and less positive compare to the never-depressed group ...(3/n)

An example is our recent paper on emotional memory in depression, where we followed up depressed, remitted, and never-depressed patients in daily life to see whether there is indeed a negative memory bias, and decreased positive bias. Using EMA, we asked the three groups to rate their mood...(2/n)

Imo, I think it comes down to what you expect stimuli to do. If you're interested in the stimulus driven activity, and not sustained activity (think oddball vs nback), an impulse function should be enough. Although consider modeling RT in this case: www.biorxiv.org/content/10.1...

Of course, this work was done with a wonderful team, of whom only @fladd.bsky.social has made it here so far! n/n

Overall, wearables offer some added value that may recover signal from EMA noise, but more work is needed to examine these effects at a larger scale. Thanks for making it this far, heres a link to the paper and a cat gif. www.jmir.org/2023/1/e39995/
#ESM #EMA #psychology #science

Finally, we wanted to see whether wearables offer any added value with respect to stress detection compared to EMA alone. While EMA models did better, combining both measures improved predictions! Moreover, idiographic approaches (LOBO) vastly outperformed nomothetic ones (LOSO). 4/n

So we did a time resolved analysis and saw that momentary arousal went up with stress...we end up seeing that arousal is of course not valence specific. Experiencing fewer positively valent emotions in the stress/exam week resulted in on average lower arousal 3/n