Excited to share this paper! My co-authors and I did a deep dive into smartphone language to understand indicators of suicide risk in adolescents. One of our goals was to compare AI indicators to clinician ratings. Take a look!

great perspective! I would love to hear your thoughts on alignment and synthesis with passive smartphone language sensing approaches.

Perspective:
Open-ended responses can improve ESM data by grounding it in real-world experiences and phenomena as they are experienced in everyday life. Handbooks and guidelines on ESM should include sections on collecting and analyzing open-ended text items.
www.nature.com/articles/s44...

When attention falters: brain, breathing, and behavioral signals of lapses in interoceptive attention https://www.biorxiv.org/content/10.64898/2026.02.07.704566v1

Thank you to my wonderful co-authors. @skhalsa.bsky.social @ch-verdonk.bsky.social @sgoldbergphd.bsky.social @jepusto.bsky.social and others

c) studies could incorporate objective measures of interoceptive accuracy (e.g. respiratory threshold discrimination) to examine relationships between subjective experience and objective signals.

b) the vast majority of studies employed the MAIA self-report scale, making it difficult to disentangle differences in subjective experiences

a) no studies incorporated youth, which is a critical period for emotional and cognitive development.

We make a number of recommendations for future mindfulness research:

Instead, attitudes may be key. Namely, a more accepting and non-judgmental attitude towards all experiences.

We also were unable to find any effects of intervention length, dosage, or quality. Effects were similar for interventions compared to active or passive controls. This suggests that the meditation practice itself might not be key for interoception benefits.

However, there were a number of questions that we were unable to conclusively answer. We don't currently know what aspects of subjective interoception change the most, who benefits the most, and how changes are related to decreases in mental distress.

Our meta-analysis strongly supports this theory, finding consistent, low heterogeneity effects of mindfulness trainings beyond control conditions on self-reported interoception. We are also confident in this finding because there was no evidence of publication bias.

Self-reported interoception refers to the subjective experience, regulation, and awareness and of internal bodily signals. Mindfulness may have mental health benefits because it helps cultivate more adaptive and positive subjective interoception.

How does mindfulness change body awareness? Excited to share our meta-analysis on mindfulness training and #interoception. www.nature.com/articles/s41...

Nice to see this out! Congratulations!

Thanks to my co-authors! @sgoldbergphd.bsky.social @paulbloom.bsky.social @auerbachlab.bsky.social

Future studies could consist of developing better personalization algorithms to improve signals of focus and calm and conduct observational studies to witness the use of devices in the wild.

My fellow researchers and I are bullish about the possibility of technology making mindfulness more accessible and fun for new people, but bearish about the ability of consumer-grade technology to measure signals with high fidelity. Lab-based studies with fMRI and EEG may perform better.

Small decreases in stress are observed, but no effects on cognition, mindfulness, brain measures from the devices, and physiology are observed. Placebo effects have not been conclusively controlled for.

Our findings show that there is so far comparatively little evidence that you can enhance your meditation with devices like MUSE.

There is a lot of hype out there for devices like MUSE that can show you your real-time brainwaves while you meditate. These devices convert brainwaves into signals of 'calm' or 'focus', and the goal is to modulate them.

Can neurotech devices help you meditate???
Our meta-analysis analyzing the results of RCTs and randomized studies of neurotechnology assesses that question, published in #JMIR, www.jmir.org/2025/1/e68204

This research could not have been conducted without the support of my co-authors , as well as NIH funding.

For the sake of neuromodulation paradigms to enhance attention and facilitate meditation, we may want to focus more on states that are sensitive to actual performance.

This suggests a dissociation between brain correlates of overt task differences (motivation, arousal) vs brain correlates of attentional focus .

This supports the possibility of different network configurations during tasks. Interestingly, a set of different states (like a globally connected state) was related to fluctuations within-people during the task itself.

Then we examined breathing vs rest, as is typical in most studies, but also within-task fluctuations in attention, which is a more novel step. Using dynamic brain state analysis, we uncovered states that were more present during the breath counting , like a state involving the default-mode network.

We asked seventy adolescents to spend 20 minutes counting their breaths in the MRI scanner (breath counting is an important stabilizing practice in many meditation traditions), and then also imaged them at rest. We collected performance and physiology during the breath counting.